Microscopía virtual en el diagnóstico de rutina y la ...

Microscopía virtual en el diagnóstico de

rutina y la docencia en un hospital

universitario

TESIS DOCTORAL

Adela Saco Álvarez

Microscopía virtual en el diagnóstico rutinario y la docencia

[2]

Tesis Doctoral. Adela Saco Álvarez

[3]

Facultad de Medicina

Departament de Fonaments Clínics

Microscopía virtual en el diagnóstico de

rutina y la docencia en un hospital

universitario

Tesis Doctoral presentada por: Adela Saco Álvarez

Directores: Jaume Ordi Maja y José Ramírez Ruz

Barcelona, 2017


[5]

Jaume Ordi Maja y José Ramírez Ruz, profesores titulares de Anatomía Patológica de la

Universidad de Barcelona, certifican que la tesis doctoral titulada “Microscopía virtual

en el diagnóstico de rutina y la docencia en un hospital universitario” y presentada

por Adela Saco Álvarez, ha sido realizada bajo su dirección y cumple todos los

requisitos que dicta la normativa vigente para la presentación de tesis doctorales como

compendio de publicaciones en la Facultad de Medicina de la Universidad de

Barcelona.

Dr. Jaume Ordi Maja Dr. José Ramírez Ruz


[6]


[7]

Agradecimientos En primer lugar, me gustaría dar las gracias a Jaume y Rami por su inestimable ayuda y

porque sin ellos este trabajo no sería posible.

También me gustaría agradecer a mi familia, tanto gallega como leridana, y en especial

a Jose por apoyarme siempre y a Amelia por sus constantes sonrisas.


[8]


[9]

PUBLICACIONES INTERNACIONALES QUE COMPONEN ESTA TESIS DOCTORAL

Estudio 1

“Validation of Whole-slide imaging for histopathological diagnosis: current state”

Adela Saco, José Ramirez, Natalia Rakislova, Aurea Mira, Jaume Ordi

Pathobiology 2016; 83:89 – 98

Factor de impacto (2016): 1.703

Ranking (2016): 60/193, segundo cuartil

Estudio 2

“Validation of Whole-Slide Imaging in the Primary Diagnosis of Gynecological

Pathology in a University Hospital”

Jaume Ordi, Paola Castillo, Adela Saco, Marta del Pino, Oriol Ordi, Leonardo Rodríguez-

Carunchio, Jose Ramírez

Journal of Clinical Pathology 2015 Jan; 68(1): 33 - 9


Ranking (2016): 33/193, primer cuartil


[10]

Estudio 3

“Validation of Whole-Slide Imaging in the Primary Diagnosis of Liver Biopsies in a

University Hospital”

Adela Saco, Alba Diaz, Monica Hernandez, Daniel Martinez, Carla Montironi, Paola

Castillo, Natalia Rakislova, Marta del Pino, Antonio Martinez, Jaume Ordi

Dig Liver Dis. 2017 Jul 19. pii: S1590-8658(17)30977-5.

doi: 10.1016/j.dld.2017.07.002. [Epub ahead of print]


Ranking (2016): 35/134, Segundo cuartil

Estudio 4

“Current Status of Whole-Slide Imaging in Education”

Adela Saco, Josep Antoni Bombi, Adriana Garcia, José Ramirez, Jaume Ordi




https://www.ncbi.nlm.nih.gov/pubmed/28780052


[11]

Estudio 5

“Virtual Microscopy in the Undergraduate Teaching of Pathology”

Oriol Ordi, Josep Antoni Bombí, Antonio Martínez, Josep Ramírez, Llúcia Alòs,

Adela Saco, Teresa Ribalta, Pedro L. Fernández, Elias Campo, Jaume Ordi

Journal of Pathology Informatics 2015 Jan 29; 6:1

Factor de impacto (2016): 0


[12]

ÍNDICE

Lista de abreviaciones utilizadas…………………………………………………………………………. 15

INTRODUCCIÓN…………………………………………………………….…..………………………………………….. 17

1. Historia de la microscopía virtual…………………....................................... 19

2. Ventajas de la microscopía virtual

2.1 Ventajas de la microscopía virtual en el diagnóstico

rutinario……………………………………………………………………………

21

2.2 Ventajas de la microscopía virtual en la docencia……………. 23

3. Inconvenientes de la microscopía virtual

3.1 Inconvenientes de la microscopía virtual en el diagnóstico

de rutina..…………………………………………………………………………

25

3.2 Inconvenientes de la microscopía virtual en la docencia….. 26

4. Microscopía virtual en el diagnóstico de rutina

4.1 Integración del sistema de microscopía virtual en un

servicio de Anatomía Patológica……………………………………….

27

4.2 Validación de la microscopía virtual en el diagnóstico

rutinario……………………………………………………………………………

29

4.3 Estado de la validación en el momento actual………….……... 31

4.4 Aspectos técnicos en la digitalización de rutina………………… 32


[14]

4.5 Almacenamiento de archivos……………………………………………. 35

5. Microscopía virtual en la docencia……………………………………………………. 35

6. Microscopía virtual en comités multidisciplinares……………………………… 36

7. Microscopía virtual en la teleconsulta………………..……………………………… 37

HIPOTESIS DE LOS TRABAJOS……………………………………………………………..…………………………. 39

OBJETIVOS…………………………………………………………………….………………………………………………. 43

TRABAJOS REALIZADOS, MÉTODOS Y RESULTADOS………................................................. 47

ESTUDIO 1: Validation of Whole-slide Imaging for Histopathological Diagnosis:

Current State……………………………………………………………………………………………………

51

ESTUDIO 2: Validation of Whole-Slide Imaging in the Primary Diagnosis of

Gynecological Pathology in a University Hospital…………………………………………….

63

ESTUDIO 3: Validation of Whole-Slide Imaging in the Primary Diagnosis of

Liver Biopsies in a University Hospital………………………………………………………………

73

ESTUDIO 4: Current Status of Whole-Slide Imaging in Education…………………….. 83

ESTUDIO 5: Virtual Microscopy in the Undergraduate Teaching of Pathology…. 95

DISCUSIÓN…………………………………………………………………………………………………………………..... 103

CONCLUSIONES………………………………………………………………….………………………………………… 119

BIBLIOGRAFIA…………………………………………………………………………………………………................ 123


[15]

LISTA DE ABREVIACIONES UTILIZADAS

FISH: Hibridación in situ con fluorescencia (del ingles: fluorescence in situ hybridization)

H&E: Tinción de hematoxilina-eosina

H-SIL: lesión escamosa intraepitelial de alto grado (del inglés: high grade squamous

intraepithelial lesion)

LIS: Sistemas de información de laboratorio (del inglés: laboratory information

systems)

L-SIL: lesión escamosa intraepitelial de bajo grado (del inglés: low grade squamous

intraepithelial lesion)

MC: Microscopía convencional

MV: Microscopía virtual

QR: Respuesta rápida (del inglés: quick response)


[16]


[17]

I. Introducción


[18]


[19]

1. Historia de la microscopía virtual

La microscopía óptica convencional (MC) está ligada a la Anatomía Patológica

desde el inicio de esta especialidad. En los inicios, la evaluación de las muestras se

limitaba al estudio macroscópico e histológico, siendo la MC la base, y prácticamente

la única herramienta para realizar el diagnóstico de rutina de las biopsias. En las

últimas décadas se han incorporado diferentes técnicas útiles para el diagnóstico como

la inmunohistoquímica, la cual también se evalúa usando MC, y otras como la

patología molecular que no requiere ya del uso del microscopio.

Este método de diagnóstico hace que el patólogo dependa totalmente del uso

del microscopio óptico y de la presencia física de las preparaciones histológicas

montadas sobre portaobjetos de cristal. La posibilidad de consultar casos con otros

especialistas solo se puede realizar enviando las preparaciones o los bloques, lo que se

traduce en un importante aumento del tiempo en el diagnóstico y el riesgo de pérdida

o daño del material. El uso del microscopio óptico limita también la visualización

conjunta de casos, práctica especialmente necesaria en la docencia, y requiere el uso

de microscopios con múltiples cabezales. La solución de estos inconvenientes

representaría un gran avance, tanto en el diagnóstico rutinario como en la práctica de

consultas diagnósticas, sesiones y comités conjuntos con otras especialidades, o en la

docencia tanto de pre como de post-grado.

Este escenario empezó a cambiar hace algunas décadas con el desarrollo de la

informática, al aparecer los primeros ordenadores personales [1–3]. Las imágenes

estáticas del material histológico, adquiridas mediante máquinas fotográficas

acopladas al MC, fueron el primer avance digital en aparecer, estando dirigidas

principalmente a la docencia y en menor grado a la teleconsulta. Este método

presentaba grandes inconvenientes como la deficiente calidad de las imágenes y la

imposibilidad de navegar o emplear distintos objetivos, por lo que su uso para el

diagnóstico estaba muy limitado [4]. Posteriormente aparecieron sistemas de

telepatología dinámica en tiempo real con videocámaras integradas al MC. Su uso se

destinó casi exclusivamente a la visualización remota de cortes histológicos obtenidos

de forma convencional o, en la mayoría de los casos en congelación, lo que permitía el


[20]

diagnóstico de biopsias peroperatorias aún cuando el patólogo se encontrara situado

en otro centro distante. Esta tecnología resultó también muy útil en hospitales

pequeños en los que el patólogo especializado en algún área concreta no se

encontraba en el mismo centro. La telepatología dinámica en tiempo real permitía

realizar el diagnóstico de biopsias complicadas, y disminuía así la variabilidad

dependiente del centro [5–12]. Sin embargo, a pesar de representar un gran avance,

esta tecnología no permitía la navegación remota, requiriendo personal que se

encargase de mover la platina del microscopio con el portaobjetos y de cambiar el

objetivo óptico; además, la calidad de la imagen continuaba sin ser la óptima para

realizar el diagnóstico primario.

Los rápidos avances informáticos y tecnológicos en las últimas décadas

permitieron el desarrollo de los primeros escáneres capaces de crear una reproducción

digital a partir de una preparación histológica [2,3]. Estos escáneres son la base de la

patología digital o MV, la cual permite la navegación por la preparación histológica a

diferentes objetivos. La calidad de imagen y velocidad de los primeros escáneres

dificultaban el diagnóstico rutinario; además el coste económico era muy elevado, por

lo que esta tecnología se empleaba casi exclusivamente en ciertas áreas, como en

consultas diagnósticas o en docencia, y excluía el diagnóstico rutinario [10,13–17]. A

pesar de todos estos inconvenientes, esta tecnología abría la puerta a la posibilidad

real del diagnóstico virtual de rutina con todos los beneficios que podía conllevar.

En los últimos años han aparecido en el mercado numerosos escáneres con un

coste económico más bajo, con una gran calidad de imagen y con una adecuada

velocidad en la visualización. Se desarrollaron además múltiples programas

informáticos que permiten la visualización de las preparaciones virtuales usando

distintos objetivos y con numerosas herramientas que permiten tomar anotaciones,

poner marcas o realizar medidas [17–22]. Estas mejoras han acelerado la expansión de

esta tecnología y su uso tanto en docencia como en el diagnóstico rutinario en los

servicios de Anatomía Patológica.

A pesar de la gran calidad de imagen de las preparaciones virtuales, el uso de

esta tecnología ha mostrado algunos problemas que dificultan su implementación para


[21]

el diagnóstico rutinario en los servicios de Anatomía Patológica. La principal de ellas es

la fiabilidad de los diagnósticos emitidos con este método y la necesidad de realizar

validaciones antes de realizar el diagnóstico primario de las biopsias.

2. Ventajas de la microscopía virtual

2.1. Ventajas de la microscopía virtual en el diagnóstico rutinario

La MV presenta múltiples ventajas que resuelven gran parte de los problemas

que plantea el uso de la MC. Entre los principales destaca la imposibilidad de

visualización de preparaciones por un grupo amplio de usuarios, especialmente si

estos se encuentran separados físicamente. La MV permite la navegación y

visualización en una pantalla a tiempo real de las preparaciones, solucionando así este

inconveniente. Esta característica también influye en la presentación de casos en

sesiones o comités multidisciplinares, haciéndola mucho más sencilla y mejorando la

calidad, por lo que, también se favorece la interacción entre especialistas.

Los visores digitales tienen gran número de prestaciones y permiten un rango

mucho más amplio de aumentos a los que se pueden visualizar las preparaciones, lo

que facilita la navegación. En especial el estudio a muy pequeños aumentos (<10x)

puede resultar muy útil en la evaluación de los especímenes quirúrgicos. La navegación

se ve facilitada por la presencia de una imagen del material de pequeño tamaño que

permite conocer la localización exacta del área que se visualiza en la pantalla. La

presencia de un thumbnail permite ver la imagen de la preparación convencional para

asegurar que la totalidad del material se encuentra digitalizado. La MV también

presenta varias herramientas informáticas que hacen posible rotar las imágenes,

realizar fotografías con la simple selección de un área, tomar mediciones precisas y

hacer marcas de áreas de interés o anotaciones.

Una de las ventajas principales de la MV es la visualización simultánea en la

misma pantalla de varias preparaciones, que se pueden mover y cambiar de magnitud

de forma sincrónica. Esto representa una gran mejora a la hora de comparar tinciones

inmunohistoquímicas o de localizar zonas a estudio en las distintas preparaciones.


[22]

El desarrollo de la MV ha representado también un gran avance en la

teleconsulta, debido a la facilidad a la hora de compartir casos y realizar consultas

diagnósticas a otros patólogos que se encuentran en centros alejados. El transporte de

las preparaciones deja de ser necesario, lo que conlleva una disminución de riesgo de

ruptura o pérdida del material, así como un ahorro económico al prescindir del servicio

de mensajería. Esto también implica una disminución drástica del tiempo de respuesta

por parte del patólogo consultor, pudiendo llegar a ser de minutos en lugar de días.

La MV permite la visualización no solo desde la pantalla de un ordenador, sino

que posibilita el uso de dispositivos portátiles como smartphones o tablets. Esta

tecnología en constante desarrollo permitirá consultar casos o realizar diagnósticos de

una forma más rápida y sencilla, sin importar la localización del patólogo [23–25].

Existen otras ventajas derivadas del uso de la MV en la rutina diagnóstica, como

la facilidad de almacenamiento y recuperación de biopsias antiguas. Esta característica

hace innecesaria la búsqueda de preparaciones histológicas convencionales en el

archivo; lo que resulta especialmente útil en las patologías crónicas, donde la revisión

de biopsias previas es muy frecuente. A esta ventaja hay que añadir el ahorro de

tiempo y la disminución de la posibilidad de pérdida o daño del material. Además, las

preparaciones digitales mantienen en el tiempo sus características, al contrario que las

convencionales cuyas tinciones pierden color y se deterioran.

Diversos estudios muestran una buena valoración de la MV por parte de los

patólogos, destacando la calidad de imagen, la presencia de herramientas

informáticas, así como la ergonomía de los puestos de trabajo [22]. A pesar de lo cual,

aún existen reticencias al uso de la MV en el diagnóstico de rutina por parte de algunos

facultativos.

La MV también ha permitido la creación de algoritmos diagnósticos para la

cuantificación automática de células positivas con tinciones inmunohistoquímicas; lo

que conlleva resultados más objetivos, al disminuir la variabilidad intra e inter-

observador. Esta característica es extremadamente útil en algunas patologías donde

pequeñas variaciones en estos resultados pueden conllevar un cambio en el

tratamiento o pronóstico del paciente [26–29]. En la actualidad se están desarrollando


[23]

nuevas herramientas que permiten el reconocimiento de patrones histológicos que

remarcan de forma automática áreas sugestivas de patologías concretas o de

infiltración del estroma.

El flujo de trabajo del personal técnico también se ve modificado con la

introducción de la MV, y aunque la carga y descarga de preparaciones en el escáner

representa un incremento del trabajo técnico, otras tareas se ven reducidas. El reparto

de preparaciones por las distintas subespecialidades deja de ser necesario y la

búsqueda de preparaciones en el archivo se ve reducida debido a que el visor permite

la visualización de preparaciones antiguas.

El archivo de preparaciones de cristal también presenta grandes inconvenientes.

El gran espacio físico necesario y las horas de trabajo del personal técnico destinadas a

esta tarea suponen un gran gasto económico. La MV podría representar una

importante mejora si no fuese necesario el almacenamiento de algunas preparaciones

convencionales al existir la imagen virtual.

2.2. Ventajas de la microscopía virtual en la docencia

La aplicación de la MV presenta grandes ventajas, tanto en la docencia de pre

como de post-grado. En lo que respecta a la docencia de pre-grado, la MV ha supuesto

un gran avance a la hora de realizar prácticas de microscopía en asignaturas como

Histología o Anatomía Patológica. Una de las principales ventajas es que la MV permite

la visualización desde cualquier ordenador, desapareciendo la necesidad de

microscopios convencionales. El hecho de no necesitar aulas de MC hace que se

reduzca el gasto económico dirigido a su mantenimiento. A su vez, también mejora de

forma indirecta la calidad del material docente, pues normalmente los microscopios

ópticos destinados a docencia suelen tener una menor calidad [30,31]. Además, las

imágenes virtuales se encuentran siempre enfocadas y con la iluminación óptima, lo

que contribuye también a mejorar de la calidad del material docente.

A la hora de evaluar el conocimiento adquirido por los estudiantes la MV

también supone un avance, pues el hecho de no usar imágenes estáticas hace que los


[24]

alumnos reconozcan las características histológicas y no las de la imagen (forma,

tamaño del tejido, etc), evaluando realmente el conocimiento adquirido [32].

Entre las ventajas de la MV cabe destacar que no es necesario tener

conocimientos previos en el uso del microscopio convencional, lo que resulta

especialmente beneficioso, pues los estudiantes pueden centrarse en las imágenes

histológicas y no en el uso de la MC. Así mismo, la aceptación de la MV es muy buena,

lo cual es debido a que los estudiantes tienen una amplia experiencia en el uso de

dispositivos informáticos.

La navegación por la preparación digital resulta sencilla pues dispone de un

thumbnail y presenta un mayor rango de magnitudes que permite pequeños

aumentos; estas características facilitan la orientación dentro del tejido, lo que es de

especial importancia cuando no se dispone de experiencia previa. La visualización de

varias preparaciones de forma simultánea en una misma pantalla es posible, lo que

facilita la comparación e interpretación de distintas tinciones como por ejemplo de las

técnicas inmunohistoquímicas. La MV presenta también herramientas informáticas

que permite realizar marcas y anotaciones, éstas pueden ser creadas por el profesor

para indicar puntos de interés, o bien por el alumno para señalar dudas; lo cual hace

que el aprendizaje sea mucho más dirigido. Algunos estudios muestran mejores

resultados cuando los docentes realizan marcas con anotaciones sobre puntos clave

para el diagnóstico, respecto a preparaciones sin marcar [33–35].

Los programas para visualizar las preparaciones digitales destinados a la

docencia permiten completar la imagen histológica con datos de la historia clínica,

pruebas de imagen, fotografías macroscópicas o técnicas adicionales como las

tinciones inmunohistoquímicas, FISH o inmunofluorescencia [2]. Esto ofrece una visión

más amplia y completa de los casos, con el fin de acercarse en la mayor medida de lo

posible a la práctica clínica habitual.

La posibilidad de visualización simultanea por parte de varias personas de las

mismas imágenes también ayuda a homogeneizar el aprendizaje de los alumnos. El

hecho de que los alumnos puedan ver la misma imagen a la vez mejora la cooperación

entre ellos y con el profesor, contribuyendo así al proceso de aprendizaje [31,36,37].


[25]

Con la MC era necesaria la creación de nuevas preparaciones convencionales, las

cuales tenían que ser reemplazadas cada cierto tiempo debido al deterioro por su uso.

Las imágenes generadas con la MV conservan siempre la misma calidad, haciendo

innecesario nuevos cortes histológicos adicionales con la subsecuente pérdida de

material. Gracias a esta característica, es posible incluir dentro del material docente

citologías y biopsias de pequeño tamaño, cuyo estudio se encontraba limitado por el

temor a perder material o a que éste fuese requerido para ampliar las pruebas

diagnósticas del paciente [38]. Lo mismo ocurre con los casos a consulta, en los que se

puede devolver el material una vez digitalizado con fines docentes. También cabe

remarcar el ahorro de tiempo del personal técnico a la hora de archivar y realizar

cortes histológicos destinados a la docencia.

Otra ventaja muy valorada por el alumnado es la posibilidad de acceso a las

preparaciones desde cualquier ordenador y en cualquier momento, facilitando el

estudio y eliminando las restricciones del uso de laboratorios de microscopía después

de las horas lectivas [33–35,39,40].

Por último, la MV permite la confección de series de casos de forma fácil y sin

gastar material de las biopsias. Este material docente resulta muy beneficioso

especialmente en estudios de post-grado y para la formación continuada de los

especialistas en Anatomía Patológica. Los patólogos en formación pueden centrarse en

reconocer patrones histológicos importantes para realizar los distintos diagnósticos. A

su vez, es posible compartir casos de patologías poco frecuentes para ayudar a

homogeneizar aún más el aprendizaje entre los distintos centros.

3. Inconvenientes de la microscopía virtual

3.1. Inconvenientes de la microscopía virtual en el diagnóstico de rutina

La principal limitación de la MV es la inversión económica que conlleva, tanto en

su implementación en un Servicio de Anatomía Patológica como en su posterior

mantenimiento. Aunque en los últimos años los precios de los escáneres han

disminuido, todavía representan un importante gasto económico, a lo que hay que

sumar la creación de puestos de trabajo que contengan un ordenador con un soporte


[26]

adecuado para esta tecnología y una pantalla de alta resolución que permita su

correcta visualización [41,42].

El personal técnico es una pieza indispensable en el proceso de digitalización, por

lo que resulta necesario que al menos dos personas tengan conocimientos en MV y

dediquen parte de su tiempo a la carga y descarga de preparaciones en el escáner, así

como a la resolución de incidencias que puedan presentarse durante el proceso. Es

necesario tener en cuenta que el incremento del trabajo del personal técnico se ve en

parte contrarrestado por ciertos beneficios comentados anteriormente como el ahorro

de tiempo a la hora de repartir preparaciones entre las diferentes áreas o en la

búsqueda de preparaciones antiguas en el archivo.

Una vez instaurado todo el sistema, el principal inconveniente es la necesidad de

espacio virtual para el almacenamiento de los archivos. Las imágenes generadas por el

escáner presentan una gran calidad, lo que se traduce en un gran tamaño de archivo,

frecuentemente más de 2 GB por preparación. Son necesarios servidores de gran

capacidad, así como estrategias de reducción de tamaño de los archivos, como la

compresión de archivos o el hecho de escanear con el objetivo de menor magnitud

que permita un correcto diagnóstico [41].

Otro problema añadido es la reticencia de algunos patólogos a abandonar el uso

del MC, aunque las ventajas de la MV son numerosas y cada vez mejor conocidas. La

mayoría de las principales quejas giran en torno al aumento de tiempo empleado en la

visualización de las preparaciones virtuales, respecto a las convencionales. Algunos

estudios han puesto de manifiesto que existe una curva de aprendizaje en el uso de la

MV, mejorando sustancialmente el tiempo cuando el patólogo se familiariza con esta

tecnología [43–48].

3.2. Inconvenientes de la microscopía virtual en la docencia

Nuevamente la mayor desventaja es el coste económico, tanto de implantación

de la MV como de mantenimiento. Una posible solución es el uso de escáneres que ya

estén siendo usados para el diagnóstico rutinario y el uso de un programa informático

de bajo coste [33]. El almacenamiento de archivos también representa una alta


[27]

inversión económica por lo que es recomendable escanear las preparaciones con un

objetivo máximo de 200x y existe la posibilidad de alquilar terabytes en un servidor

externo.

A nivel de material docente se precisa un aula con ordenadores que dispongan

de una conexión a internet de alta velocidad, y a pesar de que es necesario realizar

mantenimiento, resulta mucho más versátil y económico que un aula con microscopios

ópticos [30,31].

Otra de las principales críticas de la MV en la docencia es que los estudiantes no

aprenden el uso de los microscopios ópticos. Sin embargo, la MV les permite centrarse

en la histología y el reconocimiento de patrones histológicos y no en el uso de la

herramienta, lo que lleva a un mejor aprovechamiento del tiempo dedicado al estudio

[31,49,50].

La interacción entre alumnos y profesor puede verse afectada por el uso de la

MV, ya que la posibilidad de visualización a distancia puede disminuir el contacto entre

ellos. También es necesario realizar controles de calidad periódicos para evaluar el uso

del sistema fuera de horario docente y valorar que imágenes y áreas son las más

visualizadas por los alumnos, con el objetivo de reunir la mayor información posible

sobre el proceso de adquisición de conocimientos por parte de los alumnos [33].

4. Microscopía virtual en el diagnóstico de rutina

4.1. Integración del sistema de microscopía virtual en un servicio de Anatomía

Patológica

La MV requiere la transformación de las preparaciones de cristal convencionales

en imágenes virtuales. En el momento actual existen en el mercado varios escáneres

capaces de realizar este proceso de forma muy eficiente sobre gran número de

preparaciones, permitiendo así procesar toda la actividad de grandes servicios de

Anatomía Patológica [17,18,20,22,51]. El proceso de digitalización de preparaciones

convencionales se realiza de forma automática, incluyendo la selección del área de la

preparación que contiene el tejido, la distribución de los puntos de enfoque y la


[28]

calibración. Cuando una preparación contiene secciones seriadas, el sistema escanea

todos los cortes presentes en el cristal. Las preparaciones deben mantener unos

mínimos de calidad en el montaje para poder ser escaneadas correctamente, lo que

también contribuye al mantenimiento del control de calidad.

Para facilitar el diagnóstico, las imágenes virtuales se encuentran ligadas a cada

uno de los pacientes. Esto sucede porque el sistema LIS de los servicios de Anatomía

Patológica se encuentra vinculado al escáner, generando un código QR que el escáner

liga con la información referente a cada caso. Este mecanismo permite ver los datos

del paciente desde el visor, facilitando su reconocimiento y limitando la necesidad de

buscar información adicional en el sistema LIS. A su vez, algunos programas

informáticos encargados del LIS disponen de un acceso directo al visor de MV,

permitiendo la visualización de una preparación virtual de forma directa.

Las preparaciones pueden ser escaneadas a distintas magnitudes; la mayor parte

de los escáneres disponen de objetivos de 20x, 40x y en algunos casos 60x. Esto

permite crear imágenes de gran calidad que pueden ser visualizadas posteriormente a

200x, 400x o 600x sin perder definición. Los programas informáticos empleados como

visores permiten realizar un zoom digital añadido a estas magnitudes, permitiendo una

visualización de hasta 400x cuando la preparación ha sido escaneada a 20x y hasta

600x cuando lo ha sido a 40x.

Los visores de las preparaciones digitales son programas especialmente

diseñados para esta función, imitando en gran medida a los microscopios ópticos

convencionales. Permiten la visualización y magnificación a tiempo real de las

preparaciones con un número de aumentos mucho más amplios que la MC. Las

imágenes se encuentran siempre enfocadas, y presentan un contraste y una

iluminación óptima en todo momento. La mayor parte de estos visores disponen de un

thumbnail de la preparación que permite comprobar que la totalidad del material ha

sido correctamente digitalizado y también presentan una imagen a pequeño aumento

para facilitar la navegación por la preparación digital. Así mismo, incorporan múltiples

herramientas digitales capaces de realizar marcas y anotaciones sobre las imágenes,

tomar medidas precisas, fotografías o incluso realizar cuantificaciones automáticas.


[29]

Todas estas imágenes son almacenadas en servidores que permiten el acceso a

los casos desde el visor, incluso después de ser validados. Debido al gran tamaño de

los archivos en muchos centros existen dos sistemas de almacenaje: un servidor se

encarga de los casos recientes permitiendo un acceso rápido y otro almacena los casos

más antiguos, con un acceso a las imágenes que tarda algunos segundos más que el

anterior. De esta forma se optimiza el espacio de disco destinado al almacenamiento

de archivos.

Estos escáneres suelen disponer de un módulo de auditoría que registra el

tamaño de cada archivo, el tiempo empleado por el escáner para digitalizar cada caso

en particular y cada uno de los accesos a las imágenes virtuales.

4.2. Validación de la microscopía virtual en el diagnóstico rutinario

El uso de esta nueva tecnología para el diagnóstico primario fue cuestionado por

muchos especialistas, principalmente por las escasas evidencias científicas sobre su

fiabilidad diagnóstica existentes al inicio de su uso. Aunque cada vez existen más

publicaciones demostrando una buena reproducibilidad entre los diagnósticos

realizados con MV y MC, su número sigue sin ser aún elevado y existen

subespecialidades en las que no existen estudios o éstos presentan deficiencias en su

diseño.

Con el objetivo de aclarar estas cuestiones y de normativizar la implementación

de la MV, la American Telemedicine Association, el College of American Pathologist y la

Canadian Association of Pathologist fueron los primeros en realizar una revisión de la

bibliografía existente en ese primer momento y elaboraron la primera guía de

recomendaciones para la realización de una correcta validación de la microscopía

virtual [3,17]. En esta revisión, realizada en 2013, se analizaban los resultados de 767

publicaciones, de las cuales solamente 112 prestaban una metodología adecuada. El

nivel de concordancia entre diagnósticos era muy bueno, variando entre el 73% y el

98% en los distintos estudios; por lo que se concluía que la MV es una tecnología

adecuada para realizar el diagnóstico rutinario. Las discrepancias mayores se

encontraban entre el 3% y el 7%, cifras bastante dependientes de las variaciones en la

metodología de los estudios, por ejemplo si la concordancia entre diagnósticos era


[30]

evaluada intra o inter-observador. Estos resultados reflejaban la necesidad de unas

recomendaciones claras sobre el proceso de validación; motivo por el cual crearon las

primeras guías sobre la validación de la MV en el diagnóstico primario. En ellas

recomiendan la realización de una validación interna en cada centro en el que se

introduzca la MV. Este proceso tiene que englobar una variedad de muestras que

refleje la complejidad de la práctica real del centro (congelados, preparaciones teñidas

con H&E e inmunohistoquímica, citologías, etc). Se considera innecesario incluir una

representación de todos los órganos y subespecialidades, pues los resultados de una

subespecialidad pueden ser usados en otra con características similares. Respecto al

número de casos necesarios, la guía sugiere 60 especímenes de cada tipo de muestras

para alcanzar una precisión cercana al 90% y una concordancia del 95%; en los

estudios analizados un incremento en el número de casos no representaba un

aumento sustancial de la precisión en los resultados.

No es necesario realizar la validación de cada uno de los componentes del

sistema de digitalización, sino que se realiza en conjunto. De la misma forma, cuando

se modifica algún componente del sistema, éste tiene que ser reevaluado en su

conjunto.

La forma más adecuada de hallar la concordancia entre los diagnósticos es con el

cálculo de la reproducibilidad intra-observador y se recomienda un periodo mínimo de

descanso entre ambas visualizaciones de dos semanas. El orden de la evaluación con

MV o MC no influye en el resultado final, por lo que puede ser aleatorio.

Dentro de los aspectos técnicos, resulta necesario comprobar que la totalidad

del material de las preparaciones haya sido escaneado; esta tarea se ve facilitada por

la incorporación en muchos de los visores digitales de una imagen en miniatura de la

preparación convencional. A su vez, se debería comprobar que las imágenes generadas

por el escáner son iguales a las recibidas por el especialista, especialmente cuando se

usan sistemas de compresión para reducir el tamaño de éstas, pues la calidad de la

imagen final puede verse comprometida, especialmente cuando se emplean

compresiones del tipo irreversible.


[31]

Se recomienda que el proceso de validación incluya a todo el personal que

posteriormente se verá involucrado en la digitalización rutinaria, siendo aconsejable

contar con la asistencia de un patólogo con experiencia en el uso de la MV. La

validación se debe llevar a cabo siguiendo los estándares más actualizados en cada

momento y se recomienda generar un documento que recoja todo el proceso por si

resulta necesario efectuar comprobaciones posteriormente.

El Libro Blanco de la Anatomía Patológica en España publicado en 2015 recoge

las recomendaciones de la Sociedad Española de Anatomía Patológica sobre la

validación de la MV. Éstas coinciden con las aconsejadas por las guías de la American

Telemedicine Association, el College of American Pathologists y la Canadian

Association of Pathologists aunque se propone un proceso de validación menos

estricto debido a la gran necesidad de tiempo y recursos requeridos para su

realización, siendo poco compatible con la práctica diaria. Puesto que cada vez existen

más evidencias científicas sobre la buena correlación entre los diagnósticos con MV y

MC, se pueden realizar validaciones menos estrictas o incluso adoptar un sistema

digital validado con anterioridad en otro centro [52]. Sin embargo, es necesario tener

en cuenta que la literatura publicada sobre la validación en el diagnóstico presenta

deficiencias en algunos ámbitos de la patología, por lo que sigue siendo necesario

completar estos estudios para facilitar la implementación de este sistema en los

servicios de Anatomía Patológica.

Otra característica a destacar es la presencia de una curva de aprendizaje en el

uso de la MV, por lo que al inicio todos los patólogos pasan por un periodo de

adaptación en el que resulta necesario compaginar ambas tecnologías para realizar el

diagnóstico, hasta adquirir una mayor seguridad con el uso de la MV.

4.3. Estado la de validación en el momento actual

En los últimos años han aparecido numerosos estudios demostrando muy buena

correlación entre los diagnósticos realizados con MV y MC. Estos buenos resultados

sumados a las numerosas ventajas de la MV, hacen de esta tecnología una excelente

candidata para el diagnóstico de rutina. Sin embargo, cuando se dividen por

subespecialidades aparecen deficiencias en muchos ámbitos de la Anatomía


[32]

Patológica, pues existen áreas que no disponen de estudios de validación o éstos

tienen una metodología inadecuada. Antes de utilizar la MV para el diagnóstico

primario de un tipo de biopsias determinadas es fundamental asegurar la existencia de

estudios de validación que engloben dicho tipo de muestras u otras con características

similares. Por esta razón resulta necesario identificar cuáles son las áreas de la

patología en las que aún no existe suficiente evidencia científica sobre el uso de la MV,

para llevar a cabo la realización de nuevos estudios de validación con el objetivo de

que esta tecnología pueda ser empleada en el diagnóstico de rutina.

4.4. Aspectos técnicos de la digitalización de rutina

Tras la instauración de la MV para el diagnóstico rutinario en un Servicio de

Anatomía Patológica es necesario tener en cuenta ciertas consideraciones técnicas que

pueden influir en el buen funcionamiento del sistema.

El uso de la MV en la práctica clínica requiere una digitalización eficiente de las

preparaciones, lo que se traduce en reducir el tiempo desde que las preparaciones

convencionales son montadas hasta que se digitalizan, para ello es necesario

escanearlas de forma frecuente, preferiblemente en más de una ocasión durante la

jornada laboral. Para que el proceso diagnóstico se ralentice lo menos posible es

recomendable que los centros dispongan de más de un escáner, pues al poder dividir

las preparaciones el tiempo se reduce sustancialmente. Además, el segundo escáner

garantiza la continuidad de la MV durante las eventuales averías que puedan surgir;

evitando cambiar el hábito diagnóstico de los patólogos que usan esta tecnología.

También es necesario tener en consideración que existen ciertos casos en los

que, debido a las características de la muestra o al estado del paciente, es

imprescindible realizar un diagnóstico lo más rápido posible, por lo que es conveniente

la creación de estrategias para priorizar biopsias urgentes.

Otra de las cuestiones que han generado controversia desde el comienzo de la

MV es la magnitud a la que se deben de escanear las preparaciones. La imagen

resultante tiene que permitir un correcto diagnóstico, minimizando al máximo el

tamaño del archivo generado para facilitar su almacenamiento. Varios estudios de


[33]

validación de diferentes subespecialidades señalan al objetivo de 20x como adecuado

para el diagnóstico, aunque existen ciertas excepciones [22,41]. Algunas

subespecialidades como la patología renal, biopsias pequeñas cardiacas o hepáticas,

así como ciertas biopsias dermatológicas necesitan imágenes de una mayor calidad,

pues pequeñas variaciones pueden condicionar un cambio en el tratamiento o

pronóstico del paciente; por lo que, aunque todavía existen escasas publicaciones, la

recomendación más extendida es escanear este tipo de biopsias con un objetivo

mínimo de 40x.

La MV también involucra en gran medida al personal técnico, que ve modificado

su flujo de trabajo. La digitalización rutinaria de preparaciones supone una nueva tarea

de carga y descarga del escáner; así como de comprobación de la correcta

digitalización de las preparaciones, que en ocasiones requiere la modificación o re-

escaneando de aquellas que han presentado algún incidente o problema de enfoque

durante el proceso [3]. Por otro lado, ciertas tareas clásicamente atribuidas al personal

técnico desaparecen o disminuyen en gran medida, como el reparto de preparaciones

por las distintas subespecialidades o la búsqueda de preparaciones antiguas en el

archivo, pues el visor virtual permite la visualización de casos previos.

El diagnóstico digital de preparaciones de material congelado de biopsias

peroperatorias, no solo es posible, sino que presenta una muy buena concordancia con

el diagnóstico con MC según múltiples estudios, con un tiempo total de respuesta que

varía entre 14 y 20 minutos [7,42,53]. El personal técnico y facultativo deben de tener

experiencia en el uso de la MV, con el objetivo de reducir este tiempo al máximo. Una

ventaja de esta tecnología a tener en cuenta en el diagnóstico de biopsias

peroperatorias es la posibilidad de visualización remota, permitiendo el diagnóstico

por parte de un patólogo especialista, aunque éste no se encuentre en el centro en ese

momento.

La MV presenta algunos inconvenientes técnicos, pues generalmente no permite

ajustar el enfoque a distintos planos de la preparación, lo que dificulta su uso para el

diagnóstico en el ámbito de la citología. Algunos escáneres presentan la posibilidad de

escanear usando un plano adicional (plano z) que simula este enfoque en profundidad,


[34]

pero conlleva un significativo incremento del tamaño de los archivos, lo que resulta en

un gran inconveniente pues dificulta su almacenamiento.

Las preparaciones convencionales deben presentar unas rigurosas condiciones,

tanto de tinción como de montaje, con el fin de asegurar una correcta digitalización;

motivo por el cual resulta necesario realizar un control de calidad periódico sobre ellas.

Así mismo, todo el sistema de MV también debe de ser sometido a un proceso

de control de calidad para asegurar su correcto funcionamiento. Las principales guías

recogen una serie de recomendaciones muy detallas sobre cómo realizar este proceso

de la forma más adecuada dentro de cada centro [3]. Entre ellas destacan la creación

de un comité específico formado por personal involucrado en el proceso de

digitalización (facultativos, personal técnico y dirección del servicio), con la función de

supervisar este proceso. Una de las funciones de este comité es realizar una

comprobación sistemática de la política sobre el uso de la MV en busca de posibles

actualizaciones. También es necesario crear una guía informativa sobre el uso del

programa informático y el hardware que debe de estar disponible para los usuarios en

todo momento. Así mismo, resulta necesaria la creación de un mecanismo para la

detección y resolución de problemas, con una respuesta lo más rápida posible. Se

recomienda la documentación de parámetros de importancia como el tiempo y la

concordancia en el diagnóstico de biopsias peroperatorias, el porcentaje de casos que

requieren revisión de la preparación convencional durante el diagnóstico y el

porcentaje de re-escaneos. Por último, proponen la revisión de un 10% de los casos

diagnosticados de forma virtual, tanto propios del centro como de consulta

diagnóstica.

Al igual que sucede en la validación de la MV para el diagnóstico, estas

recomendaciones resultan muy estrictas y su completa realización representaría un

gran consumo de tiempo y recursos por parte del personal y del centro. Algunas de

ellas, como la revisión de un 10% de los casos también están recomendadas en el uso

de la MC, pero en la práctica real su aplicación es muy baja [3,52].


[35]

4.5. Almacenamiento de archivos

Uno de los principales desafíos al que se enfrenta el diagnóstico rutinario con MV

es el gran tamaño de las imágenes, el cual genera una gran necesidad de espacio de

disco para almacenar preparaciones digitales. Cuando este archivo se suma al de

preparaciones convencionales, el gasto económico aumenta significativamente,

representando uno de los principales inconvenientes del uso de la MV en el

diagnóstico rutinario.

Se han propuesto múltiples estrategias con el objetivo de minimizar este gasto,

como la compresión de imágenes, el archivo exclusivo de las preparaciones virtuales

más representativas de cada caso, la eliminación de imágenes o incluso de

preparaciones convencionales tras un tiempo después del diagnóstico; teniendo en

cuenta que en todos los casos los bloques de parafina serían conservados.

En el momento actual no existen apenas estudios que traten este tema, ni

tampoco un marco legal que defina cuál es el tratamiento más correcto para estos

archivos; aunque algunos estudios sugieren un tiempo mínimo de conservación de las

imágenes de 6 meses [22].

5. Microscopía virtual en la docencia

En diversos ámbitos académicos como Medicina, Odontología o Veterinaria el

estudio de la Histología y la Anatomía Patológica tienen una gran relevancia.

Clásicamente la única alternativa para llevarlo a cabo era la utilización de MC y

posteriormente el uso de imágenes estáticas como fotografías de las muestras

histológicas. Este panorama cambió con la aparición de la MV, debido a que ciertas

características intrínsecas a esta tecnología representan numerosas ventajas al ser

aplicadas a la docencia. Algunas de las principales son la posibilidad de visualización de

una preparación por parte de un amplio grupo de personas, el acceso remoto en

cualquier momento del día, la facilidad en su uso y la mejora en la uniformidad del

contenido, pues el material docente es el mismo para todos los alumnos.


[36]

Existen numerosas publicaciones que muestran buenos resultados con el uso de

la MV; así como una valoración muy positiva tanto por parte de los docentes como del

alumnado [30,40,54–59].

A pesar de que estas características convierten la MV en una excelente

herramienta para el aprendizaje de la Anatomía Patológica, al igual que ocurre en el

diagnóstico rutinario, resulta necesario confirmar la no inferioridad de esta

herramienta respecto a la MC para asegurar una correcta formación de los alumnos

tanto de pre como de post-grado.

6. Microscopía virtual en comités multidisciplinares

La incorporación de la MV a los servicios de Anatomía Patológica ha

representado un cambio a la hora de la realización de comités multidisciplinares, pues

permite la visualización a tiempo real de las preparaciones histológicas requiriendo

únicamente un ordenador y una pantalla. De esta forma, se disminuyen drásticamente

las barreras entre las distintas especialidades, facilitando la interacción entre

patólogos y clínicos [22,42]. Además, permite un estudio más completo de los

pacientes al incluir las imágenes histológicas a la hora de presentar los casos, al igual

que ocurrió hace algunos años con la incorporación de las imágenes diagnósticas.

Otra ventaja es la posibilidad de participar en comités de otros centros o en

sesiones conjuntas a distancia; lo permite un mayor flujo de información, dado que

posibilita el compartir casos interesantes o realizar consultas a otros especialistas.

Existen estudios cuyos resultados revelan un impacto positivo de la MV a la hora

de preparar los casos para los comités, pues el tiempo empleado por el patólogo

disminuye aproximadamente al 50% (entre 30 minutos y 1 hora semanal de ahorro).

Esto es debido a que las preparaciones se pueden visualizar en tiempo real, por lo que

no resulta necesario realizar fotografías o movilizar material; con el fin de agilizar aún

más la visualización se pueden hacer anotaciones sobre las zonas de interés

diagnóstico [22].


[37]

7. Microscopía virtual en la teleconsulta

Una de las primeras aplicaciones de la MV fue en la teleconsulta, lo que

representó un gran avance, dado que hasta su aparición estaba basada únicamente en

imágenes estáticas que no permitía la navegación ni el uso de distintas magnitudes

[1,53,60–65]. La MV permite la consulta de casos con dificultad diagnóstica entre

distintos hospitales cuando éstos se encuentran en localizaciones remotas, áreas

rurales o cuando los expertos no están en el centro [14,66]. Además, con la reciente

expansión de la MV existe la posibilidad de establecer redes diagnósticas entre los

hospitales regionales y los centros de referencia, lo que se traduce en una significativa

mejora en el proceso diagnóstico. Estas redes permiten que grupos de expertos

realicen el diagnóstico en casos complicados, incrementando así de forma significativa

la posibilidad de alcanzar un correcto diagnóstico, lo cual es una gran mejora para los

pacientes pues se alcanza la igualdad entre ellos sin importar su localización. Otra

ventaja es que este diagnóstico experto no supone un incremento significativo del

tiempo de respuesta, pues no resulta necesaria la movilización de material, y la

comunicación entre profesionales es mucho más dinámica al poder visualizar los casos

a tiempo real. Algunas publicaciones ponen de manifiesto la gran disminución de este

tiempo de respuesta en las consultas, llegando a ser de hasta 18 minutos en algunos

estudios [66].

Otro beneficio secundario es la disminución de la necesidad de movilizar bloques

de parafina o preparaciones histológicas, evitando así la posible pérdida o daño del

material; a lo que hay que añadir un ahorro del coste de mensajería [14,65–67].


[38]


[39]

II. Hipótesis


[40]


[41]

En los últimos años hemos asistido a una creciente expansión la MV debido sus

claras de ventajas respecto a la MC, tanto en el diagnóstico primario como en la

docencia. Esta tecnología resuelve una gran parte de los inconvenientes intrínsecos al

uso de la MC. Sin embargo, esta tecnología está generando algunas dudas y reticencias

por parte de los patólogos. La mayoría de éstas surgen de la novedad de la

herramienta y de la relativa inseguridad de poder alcanzar un diagnóstico fiable con

ella. Las escasas guías publicadas sobre su uso por el College of American Pathologists,

la Canadian Association of Pathologists y la American Telemedicine Association

proponen la realización de una validación interna en cada uno de los centros en los

que se pretenda implantar esta herramienta. Esta validación podría ser una medida

innecesaria si existiese suficiente evidencia científica de su no inferioridad respecto a

la MC. Sin embargo, existen numerosas subespecialidades en las que las evidencias

existentes sobre su validación son escasas o incluso nulas, por lo que se requieren

estudios de validación adicionales.

Nuestra hipótesis en el estudio número 1 es que existe suficiente evidencia

científica sobre la no inferioridad de la MV con respecto a la MC para el diagnóstico

primario en numerosas áreas, pero que es preciso identificar las áreas de la patología

en las que los estudios de validación sean escasos o inexistentes, en los cuales resulte

necesaria la realización de estudios de validación adicionales antes de su

implementación en el diagnóstico primario. En este estudio inicial se detectaron

algunas áreas en las que la evidencia era escasa o nula. Por ello hemos llevado a cabo

los estudios número 2 y número 3 con el objeto de validar respectivamente la MV

para el diagnóstico primario de biopsias ginecológicas y biopsias hepáticas con aguja.

En ambos casos la principal hipótesis es que el diagnóstico histológico rutinario de

estas biopsias con MV no es inferior al realizado con MC. También planteamos como

hipótesis adicional la existencia de una curva de aprendizaje al inicio del uso de la MV,

pero que rápidamente tanto el diagnóstico como el tiempo empleado por el/la

patólogo/a se equipara con la MC.

Por otro lado, son más numerosas las evidencias de las ventajas que la MV

aporta a la docencia tanto de pre como de post-grado. Sin embargo, resulta también

fundamental la validación de la MV en la docencia. La comparación entre los


[42]

resultados de los exámenes tras la realización de las prácticas usando ambas

herramientas es uno de los parámetros más útiles a la hora de asegurar que la MV y la

MC son equiparables. Otro parámetro a tener en cuenta es la valoración por parte del

alumnado y de los docentes; lo que también puede ayudar a conocer cuáles son las

necesidades y la forma de llevar a cabo el proceso educativo en la actualidad. Sobre

estas premisas se realizó el estudio número 4 con la hipótesis de que, al igual que en el

diagnóstico primario, los estudios sobre el uso de la MV en la docencia permiten

confirmar su utilidad para la docencia, tanto en alumnos de pre-grado como en la

formación post-grado. Las conclusiones de este estudio encauzaron la implementación

de la MV en la enseñanza de la Anatomía Patológica en los alumnos de pregrado de la

Facultad de Medicina de la Universidad de Barcelona. El estudio número 5 tiene como

hipótesis principal que los resultados del uso de la MV en docencia no son

equiparables a los alcanzados con la MC y que la valoración por parte de los alumnos

de la MV es positiva.


[43]

III. Objetivos


[44]


[45]

El objetivo general de la presente tesis es estudiar el uso de la MV en el diagnóstico

rutinario de biopsias en un Servicio de Anatomía Patológica, así como, su uso en la

docencia; con el fin de determinar la utilidad de esta tecnología en ambos ámbitos,

que permita planificar de forma adecuada su implementación en ambos campos.

De forma particular se han planteado los siguientes objetivos específicos:

1. Evaluar en los artículos publicados que la MV y la MC presentan resultados

equiparables en el diagnóstico primario de las distintas subespecialidades de la

Anatomía Patológica (estudio 1)

2. Determinar si existen suficientes estudios de validación que engloben la

totalidad de las diferentes subespecialidades de la Anatomía Patológica, en

especial aquellas con características diferentes (estudio 1)

3. Evaluar la concordancia inter-observador entre los diagnósticos realizados con

MV y con MC en biopsias ginecológicas de rutina, entre dos especialistas con

experiencia en el área (estudio 2)

4. Determinar si existe una curva de aprendizaje en el uso de la MV y sus

características (estudio 2)

5. Establecer cuál es el aumento de escaneo con una mejor relación coste-beneficio

para realizar un diagnóstico adecuado en biopsias ginecológicas (estudio 2)

6. Evaluar la concordancia inter e intra-observador entre los diagnósticos realizados

con MV y con MC en biopsias hepáticas con aguja, tanto provenientes de hígados

nativos como de trasplante (estudio 3)

7. Establecer la estrategia de escaneo más adecuada para alcanzar un diagnóstico

correcto en estas biopsias (estudio 3)

8. Determinar si existe suficiente evidencia sobre la adecuación de la MV en la

docencia de Anatomía Patológica, tanto de pre como de postgrado (estudio 4)


[46]

9. Determinar si el paso de la MC a la MV en la asignatura de Anatomía Patológica

tiene impacto en los resultados de los estudiantes de Medicina de la Universidad

de Barcelona (estudio 5)

10. Analizar las impresiones de los estudiantes sobre el uso de la MV y valorar de

forma objetiva cómo influye esta herramienta en el proceso de aprendizaje

(estudio 5)


[47]

IV. Trabajos

realizados, métodos

y resultados


[48]


[49]

La descripción de las muestras, así como la metodología utilizada en los trabajos

realizados, se encuentran detalladamente descritas en las secciones de “Material y Métodos”

de cada uno de los artículos que constituyen el cuerpo doctrinal de la presente Tesis Doctoral.

Dichos artículos se incluyen a continuación tal y como han sido publicados en la

literatura científica.


[50]


[51]

Estudio número 1

“Validation of Whole-slide Imaging for Histopathological

Diagnosis: Current State”

Adela Saco, José Ramirez, Natalia Rakislova, Aurea Mira, Jaume Ordi

Pathobiology 2016; 83: 89 – 98




[52]

E-Mail [email protected]

Original Paper

Pathobiology 2016;83:89–98 DOI: 10.1159/000442823

Validation of Whole-Slide Imaging for Histolopathogical Diagnosis: Current State

Adela Saco a Jose Ramírez a Natalia Rakislova a Aurea Mira a Jaume Ordi a, b

a Department of Pathology, Hospital Clínic, University of Barcelona School of Medicine, and b ISGlobal, Barcelona Center for International Health Research (CRESIB), Barcelona , Spain

annotations and measurements. WSI can be used from any device and anywhere, thereby providing great opportuni-ties for teleconsultation. New technologies such as the rec-ognition of histopathology patterns using image analysis may facilitate diagnosis and improve the reproducibility among pathologists in the future. © 2016 S. Karger AG, Basel

Introduction and Historical Perspective

For more than a century, conventional light micros-copy (CLM) has been the basic tool for tissue evaluation and has played a pivotal role in pathological diagnosis. Until the incorporation of nonmorphological molecular technologies into routine practice in recent years, the standard of diagnosis for pathologists was morphology and especially CLM-evaluated morphological criteria. In-deed, the evaluation of most specimens submitted to pa-thology laboratories today still relies on the interpretation of images by CLM, complemented by gross examination and a number of ancillary molecular techniques, mostof which [histochemistry and immunohistochemistry (IHC)] are also evaluated with CLM. Asking experts or other colleagues for diagnostic opinions required sending

Key Words

Primary diagnosis · Routine diagnosis · Validation · Virtual microscopy · Whole-slide images

Abstract

Rapid advances in informatics and technological improve-ments have led to the development of high-throughput whole-slide imaging (WSI) scanners able to produce high-quality digital images, which allow achieving a correct diag-nosis of the biopsies using virtual viewers. This technology is currently prepared to be introduced in the departments of pathology for routine diagnosis. The aim of this review is to analyze the current evidence regarding the use of WSI in pri-mary or routine diagnosis in the different subspecialties of pathology. An increasing number of studies have shown al-most perfect inter- and intraobserver agreement between the diagnoses obtained with WSI and the classical diagnoses based on conventional light microscopy. The only exception seems to be cytology, which still requires some technologi-cal development. Although validation studies are needed in some areas of pathology, growing evidence indicates that WSI is a reliable tool for routine diagnosis. Pathologists have a positive perception of the ergonomics of the workstations, the low magnification of WSI and the possibility of making

Published online: April 26, 2016

Jaume Ordi Department of Pathology, Hospital Clínic, University of Barcelona C/Villarroel 170 ES–08036 Barcelona (Spain) E-Mail jordi @ clinic.ub.es

© 2016 S. Karger AG, Basel1015–2008/16/0833–0089$39.50/0

www.karger.com/pat

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M

Saco/Ramírez/Rakislova/Mira/Ordi

Pathobiology 2016;83:89–98DOI: 10.1159/000442823

90

glass slides or paraffin blocks for examination by CLM. Teaching pathology to undergraduates and residents, and continuing medical education for certified pathologists also depended on the use of CLM.

This scenario slowly started to change a few decades ago [1–3] . Static digital images allowed teaching and, to a certain degree, teleconsultation, but limitations in image quality and, particularly, the inability to navigate and use different optical objectives made the substitution of CLM unfeasible [4] . Dynamic real-time telepathology systems with video cameras integrated into CLM were used for intraoperative frozen biopsies, because they allowed an image to be sent to an expert located remotely. This ca-pacity was extraordinarily useful for small hospitals, as it provided a quick diagnostic approach for difficult cases [5–12] . However, the relatively poor image quality and the impossibility to remotely conduct navigation through a slide made the system inadequate for routine diagnosis.

Rapid advances in informatics as well as technological improvements led to the development of scanners able to create digital reproductions from whole glass slides, which appeared one decade ago [1, 2] . These scanners are the basis of virtual microscopy or whole-slide imaging (WSI), which allows navigation across the virtual slide and visualization at different magnifications, allowing the computer to be used as a CLM. However, the image qual-

ity of the initial scanners was limited, and the costs of implementation of the technology, including the scanner, monitors and suitable computers, were very high, thereby restricting the use of WSI to certain areas, such as teach-ing and teleconsultation, and excluding routine diagnosis [10, 13–17] .

Currently, a number of high-throughput scanners able to produce high-quality images are available on the mar-ket. These scanners allow correct diagnosis of the biopsies using virtual viewers. The cost of implementation of WSI has significantly decreased, and the speed of visualization has notably increased [17–22] . Constant improvements in this technology have led to an important expansion in the use of WSI in routine diagnosis in recent years. The aim of this review is to evaluate the current evidence on the validation of WSI in routine diagnosis.

Advantages and Challenges of WSI for Routine

Diagnosis

Routine histopathological diagnosis can benefit from the multiple advantages of WSI. WSI workstations are more ergonomic ( fig. 1 ). WSI has a much larger field of vision than CLM and allows a wider range of magnifica-tions, thus providing easier navigation. In particular, WSI enables to study very low magnifications (<×100), which is very useful in the evaluation of surgical specimens. The computer tools allow making annotations and measure-ments. WSI viewers can simultaneously show and syn-chronously move several slides of a case, which is particu-larly helpful in the evaluation of IHC-stained slides ( fig. 2 ). Indeed, studies evaluating the opinion of patholo-gists have revealed a positive perception of image quality and stressed the utility of the measurement and annota-tion tools, as well as the ergonomics and usability of the viewer [22] . WSI can be used from any device and any-where, thereby providing great opportunities for telecon-sultation and remote work. Portability is certainly one of the major advantages of WSI, and this will probably be further improved in the near future when the current viewers are fully adapted to portable devices, such as tab-lets and smartphones [23–25] . Moreover, the need for standardization in the diagnosis and evaluation of IHC biomarkers predicting the outcome of specific therapies will probably boost the implementation of WSI.

Finally, WSIs allow for automatic quantification of IHC slides. These diagnostic algorithms facilitate quanti-fication of IHC positivity resulting in a more objective evaluation, which is extremely useful in the evaluation of

Fig. 1. WSI workstations for primary diagnosis typically include two screens, one displaying the WSI viewer and the other the labo-ratory information system and the clinical records or other clinical or imaging information. This physical structure has shown to be highly ergonomic. Additional advantages of WSI viewers are a much larger field of vision than CLM and the possibility of using a very low magnification.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M

Validation of WSI in Primary Diagnosis Pathobiology 2016;83:89–98DOI: 10.1159/000442823

91

some biological markers. Algorithms for the evaluation of IHC stains are variably used depending on the subspe-cialties and are particularly useful in cases of breast cancer [26–29] .

In contrast with these positive opinions, many pathol-ogists still prefer using CLM. The most criticized feature of WSI is the speed in uploading the image. Indeed, most pathologists feel that more time is required to make a di-agnosis with WSI. However, some studies have shown that although diagnosis with WSI is initially more time-consuming, this time quickly decreases as pathologists become familiar with the use of the WSI viewer [30–35] . Thus, there is a learning curve in the use of WSI and the time required for making a diagnosis, and a recent study conducted at our institution confirmed that the diagnos-tic performance improved with practice [36] . Another limitation of WSI is the relatively high costs of the equip-ment. The basic needs for a WSI system, which is ade-quate for routine diagnosis, include not only high-throughput scanners but also high-resolution monitors [37, 38] . This is a common concern since, despite the reduction in the price of the equipment in the last few years, it still represents a considerably high investment, which has a relatively low added value for many patholo-gists as the basic functions of WSI are already being con-fidently achieved with the old CLM. Finally, WSI re-quires a significant investment in high-capacity servers; the files generated by WSI scanners are huge, with sizes frequently over 2 GB per slide. Thus, strategies to reduce the size of the files, such as scanning at relatively low magnification (×200 instead of ×400 or ×600) are fre-quently used [37] .

The Need for Validation Studies

The number of studies aimed at validating WSI in pri-mary or routine diagnosis is rapidly increasing. However, whereas relatively abundant information is available in some areas, validation studies are very scant in several sub-specialties and completely absent in others. Some valida-tion studies include biopsies from several subspecialties in-stead of analyzing biopsies with similar characteristics [33, 39–43] . This relative absence of validation studies has led to reluctance in the implementation of WSI in routine clin-ical practice. Nevertheless, the number of centers imple-menting this technology is increasing due to the positive experiences reported in many departments [41, 42, 44, 45] .

Below, we review the current evidence on the valida-tion of WSI versus CLM in the different subspecialties of pathology.

Breast Pathology

WSI has been validated in the diagnosis of breast pa-thology in a number of studies conducted by different groups. Most of these studies analyzed a relatively small number of routine biopsies (between 100 and 150), in-cluding either only needle biopsies or both needle and surgical specimens [32, 46, 47] . Although scanning at ×400 was recommended in one of the studies [32] , in two of the studies a scanning magnification of ×200 was con-sidered as sufficient [46, 47] .

The intra- and interobserver agreement between CLM and WSI is excellent in all the studies, with values ranging

Fig. 2. WSI viewers may simultaneously show and synchronously move several slides of a case, which is particularly helpful in the evaluation of IHC-stained slides.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M



92

between 90 and 99%. Most of the discrepancies detected did not have clinical repercussion. Interestingly, in two of the reports, the WSI diagnosis was more frequently con-sidered as correct compared to the diagnosis performed with CLM [32, 46] . A study specifically dealing with the distinction between hyperplasia and cancer reported in-terobserver concordance in the diagnosis of 90.2%. Major discrepancies appeared in 2.3% of the cases, which, in most cases, were solved with IHC stains [48] .

A major advantage of digitization in breast pathology is the possibility to use image analysis to improve the ac-curacy and reproducibility of HER-2, estrogen and pro-gesterone receptors, and Ki67 scoring, which have a cru-cial role in the planning of treatment strategies [27–29, 49] . Moreover, the evaluation may be improved with the use of automatic quantification algorithms ( fig. 3 ).

Cytopathology

The use of WSI in cytopathology has shown someadvantages in second opinions, quality assurance, slide archiving, proficiency testing and education. However, a number of significant weaknesses of the current WSI scanners, such as the difficulties in focusing at different z-axes, are a major limitation for the introduction of this technology in routine diagnosis [50, 51] . Improvements in informatics may allow multiplane focusing using the z-axis, but they still need to be validated [21, 52, 53] .

Indeed, the current evidence of validation in cytology is almost limited to real-time dynamic digital microscopy using a video camera connected to the optical microscope

and not to WSI. The intraobserver agreement of this ap-proach with the final diagnosis is high (92%) [54] , and, in some studies, it is better than with CLM [53–55] . One study evaluating 192 liquid-based cervical cytology slides showed good intraobserver concordance (89–97%), but the interobserver concordance was better for CLM than for WSI (94 vs. 82%) [52] .

Dermatopathology

Only two studies have focused on the validation of skin biopsies evaluating routine specimens. Although both studies included a small number of cases (100 and 79, respectively), the intraobserver agreement was high (94% for WSI and 96% for CLM, respectively) [30, 56] . A study limited to tumor and tumor-like skin lesions showed agreement in the diagnosis by WSI and CLM, with a κ value of 0.93 for both methods [57] . Another study evaluated inflammatory and melanocytic lesions, with good agreement between CLM and WSI (only 1 dis-cordant diagnosis in the inflammatory biopsies and 100% concordance in the melanocytic specimens), but the number of patients included was very limited (24 cas-es). In this study, it was concluded that in most cases scanning at ×200 is sufficient to achieve a correct diag-nosis [56] .

Interestingly, WSI has shown to be suitable for tele-consultation in skin biopsies and may reduce the time of response in expert diagnosis from 5–10 days to a few hours or even minutes [57] .

Fig. 3. A major advantage of digitization in breast pathology is the possibility to use image analysis in improving the accuracy and reliability of HER-2, estrogen and pro-gesterone receptors and Ki67 scoring, which have a crucial role in the planning of treatment strategies.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M


93

Gastrointestinal Pathology

A few studies have shown that the diagnosis of gastro-intestinal biopsies using WSI or CLM provides compa-rable results [58, 59] . Two reports analyzed consecutive routine biopsies, but one was limited to gastric and co-lonic biopsies [59] . The intraobserver concordance be-tween WSI and CLM was 95% in both studies, and scan-ning at ×200 was considered as adequate. One study com-pared WSI and CLM in the evaluation of polyps in surgical specimens. Although the intra- and interobserv-er agreement was excellent for both methods in terms of diagnosis, WSI facilitated the quantification of the polyps due to the very low magnification that allows a panoram-ic view of the complete sample [60] . A study focused on Barrett’s dysplasia and neoplasia showed good diagnostic agreement between WSI and CLM, but the consensus neoplasia score was lower using WSI and the time spent in making the diagnosis was longer. These results were probably due to the lack of confidence and experience in the manipulation of the WSI viewer and seemed to im-prove with familiarity and practice [34] .

Genitourinary Pathology

Prostatic biopsies, particularly needle biopsies, are good candidates for digitization for a number of reasons: the tissue size is small and the images generated are light-er; multiple measurements are frequently required and informatics tools can facilitate these, and WSI allows a global view to more easily establish the Gleason score ( fig. 4 ) [61] . An additional advantage of WSI is the pos-

sibility to synchronize hematoxylin-eosin stains and p63 IHC in the same screen, thereby allowing the comparison of the two images and facilitating the diagnostic and teaching process [62] .

Thus, the current evidence on the validation of WSI in the diagnosis of prostatic biopsies is more extensive than in other areas. A number of studies (from 50 to over 800 cases) have been focused on the evaluation of the Gleason score in needle biopsies. Scanning at ×200 was considered sufficient to make the diagnosis. The κ values for diagno-sis ranged between 0.586 and 0.813 [63–65] , and one of the reports included only difficult biopsies with a border-line Gleason score. Concordance between WSI and CLM seems to be higher for primary (κ values 0.65–0.96) than for secondary Gleason scores (κ values 0.53–0.75), and most discordances have no impact on patient manage-ment [66] . Tumor size is better evaluated with WSI, and other parameters such as perineural invasion show simi-lar values with WSI and CLM [66] .

Two additional studies focused on genitourinary biop-sies included a mixture of prostatic and urinary tract bi-opsies and showed good intraobserver concordance (90 and 87.5%, respectively) [67, 68] .

Gynecological Pathology

Studies on the validation of WSI in gynecological bi-opsies are scant. Only one study conducted at our institu-tion analyzed interobserver agreement in 452 routine gy-necological specimens showing a κ index of 0.914 (almost perfect concordance). Interestingly, the agreement be-tween WSI and CLM increased in this study in parallel

Fig. 4. Prostatic biopsies often require mul-tiple measurements. The tools of WSI viewers make these measurements easy. WSI allows a global view to more adequate-ly establish the Gleason score.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M



94

with time, suggesting that there is a learning curve in the use of WSI and that experience in the use of WSI viewers improves the results obtained. Major discrepancies were found in only 2% of the cases, and none was related to poor image quality. Most discrepancies in this study were observed in biopsies of premalignant lesions of the uter-ine cervix, an area which has shown high inter- and in-traobserver variability rates using CLM [36] . The magni-fication used in the study was ×200, and higher magnifi-cation did not seem to be required.

A second study described the usefulness of WSI in the evaluation of 52 frozen ovarian sections showing 96% in-terobserver agreement. Interestingly, in this study, no clinical information was provided to the pathologists, and the time spent per case was 3–5 min [7] .

Head and Neck Pathology

To date, no validation study including the complete spectrum of samples of this subspecialty is available. Only one study on premalignant laryngeal lesions has been published. This study concluded that WSI is a valid alter-native to CLM. Although the correlation with the final diagnosis was slightly lower with WSI than with CLM, the differences were not statistically significant [69] .

Neuropathology

Validation studies of neuropathology are limited to in-traoperative biopsies and smears [5, 8, 70] . Agreement between the diagnosis with WSI and the final diagnosis using CLM is very good, even with low scan magnifica-tion (×100). The studies conclude that ×200 magnifica-tion is sufficient to obtain a diagnosis. In one study, the diagnosis achieved with WSI was concordant with CLM in 29 of the 30 cases evaluated, and the discordant diag-nosis did not lead to changes in the management of the patient [8] . A second study included 126 frozen sections that were evaluated by four different pathologists. The diagnosis was discordant with the final report in only 8 cases. In this study, the diagnosis of frozen sections scanned and diagnosed using WSI was compared with the final diagnosis obtained in formalin-fixed, paraffin-embedded tissue [70] .

Algorithms are currently being developed to identify the hot spots in Ki67-stained sections to automatically quantify the proliferative activity in tumors of the central nervous system [71, 72] .

Pediatric Pathology

Two studies have validated the use of WSI in pediatric pathology. One included 80 routine biopsies of patients under 18 years of age and 20 placentas. The intraobserver concordance between the diagnoses with WSI and CLM was 90% in pediatric biopsies and 93% in placental speci-mens. Major discrepancies were observed in only 2% of the cases. A scanning magnification of ×200 generated an image quality allowing correct diagnosis except for the identification of nucleated red blood cells, which is very difficult even when the slides are scanned at a magnifica-tion of ×400 [73] .

The second study evaluated WSI in 60 cases selected to include the whole spectrum of the diagnostic complexity of pediatric biopsies. The surgical specimens were digitized at ×200 magnification, whereas small biopsies and cyto-logical samples were digitized at ×400. The intraobserver agreement was almost perfect with only 1 discordant case. The scanning process of two cytological smears was unsat-isfactory because the material was very scanty [74] .

Pulmonary Pathology

One study validating WSI in the diagnosis of intraop-erative pulmonary specimens included a variety of sam-ples, with 114 frozen sections from tumors, lymph nodes and bronchial margins, 174 fine-needle aspiration slides, 3 exfoliative smears and 13 small biopsies. This study evaluated both a dynamic real-time telepathology system and WSI, and found very good agreement, which was bet-ter for WSI than for the real-time telepathology system (100% in consultation and frozen biopsies) [75] . A second study analyzed the use of WSI in 20 tumor biopsies and surgical specimens sent for consultation. Complete in-terobserver agreement was achieved in 85% of the cases, even at a scanning magnification of ×100 [75] .

Renal Pathology

Validation studies of WSI in the diagnosis of renal pa-thology biopsies are scarce and include few cases. A re-port including 50 routine renal biopsies showed complete intraobserver agreement in 84% of the cases. Five major discrepancies (with clinical repercussion for the patient) were found and in 2 cases the correct diagnosis was made with WSI. In this study, renal transplant biopsies showed significantly more discrepancies at a magnification of

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M


95

×200 [67] . Another study (using a magnification of ×400) reported good agreement in renal transplant biopsies, but the time spent for obtaining the diagnosis was longer with WSI than with CLM [76] . Finally, one study evaluated the concordance between 96 pathologists in the diagnosis of 12 renal biopsies using WSI and CLM and found no sig-nificant differences between both methods [77] .

Intraoperative Diagnosis of Frozen Sections

A number of studies have evaluated dynamic real-time telepathology in intraoperative sections, showing a good correlation with CLM diagnosis. They emphasize the learning curve in the use of the WSI technology, which typically involves a longer diagnostic time window at the beginning but rapid improvement with practice [78] . A validation study using WSI in frozen intraoperative sec-tions from different anatomical sites has shown almost perfect agreement with a κ index of 0.85. The mean time spent on diagnosis was 2 min 50 s per case. The quality of the image was considered excellent in 98% of cases [9] . Studies using WSI in frozen intraoperative sections from specific specialties have been discussed above. In another study evaluating 67 consecutive frozen intraoperative sections, diagnosis was obtained by viewing the virtual slides in a portable device (iPad tablet). The slides were scanned at ×200, and all cases were shown together with the clinical information. The concordance between the diagnoses achieved with WSI and CLM was good with a κ value of 0.85. The mean time to achieving a diagnosis using WSI was 2 min and 46 s [79] .

Surgical Pathology

A number of studies have evaluated a variety of differ-ent specimens from the routine practice in the depart-ment of pathology, including between 25 and 607 samples [22, 39–41, 80, 81] . Inter- and intraobserver agreement between WSI and CLM varied from 75 to 97.7% depend-ing on the study. Most studies concluded that a magnifi-cation of ×200 provides images with adequate quality for diagnosis [22, 39–41, 80, 81] . The interobserver agree-ment between WSI and CLM was 95%, and all discrepan-cies were minor. However, although the general opinion of the pathologists was positive, some felt that the WSI system was slower than CLM, and most of the patholo-gists interviewed were reluctant to completely move from CLM to WSI in routine diagnosis [33] . One study sug-

gested that the interobserver agreement was better for neoplastic than for nonneoplastic diseases [16] . It has been suggested that a scanning magnification of ×200 may not be sufficient to allow correct diagnosis in inflam-matory lesions [82] .

Finally, two studies included only consultation biop-sies of different organs. The interobserver agreement be-tween WSI and CLM diagnoses in these studies was great-er than 91%, and most of the discrepancies were due to the intrinsic difficulty in diagnosing some cases [15, 16, 82] .

Current Recommendations for WSI Validation

Validation of WSI at each institution has been recom-mended before its implementation in routine diagnosis. Several professional associations have developed guide-lines with recommendations for the introduction of WSI in routine diagnosis in a department of pathology. The first guidelines were developed by the College of Ameri-can Pathologists and the American Telemedicine Asso-ciation, and include some recommendations and sugges-tions to be followed before using WSI for diagnosis [2, 10, 17] . It is recommended to include a variety of different biopsies representative of the complexity of the surgical specimens usually analyzed in the center. The guidelines state that it is not necessary to validate each subspecialty because the results from one specialty can be extrapolated to others with similar features. Each specific type of spec-imen with significant differences requires an internal val-idation. The guidelines recommend measuring intraob-server agreement between WSI and CLM, using a ‘wash-out period’ of 2 weeks. Finally, it is recommended that a pathologist with experience in WSI should be involved in the process of validation.

Conclusions

In conclusion, independently of the subspecialty, all the validation studies published show a very good correla-tion between diagnoses achieved with WSI and CLM. Thus, WSI seems to be an adequate tool for histological diagnosis in routine practice and has several advantages over CLM. However, although good evidence demon-strating that WSI can be reliably used for routine diagno-sis has been provided for several specialties, there are a number of areas of pathology, such as hematopathology and liver, endocrine, bone and soft-tissue pathology, for which no study has yet been published. Although some of

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M



96

these areas may be considered similar to others already validated, specific validation studies are needed in areas with many differences such as liver biopsies or hematopa-thology. These validations are necessary before the use of WSI can be extended to these subspecialties with the aim of going fully digital in pathological services in the future.

Notwithstanding, as with many other new tools, the use of WSI has a learning curve, and the time spent on the diagnosis and, to a lesser extent, inter- and intraobserver agreement may be suboptimal in the initial phases of its use. Cytology seems to be an exception; the application of WSI in this area is more controversial due to the impos-sibility of focusing on different planes.

However, the introduction of WSI in routine diagnosis faces some difficulties, mainly related to the reluctance of pathologists to abandon CLM and to the costs associated with the acquisition of the equipment and the storage of the images generated. New technologies that allow creat-ing 3D reconstruction from 2D biopsies may help to im-prove the understanding of the growth patterns and the spatial arrangement of diseased cells [21, 83] . Another area that will markedly expand in the next few years is that of histopathology pattern recognition using image analy-sis, which can facilitate the diagnostic tasks and improve the reproducibility among pathologists in many subspe-cialties [65, 84–90] .

References

1 Al-Janabi S, Huisman A, van Diest PJ: Digital pathology: current status and future perspec-tives. Histopathology 2012; 61: 1–9.

2 Pantanowitz L, Sinard JH, Henricks WH, Fa-theree LA, Carter AB, Contis L, Beckwith BA, Evans AJ, Lal A, Parwani AV: Validating whole slide imaging for diagnostic purposes in pathology: guideline from the College of American Pathologists Pathology and Labo-ratory Quality Center. Arch Pathol Lab Med 2013; 137: 1710–1722.

3 Weinstein RS: Prospects for telepathology. Hum Pathol 1986; 17: 433–434.

4 Cross SS, Dennis T, Start RD: Telepathology: current status and future prospects in diag-nostic histopathology. Histopathology 2002; 41: 91–109.

5 Evans AJ, Kiehl TR, Croul S: Frequently asked questions concerning the use of whole-slide imaging telepathology for neuropathology frozen sections. Semin Diagn Pathol 2010; 27: 160–166.

6 Evans AJ, Chetty R, Clarke BA, Croul S, Gha-zarian DM, Kiehl TR, Ordonez BP, Ilaalagan S, Asa SL: Primary frozen section diagnosis by robotic microscopy and virtual slide telepa-thology: the University Health Network expe-rience. Semin Diagn Pathol 2009; 26: 165–176.

7 Fallon MA, Wilbur DC, Prasad M: Ovarian frozen section diagnosis: use of whole-slide imaging shows excellent correlation between virtual slide and original interpretations in a large series of cases. Arch Pathol Lab Med 2010; 134: 1020–1023.

8 Gould PV, Saikali S: A comparison of digi-tized frozen section and smear preparations for intraoperative neurotelepathology. Anal Cell Pathol (Amst) 2012; 35: 85–91.

9 Kaplan KJ, Burgess JR, Sandberg GD, Myers CP, Bigott TR, Greenspan RB: Use of robotic telepathology for frozen-section diagnosis: a retrospective trial of a telepathology system for intraoperative consultation. Mod Pathol 2002; 15: 1197–1204.

10 Pantanowitz L, Dickinson K, Evans AJ, Has-sell LA, Henricks WH, Lennerz JK, Lowe A, Parwani AV, Riben M, Smith CD, Tuthill JM, Weinstein RS, Wilbur DC, Krupinski EA, Bernard J: American Telemedicine Associa-tion clinical guidelines for telepathology. J Pathol Inform 2014; 5: 39.

11 Slodkowska J, Pankowski J, Siemiatkowska K, Chyczewski L: Use of the virtual slide and the dynamic real-time telepathology systems for a consultation and the frozen section intra-operative diagnosis in thoracic/pulmonary pathology. Folia Histochem Cytobiol 2009; 47: 679–684.

12 Wilbur DC: Digital pathology: get on board – the train is leaving the station. Cancer Cyto-pathol 2014; 122: 791–795.

13 Ayad E: Virtual telepathology in Egypt, appli-cations of WSI in Cairo University. Diagn Pathol 2011; 6(suppl 1):S1.

14 Romero LG, Cable W, Lesniak A, Tseytlin E, McHugh J, Parwani A, Pantanowitz L: Digital pathology consultations – a new era in digital imaging, challenges and practical applica-tions. J Digit Imaging 2013; 26: 668–677.

15 Wienert S, Beil M, Saeger K, Hufnagl P, Schrader T: Integration and acceleration of virtual microscopy as the key to successful im-plementation into the routine diagnostic pro-cess. Diagn Pathol 2009; 4: 3.

16 Wilbur DC, Madi K, Colvin RB, Duncan LM, Faquin WC, Ferry JA, Frosch MP, Houser SL, Kradin RL, Lauwers GY, Louis DN, Mark EJ, Mino-Kenudson M, Misdraji J, Nielsen GP, Pitman MB, Rosenberg AE, Smith RN, Sohani AR, Stone JR, Tambouret RH, Wu CL, Young RH, Zembowicz A, Klietmann W: Whole-slide imaging digital pathology as a platform for teleconsultation: a pilot study using paired subspecialist correlations. Arch Pathol Lab Med 2009; 133: 1949–1953.

17 Bernard C, Chandrakanth SA, Cornell IS, Dalton J, Evans A, Garcia BM, Godin C, God-lewski M, Jansen GH, Kabani A, Louahlia S, Manning L, Maung R, Moore L, Philley J, Slat-nik J, Srigley J, Thibault A, Picard DD, Cra-cower H, Tetu B: Guidelines from the Cana-dian Association of Pathologists for establish-ing a telepathology service for anatomic pathology using whole-slide imaging. J Pathol Inform 2014; 5: 15.

18 Hedvat CV: Digital microscopy: past, present, and future. Arch Pathol Lab Med 2010; 134: 1666–1670.

19 Ho J, Ahlers SM, Stratman C, Aridor O, Pan-tanowitz L, Fine JL, Kuzmishin JA, Montalto MC, Parwani AV: Can digital pathology result in cost savings? A financial projection for dig-ital pathology implementation at a large inte-grated health care organization. J Pathol In-form 2014; 5: 33.

20 Isaacs M, Lennerz JK, Yates S, Clermont W, Rossi J, Pfeifer JD: Implementation of whole slide imaging in surgical pathology: a value added approach. J Pathol Inform 2011; 2: 39.

21 Pantanowitz L: Digital images and the future of digital pathology. J Pathol Inform 2010; 1: 15.

22 Thorstenson S, Molin J, Lundstrom C: Imple-mentation of large-scale routine diagnostics using whole slide imaging in Sweden: digital pathology experiences 2006–2013. J Pathol Inform 2014; 5: 14.

23 Hartman DJ, Parwani AV, Cable B, Cucoranu IC, McHugh JS, Kolowitz BJ, Yousem SA, Pa-lat V, Reden AV, Sloka S, Lauro GR, Ahmed I, Pantanowitz L: Pocket pathologist: a mobile application for rapid diagnostic surgical pa-thology consultation. J Pathol Inform 2014; 5: 10.

24 Roy S, Pantanowitz L, Amin M, Seethala RR, Ishtiaque A, Yousem SA, Parwani AV, Cuco-ranu I, Hartman DJ: Smartphone adapters for digital photomicrography. J Pathol Inform 2014; 5: 24.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M


97

25 Speiser JJ, Hughes I, Mehta V, Wojcik EM, Hutchens KA: Mobile teledermatopathology: using a tablet PC as a novel and cost-efficient method to remotely diagnose dermatopathol-ogy cases. Am J Dermatopathol 2014; 36: 54–57.

26 Gavrielides MA, Conway C, O’Flaherty N, Gallas BD, Hewitt SM: Observer performance in the use of digital and optical microscopy for the interpretation of tissue-based biomarkers. Anal Cell Pathol (Amst) 2014; 2014: 157308.

27 Nassar A, Cohen C, Agersborg SS, Zhou W, Lynch KA, Barker EA, Vanderbilt BL, Thompson J, Heyman ER, Olson A, Lange H, Siddiqui MT: A multisite performance study comparing the reading of immunohisto-chemical slides on a computer monitor with conventional manual microscopy for estro-gen and progesterone receptor analysis. Am J Clin Pathol 2011; 135: 461–467.

28 Micsik T, Kiszler G, Szabo D, Krecsak L, He-gedus C, Tibor K, Molnar B: Computer aided semi-automated evaluation of HER2 immu-nodetection – a robust solution for support-ing the accuracy of anti HER2 therapy. Pathol Oncol Res 2015; 21: 1005–1011.

29 Krenacs T, Zsakovics I, Diczhazi C, Ficsor L, Varga VS, Molnar B: The potential of digital microscopy in breast pathology. Pathol Oncol Res 2009; 15: 55–58.

30 Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJ, Ten Kate FJ, van Dijk MR, van Diest PJ: Whole slide images for primary di-agnostics in dermatopathology: a feasibility study. J Clin Pathol 2012; 65: 152–158.

31 Randell R, Ruddle RA, Mello-Thoms C, Thomas RG, Quirke P, Treanor D: Virtual re-ality microscope versus conventional micro-scope regarding time to diagnosis: an experi-mental study. Histopathology 2013; 62: 351–358.

32 Krishnamurthy S, Mathews K, McClure S, Murray M, Gilcrease M, Albarracin C, Spino-sa J, Chang B, Ho J, Holt J, Cohen A, Giri D, Garg K, Bassett RL Jr, Liang K: Multi-institu-tional comparison of whole slide digital imag-ing and optical microscopy for interpretation of hematoxylin-eosin-stained breast tissue sections. Arch Pathol Lab Med 2013; 137: 1733–1739.

33 Houghton JP, Ervine AJ, Kenny SL, Kelly PJ, Napier SS, McCluggage WG, Walsh MY, Hamilton PW: Concordance between digital pathology and light microscopy in general surgical pathology: a pilot study of 100 cases. J Clin Pathol 2014; 67: 1052–1055.

34 Sanders DS, Grabsch H, Harrison R, Bateman A, Going J, Goldin R, Mapstone N, Novelli M, Walker MM, Jankowski J: Comparing virtual with conventional microscopy for the con-sensus diagnosis of Barrett’s neoplasia in the AspECT Barrett’s chemoprevention trial pa-thology audit. Histopathology 2012; 61: 795–800.

35 Randell R, Ruddle RA, Thomas RG, Mello-Thoms C, Treanor D: Diagnosis of major can-cer resection specimens with virtual slides: impact of a novel digital pathology worksta-tion. Hum Pathol 2014; 45: 2101–2106.

36 Ordi J, Castillo P, Saco A, Del Pino M, Ordi O, Rodriguez-Carunchio L, Ramirez J: Vali-dation of whole slide imaging in the primary diagnosis of gynaecological pathology in a university hospital. J Clin Pathol 2015; 68: 33–39.

37 Cornish TC, Swapp RE, Kaplan KJ: Whole-slide imaging: routine pathologic diagnosis. Adv Anat Pathol 2012; 19: 152–159.

38 Pantanowitz L, Valenstein PN, Evans AJ,Kaplan KJ, Pfeifer JD, Wilbur DC, Collins LC, Colgan TJ: Review of the current state of whole slide imaging in pathology. J Pathol In-form 2011; 2: 36.

39 Gilbertson JR, Ho J, Anthony L, Jukic DM, Yagi Y, Parwani AV: Primary histologic diag-nosis using automated whole slide imaging: a validation study. BMC Clin Pathol 2006; 6: 4.

40 Jukic DM, Drogowski LM, Martina J, Parwani AV: Clinical examination and validation of primary diagnosis in anatomic pathology us-ing whole slide digital images. Arch Pathol Lab Med 2011; 135: 372–378.

41 Li X, Liu J, Xu H, Gong E, McNutt MA, Li F, Anderson VM, Gu J: A feasibility study of vir-tual slides in surgical pathology in China. Hum Pathol 2007; 38: 1842–1848.

42 Stathonikos N, Veta M, Huisman A, van Diest PJ: Going fully digital: perspective of a Dutch academic pathology lab. J Pathol Inform 2013; 4: 15.

43 Bauer TW, Schoenfield L, Slaw RJ, Yerian L, Sun Z, Henricks WH: Validation of whole slide imaging for primary diagnosis in surgi-cal pathology. Arch Pathol Lab Med 2013; 137: 518–524.

44 Al-Janabi S, Huisman A, Nap M, Clarijs R, van Diest PJ: Whole slide images as a platform for initial diagnostics in histopathology in a medium-sized routine laboratory. J Clin Pathol 2012; 65: 1107–1111.

45 Pantanowitz L, Wiley CA, Demetris A, Les-niak A, Ahmed I, Cable W, Contis L, Parwani AV: Experience with multimodality telepa-thology at the University of Pittsburgh Medi-cal Center. J Pathol Inform 2012; 3: 45.

46 Al-Janabi S, Huisman A, Willems SM, vanDiest PJ: Digital slide images for primary di-agnostics in breast pathology: a feasibility study. Hum Pathol 2012; 43: 2318–2325.

47 Reyes C, Ikpatt OF, Nadji M, Cote RJ: Intra-observer reproducibility of whole slide imag-ing for the primary diagnosis of breast needle biopsies. J Pathol Inform 2014; 5: 5.

48 Lopez AM, Graham AR, Barker GP, Richter LC, Krupinski EA, Lian F, Grasso LL, Miller A, Kreykes LN, Henderson JT, Bhattacharyya AK, Weinstein RS: Virtual slide telepathology enables an innovative telehealth rapid breast care clinic. Semin Diagn Pathol 2009; 26: 177–186.

49 Nassar A, Cohen C, Albitar M, Agersborg SS, Zhou W, Lynch KA, Heyman ER, Lange H, Siddiqui MT: Reading immunohistochemical slides on a computer monitor – a multisite performance study using 180 HER2-stained breast carcinomas. Appl Immunohistochem Mol Morphol 2011; 19: 212–217.

50 Khurana KK: Telecytology and its evolving role in cytopathology. Diagn Cytopathol 2012; 40: 498–502.

51 Thrall M, Pantanowitz L, Khalbuss W: Tele-cytology: clinical applications, current chal-lenges, and future benefits. J Pathol Inform 2011; 2: 51.

52 Donnelly AD, Mukherjee MS, Lyden ER, Bridge JA, Lele SM, Wright N, McGaughey MF, Culberson AM, Horn AJ, Wedel WR, Ra-dio SJ: Optimal z-axis scanning parameters for gynecologic cytology specimens. J Pathol Inform 2013; 4: 38.

53 Wilbur DC: Digital cytology: current state of the art and prospects for the future. Acta Cy-tol 2011; 55: 227–238.

54 Buxbaum JL, Eloubeidi MA, Lane CJ, Varada-rajulu S, Linder A, Crowe AE, Jhala D, Jhala NC, Crowe DR, Eltoum IA: Dynamic telecy-tology compares favorably to rapid onsite evaluation of endoscopic ultrasound fine nee-dle aspirates. Dig Dis Sci 2012; 57: 3092–3097.

55 Collins BT: Telepathology in cytopathology: challenges and opportunities. Acta Cytol 2013; 57: 221–232.

56 Al Habeeb HA, Evans A, Ghazarian D: Vir-tual microscopy using whole-slide imaging as an enabler for teledermatopathology: a paired consultant validation study. J Pathol Inform 2012; 3: 2.

57 Nielsen PS, Lindebjerg J, Rasmussen J, Starklint H, Waldstrom M, Nielsen B: Virtual microscopy: an evaluation of its validity and diagnostic performance in routine histologic diagnosis of skin tumors. Hum Pathol 2010; 41: 1770–1776.

58 Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJ, Ten Kate FJ, van Diest PJ: Whole slide images for primary diagnostics of gastrointestinal tract pathology: a feasibility study. Hum Pathol 2012; 43: 702–707.

59 Molnar B, Berczi L, Diczhazy C, Tagscherer A, Varga SV, Szende B, Tulassay Z: Digital slide and virtual microscopy based routine and telepathology evaluation of routine gas-trointestinal biopsy specimens. J Clin Pathol 2003; 56: 433–438.

60 Risio M, Bussolati G, Senore C, Vigna S, Fran-gipane E, Segnan N, Cassoni P: Virtual mi-croscopy for histology quality assurance of screen-detected polyps. J Clin Pathol 2010; 63: 916–920.

61 Camparo P, Egevad L, Algaba F, Berney DM, Boccon-Gibod L, Comperat E, Evans AJ, Grobholz R, Kristiansen G, Langner C, Lo-pez-Beltran A, Montironi R, Oliveira P, Vain-er B, Varma M: Utility of whole slide imaging and virtual microscopy in prostate pathology. APMIS 2012; 120: 298–304.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M



98

62 Helin HO, Lundin ME, Laakso M, Lundin J, Helin HJ, Isola J: Virtual microscopy in pros-tate histopathology: simultaneous viewing of biopsies stained sequentially with hematoxy-lin and eosin, and alpha-methylacyl-coen-zyme A racemase/p63 immunohistochemis-try. J Urol 2006; 175: 495–499.

63 Chargari C, Comperat E, Magne N, Vedrine L, Houlgatte A, Egevad L, Camparo P: Pros-tate needle biopsy examination by means of virtual microscopy. Pathol Res Pract 2011; 207: 366–369.

64 Fine JL, Grzybicki DM, Silowash R, Ho J, Gil-bertson JR, Anthony L, Wilson R, Parwani AV, Bastacky SI, Epstein JI, Jukic DM: Evalu-ation of whole slide image immunohisto-chemistry interpretation in challenging pros-tate needle biopsies. Hum Pathol 2008; 39: 564–572.

65 Helin H, Lundin M, Lundin J, Martikainen P, Tammela T, Helin H, van der Kwast T, Isola J: Web-based virtual microscopy in teaching and standardizing Gleason grading. Hum Pathol 2005; 36: 381–386.

66 Rodriguez-Urrego PA, Cronin AM, Al-Ah-madie HA, Gopalan A, Tickoo SK, Reuter VE, Fine SW: Interobserver and intraobserver re-producibility in digital and routine micro-scopic assessment of prostate needle biopsies. Hum Pathol 2011; 42: 68–74.

67 Al-Janabi S, Huisman A, Jonges GN, Ten Kate FJ, Goldschmeding R, van Diest PJ: Whole slide images for primary diagnostics of uri-nary system pathology: a feasibility study. J Renal Inj Prev 2014; 3: 91–96.

68 Ho J, Parwani AV, Jukic DM, Yagi Y, Antho-ny L, Gilbertson JR: Use of whole slide imag-ing in surgical pathology quality assurance: design and pilot validation studies. Hum Pathol 2006; 37: 322–331.

69 Sturm B, Fleskens SJ, Bot FJ, van Velthuysen ML, Speel EJ, Slootweg PJ, van der Laak JA: Virtual microscopy is a valid alternative for the diagnostic assessment of laryngeal prema-lignancies. Histopathology 2014; 64: 602–604.

70 Wiley CA, Murdoch G, Parwani A, Cudahy T, Wilson D, Payner T, Springer K, Lewis T: In-terinstitutional and interstate teleneuropa-thology. J Pathol Inform 2011; 2: 21.

71 Grala B, Markiewicz T, Kozlowski W, Osow-ski S, Slodkowska J, Papierz W: New automat-ed image analysis method for the assessment of Ki-67 labeling index in meningiomas. Folia Histochem Cytobiol 2009; 47: 587–592.

72 Alomari YM, Sheikh Abdullah SN, MdZin RR, Omar K: Adaptive localization of focus point regions via random patch probabilistic density from whole-slide, Ki-67-stained brain tumor tissue. Comput Math Methods Med 2015; 2015: 673658.

73 Al-Janabi S, Huisman A, Nikkels PG, Ten Kate FJ, van Diest PJ: Whole slide images for primary diagnostics of paediatric pathology specimens: a feasibility study. J Clin Pathol 2013; 66: 218–223.

74 Arnold MA, Chenever E, Baker PB, Boue DR, Fung B, Hammond S, Hendrickson BW, Kah-wash SB, Pierson CR, Prasad V, Nicol KK, Barr T: The College of American Pathologists guidelines for whole slide imaging validation are feasible for pediatric pathology: a pediat-ric pathology practice experience. Pediatr Dev Pathol 2015; 18: 109–116.

75 Slodkowska J, Chyczewski L, Wojciechowski M: Virtual slides: application in pulmonary pathology consultations. Folia Histochem Cytobiol 2008; 46: 121–124.

76 Jen KY, Olson JL, Brodsky S, Zhou XJ, Na-dasdy T, Laszik ZG: Reliability of whole slide images as a diagnostic modality for renal al-lograft biopsies. Hum Pathol 2013; 44: 888–894.

77 Furness P: A randomized controlled trial of the diagnostic accuracy of internet-based tele-pathology compared with conventional mi-croscopy. Histopathology 2007; 50: 266–273.

78 Baak JP, van Diest PJ, Meijer GA: Experience with a dynamic inexpensive video-conferenc-ing system for frozen section telepathology. Anal Cell Pathol 2000; 21: 169–175.

79 Ramey J, Fung KM, Hassell LA: Use of mobile high-resolution device for remote frozen sec-tion evaluation of whole slide images. J Pathol Inform 2011; 2: 41.

80 Buck TP, Dilorio R, Havrilla L, O’Neill DG: Validation of a whole slide imaging system for primary diagnosis in surgical pathology: a community hospital experience. J Pathol In-form 2014; 5: 43.

81 Campbell WS, Lele SM, West WW, Lazenby AJ, Smith LM, Hinrichs SH: Concordance be-tween whole-slide imaging and light micros-copy for routine surgical pathology. Hum Pathol 2012; 43: 1739–1744.

82 Dangott B, Parwani A: Whole slide imaging for teleconsultation and clinical use. J Pathol Inform 2010; 1: 7.

83 Song Y, Treanor D, Bulpitt AJ, Magee DR: 3D reconstruction of multiple stained histology images. J Pathol Inform 2013; 4:S7.

84 Caie PD, Turnbull AK, Farrington SM, Onis-cu A, Harrison DJ: Quantification of tumour budding, lymphatic vessel density and inva-sion through image analysis in colorectal can-cer. J Transl Med 2014; 12: 156.

85 Murakami Y, Abe T, Hashiguchi A, Yamagu-chi M, Saito A, Sakamoto M: Color correction for automatic fibrosis quantification in liver biopsy specimens. J Pathol Inform 2013; 4: 36.

86 Neil DA, Roberts IS, Bellamy CO, Wigmore SJ, Neuberger JM: Improved access to histo-pathology using a digital system could in-crease the organ donor pool and improve al-location. Transpl Int 2014; 27: 759–764.

87 Neltner JH, Abner EL, Schmitt FA, Denison SK, Anderson S, Patel E, Nelson PT: Digital pathology and image analysis for robust high-throughput quantitative assessment of Alz-heimer disease neuropathologic changes. J Neuropathol Exp Neurol 2012; 71: 1075–1085.

88 Riber-Hansen R, Vainer B, Steiniche T: Digi-tal image analysis: a review of reproducibility, stability and basic requirements for optimal results. APMIS 2012; 120: 276–289.

89 Webster JD, Michalowski AM, Dwyer JE, Corps KN, Wei BR, Juopperi T, Hoover SB, Simpson RM: Investigation into diagnostic agreement using automated computer-assist-ed histopathology pattern recognition image analysis. J Pathol Inform 2012; 3: 18.

90 Webster JD, Dunstan RW: Whole-slide imag-ing and automated image analysis: consider-ations and opportunities in the practice of pa-thology. Vet Pathol 2014; 51: 211–223.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

12:

01:4

3 P

M


[63]

Estudio número 2

“Validation of Whole-Slide Imaging in the Primary Diagnosis of

Gynecological Pathology in a University Hospital”

Jaume Ordi, Paola Castillo, Adela Saco, Marta del Pino, Oriol Ordi,

Leonardo Rodríguez-Carunchio, Jose Ramírez

Journal of Clinical Pathology 2015 Jan; 68 (1): 33 – 9


Ranking (2016): 33/193, primer cuartil


[64]

Validation of whole slide imaging in the primarydiagnosis of gynaecological pathology in aUniversity HospitalJaume Ordi,1,2,3 Paola Castillo,1,3 Adela Saco,1 Marta del Pino,4 Oriol Ordi,2

Leonardo Rodríguez-Carunchio,1 Jose Ramírez1,2

▸ Additional material ispublished online only. To viewplease visit the journal online(http://dx.doi.org/10.1136/jclinpath-2014-202524).1Department of Pathology,Hospital Clínic, Barcelona,Spain2University of Barcelona,School of Medicine, Barcelona,Spain3Centre de Recerca en SalutInternacional de Barcelona(CRESIB), Barcelona, Spain4Institute of Gynecology,Obstetrics and Neonatology,Hospital Clínic—Institut d´Investigacions BiomèdiquesAugust Pi I Sunyer (IDIBAPS),Faculty of Medicine-Universityof Barcelona, Barcelona, Spain

Correspondence toProfessor Jaume Ordi,Department of Pathology,CRESIB (Centre de Recerca enSalut Internacional deBarcelona) -Hospital Clínic,University of Barcelona Facultyof Medicine, Barcelona, Spain,C/Villarroel 170, 08036-Barcelona, Spain;[email protected]

Presented in part at the annualmeeting of the USA andCanadian Academy ofPathology; March 2014; SanDiego, California, USA.

Received 18 July 2014Revised 23 September 2014Accepted 8 October 2014Published Online First29 October 2014

To cite: Ordi J, Castillo P,Saco A, et al. J Clin Pathol2015;68:33–39.

ABSTRACTAims Experience in the use of whole slide imaging(WSI) for primary diagnosis in pathology is very limited.We aimed to determine the accuracy of interpretation ofWSI compared with conventional light microscopy (CLM)in the diagnosis of routine gynaecological biopsies.Methods All gynaecological specimens (n=452)received over a 2-month period at the Department ofPathology of the Hospital Clinic of Barcelona wereanalysed blindly by two gynaecological pathologists, oneusing CLM and the other WSI. All slides were digitisedin a Ventana iScan HT (Roche diagnostics) at 200×.All discrepant diagnoses were reviewed, and a finalconsensus diagnosis was established. The results wereevaluated by weighted κ statistics for two observers.Results The level of interobserver agreement betweenWSI and CLM evaluations was almost perfect (κ value:0.914; 95% CI 0.879 to 0.949) and increased duringthe study period: κ value 0.890; 95% CI 0.835 to0.945 in the first period and 0.941; 95%; CI 0.899 to0.983 in the second period. Major discrepancies(differences in clinical management or prognosis) wereobserved in 9 cases (2.0%). All discrepancies consistedof small lesions (8 high grade squamous intraepitheliallesions of the uterine cervix, one lymph nodemicrometastasis of an ovarian carcinoma)underdiagnosed or missed in the WSI or the CLMevaluation. Discrepancies with no or minor clinicalrelevance were identified in 3.8% of the biopsies. Nodiscrepancy was related to the poor quality of the WSIimage.Conclusions Diagnosis of gynaecological specimens byWSI is accurate and may be introduced into routinediagnosis.

INTRODUCTIONWhole slide imaging (WSI), also referred to asvirtual microscopy or digital pathology allows digit-isation of entire glass slides to achieve the diagnosisof pathological specimens. WSI scanners create adigital slide of the tissue section which, with theuse of specific software, can be viewed and magni-fied in real time across the web very much like theuse of a conventional light microscope (CLM).1–3

WSI has been shown to have many practical appli-cations including education,4–6 teleconsultation forsecond opinions,7–10 intraoperative frozen sectionconsultation10 11 and quality assurance.12 13

Potential additional benefits of WSI includeimprovement of the workflow by eliminating thetask of slide distribution and facilitating slide filingand retrieval.1–3

The rapid advances in this technology and itsmany potential benefits will probably result in aprogressive shift from conventional to virtualmicroscopy in the routine diagnosis in pathology.Currently, several commercially available systemsare able to digitise glass slides containing tissue sec-tions and produce virtual slides of excellent quality.However, although WSI has been around for morethan a decade, its widespread application inprimary histological diagnosis still awaits validationas opposed to traditional CLM. Moreover, despiteseveral pilot studies suggesting that WSI is as usefulas CLM for diagnostic purposes,14–22 WSI-baseddiagnosis has yet to be integrated in routine patho-logical studies, with a very small number of excep-tions. The use of WSI in the routine practice ofpathology laboratories is still not common becauseof difficulties in the integration between the WSIsoftware and the laboratory information systems(LIS), insufficient scanning speed and robustness,and limitations in storage capacity and/or excessivecosts of file storage. The lack of systematic valid-ation studies on their use in primary diagnosis isalso a major concern that hampers the introductionof this technique.1 3

The College of American Pathologists Pathologyand Laboratory Quality Center has recently pub-lished the guidelines for validating the use of WSIfor diagnostic purposes.3 According to these guide-lines all laboratories should conduct their own val-idation studies, and the validation should include atleast 60 samples. However, very few large valid-ation studies have focused on pathology subspecial-ties. Indeed, only one study has been published ongynaecological disease, evaluating the correlationbetween WSI and CLM in the assessment of thediagnoses of frozen sections of 52 ovarianlesions.11 No previous studies have evaluated theaccuracy of WSI diagnosis in the routine practice ofgynaecological pathology. In the present study weaimed to determine the accuracy of interpretationusing WSI as compared with CLM in a series ofconsecutive gynaecological specimens, representa-tive of the spectrum of specimen types and diagno-ses encountered in the routine practice of a largeacademic department.

MATERIALS AND METHODSCharacteristics of the institutionThis study was performed at the Department ofPathology, Hospital Clinic (Barcelona, Spain), alarge academic department composed of 15 staffpathologists, 8 residents and additional fellows.

Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524 33

Original article

group.bmj.com on September 2, 2017 - Published by http://jcp.bmj.com/Downloaded from

http://dx.doi.org/10.1136/jclinpath-2014-202524



http://crossmark.crossref.org/dialog/?doi=10.1136/jclinpath-2014-202524&domain=pdf&date_stamp=2014-10-29

http://jcp.bmj.com/

http://www.pathologists.org.uk/

http://jcp.bmj.com/

http://group.bmj.com

There are 14 subspecialties, and most pathologists limit theirpractice to one or two subspecialty areas. In 2013 theDepartment handled 41 928 specimens with 83 619 blocks. Thenumber of gynaecological specimens analysed during 2013 was4909, with 16 805 blocks of gynaecological specimens and witha median number of slides per case of 1.

Sample size calculationBased on previous reports,15 23 the major rate of discrepancybetween the original diagnosis by CLM and that by WSI wascalculated to be 3%, with a non-inferiority margin for WSIreview of 4%. A one-sided binomial test was used for compari-son at a level of significance of 0.05. The power to be achievedwas 80%, and the level of significance was 0.05. Based on theseassumptions, it was calculated that 450 cases would need to bereviewed to establish non-inferiority.

Specimens included in the studyAll gynaecological specimens consecutively received over a2-month period ( July–August 2013) were included in the studyat the Department of Pathology of the Hospital Clinic ofBarcelona (n=452). This represented 9.21% (452/4909) of thetotal number of gynaecological specimens evaluated in 2013:353/452 (78.1%) specimens were evaluated in the 1st month( July) and 99/452 (21.9%) in the 2nd month (August). Table 1shows a summary of the location and type (either biopsy or resec-tion) of the specimens, and the number of cases included foreach type. The number of blocks per case ranged from 1 to 30(median 1, IQR 1–3). The overall number of slides scanned was1253.

To evaluate possible changes associated with increasingexperience in the use of WSI over the study period and theagreement between WSI and CLM observers, the 456 specimenswere divided into two periods: one including the first 226 speci-mens and the second including the last 226 specimens.

The Department of Gynecology of our hospital has a veryactive referral Colposcopy Clinic.24 Consequently, the seriesincluded 157 biopsies or excisions of the uterine cervix frompatients referred to colposcopy because of an abnormal Papsmear. In addition to evaluation in the general analysis, thesespecimens were evaluated independently.

The study was approved by the Hospital Clinic InstitutionalEthical Review Board.

Scanning process and characteristics of the WSI displayIn July and August 2013 all the slides of gynaecological path-ology were scanned daily after diagnosis by light microscopy.Scanning was performed on a Ventana iScan HT (RocheDiagnostics, Sant Cugat, Spain) at a magnification of 200×. Thesystem creates high-resolution digital images of the tissue sec-tions. The whole scanning process runs automatically (includingselection of the area of the slide that contains tissue, placingfocus points, calibration, etc). In cases with step sections of asample on a single slide the system scans all the sections. Nospecific quality control of the slides scanned was made by thetechnicians prior to evaluation by the pathologist. The WSI pro-duced are stored in a dedicated mass storage environment andlinked to the pathology report, based on the recognition of aquick response (QR) code on the slide label. Although WSI canbe accessed through the pathology LIS software (Novopath,Vitrosoft, Sevilla, Spain), for the purposes of this study theaccession to the WSI was made through the viewer.

The images are viewed in the Virtuoso viewer (Roche), whichworks as a web browser and simulates a conventional micro-scope. The images are shown using the same structure providedby the LIS. No specific software installation is required to visual-ise the WSI. The images scanned can be viewed up to a realmagnification of 200× and up to 400× with digital zoom, arealways in focus, with optimised contrast and adjusted illumin-ation. The viewer shows a thumbnail of the whole slide, whichallows confirmation that all the material present on the glassslide has been included in the digital image and helps in thenavigation through the slide. Figure 1 shows the appearance ofthe virtual microscope display.

The WSI were displayed on a 30” Coronis fusion MDC4130monitor which has a resolution of four megapixels (BarcoElectronic Systems, Barcelona, Spain).

WSI and CLM diagnosisAll cases were analysed blindly by two gynaecological patholo-gists, one using CLM and the other WSI. The pathologist doingthe WSI evaluations had previously had a 1-week trainingcourse on the use of WSI. WSI were presented to the patholo-gist per case, together with the original clinical information inorder to emulate the real clinical environment, and blinded tothe original report based on CLM. For the purposes of thisstudy only the H&E slides were evaluated.

The original CLM and the WSI-based diagnoses were com-pared by an independent gynaecologist, who judged the con-cordance of the two diagnoses as: (A) complete agreementbetween the original diagnosis and that determined with WSI;(B) minor discrepancy (mild differences which would not have

Table 1 Location and type (either small biopsies or surgicalresections) of the specimens evaluated in the study

Location Number (%) Biopsies Surgical specimens

Vulva 19 (4) 13 6Vagina 12 (3) 12 1Uterine cervix 125 (28) 93 32Endocervix 46 (10) 46 0Endometrium 60 (13) 52 8Uterus 69 (15) 0 69Fallopian tube 12 (3) 0 12Ovary 36 (8) 0 36Lymph nodes 44 (10) 0 44Peritoneum 14 (3) 0 14Other * 15 (3) 0 15

*Includes biopsies of the abdominal wall (1), large bowel (9), urinary bladder (4) andureter (1). Figure 1 Screenshot of the virtual microscope display.

34 Ordi J, et al. J Clin Pathol 2015;68:33–39. doi:10.1136/jclinpath-2014-202524

Original article


http://jcp.bmj.com/


any clinical or prognostic implications); and (C) major discrep-ancy (differences with clinical and/or prognostic implications forthe patient).

Final gold standard diagnosisThe gold standard was considered as the concordant diagnosisin all cases with complete agreement in both evaluations. Eachcase with a discrepant result was reviewed by the two patholo-gists involved in the study. The revision was made using CLM,and a final consensus diagnosis was established. In this finaladjudication process, the H&E slides, as well as the immunohis-tochemical stains (when necessary) were used to achieve thefinal diagnosis.

Statistical analysisThe SPSS statistical programme (SPSS TM140, V.18, Chicago,Illinois, USA) was used for statistical analysis. The results for cat-egorical variables are expressed as absolute numbers and percen-tages and 95% CIs. The χ2 or the Fisher’s exact tests were usedto compare the variables. The results were evaluated byweighted κ statistics for two raters. This measure calculates thedegree of agreement in classification over that which would beexpected by chance and is scored as a number between 0 and1. Following the Landis-Koch benchmarks the strength of agree-ment of the κ values is: 0 poor; 0–0.20 slight; 0.21–0.40 fair;0.41–0.60 moderate; 0.61–0.80 substantial; 0.81–1.00 almostperfect.25 For the purposes of weighted κ calculation the diag-noses were categorised from normal to cancer as 1: normaltissue or reactive lesions; 2: benign tumours; 3: low-grade pre-malignant lesions; 4: high-grade premalignant lesions; and 5:malignant tumours.

RESULTSFinal diagnosesOverall, 218/452 specimens (48.2%) were evaluated as beingcomposed of normal tissue or showing reactive lesions; 130(28.8%) were benign tumours, 18 (4.0%) were low-grade pre-malignant lesions, 48 (10.6%) high grade premalignant lesionsand 38 (8.4%) showed malignant tumours. Table 2 shows thedistribution of these diagnoses for each specific site.

Agreement between WSI and CLM evaluationsInterpretations by WSI and CLM completely agreed in 94.2%of the biopsies (95% CI 91.7 to 96.0) while major discrepancieswere observed in 9/452 (2.0%) and minor discrepancies wereidentified in 17/452 (3.8%) of the biopsies.

The final consensus diagnosis achieved after the adjudicationmeeting was in agreement with the CLM evaluation in 7/9(77.8%) major discrepancies and in 11/17 (64.7%) minor dis-crepancies. Discrepancy in interpretations between WSI andCLM evaluations occurred in only two settings. Eight out of thenine major discrepancies (88.9%) observed in the study wererelated to the diagnosis of a high-grade squamous intraepitheliallesion (H-SIL) of the uterine cervix, as a low-grade squamousintraepithelial lesion (L-SIL) or as negative (four cases each,figure 2). The consensus diagnosis was in keeping with theCLM evaluation in six of eight cases and with the WSI evalu-ation in two of eight cases. The last case was a small lymphnode metastasis of an ovarian carcinoma missed in the WSIevaluation (figure 3). In this latter case a small lipogranulomawas identified close to the small metastatic nest missed in theWSI evaluation. Thirteen out of the 17 minor discrepancieswere related to overdiagnosis or underdiagnosis of L-SIL. In 10cases the L-SIL lesions were missed in the evaluation (8 in the

WSI and 2 in the CLM evaluation). Three cases were reactivechanges in the uterine cervix overdiagnosed as L-SIL (threebiopsies, two overdiagnosed in the WSI and one in the CLMevaluation). Two cases of endometrial polyps were missed (onecase missed in the WSI evaluation) or overdiagnosed (one case,overdiagnosed in the CLM evaluation). The other two minordiscrepancies were two small foci of endometriosis (one in theovary, one in the Fallopian tube) missed in the CLM evaluation.None of the discrepancies was related to the poor quality of theWSI image or to insufficient magnification.

Overall the level of interobserver agreement between the WSIand CLM evaluations was almost perfect (κ value: 0.914; 95%CI 0.879 to 0.949).

Concordance in biopsies of the uterine cervixand in other samplesIn the subset of 157 biopsies or excisions of the uterine cervixfrom patients referred to colposcopy because of abnormal Papsmear, complete agreement was observed between the WSI andCLM interpretations in 86.6% (95% CI 80.3 to 91.5) of thebiopsies. Major discrepancies were observed in 8/157 (5.1%)and minor discrepancies in 13/157 (8.3%) of the samples. Theκ value for this subset of samples was 0.832 (95% CI 0.757 to0.906).

In the 295 gynaecological specimens representing tissues otherthan cervical biopsies and excisions, complete agreement betweenWSI and CLM was observed in 98.3% (95% CI 96.1 to 99.4) ofthe biopsies. Major discrepancies were observed in 1/295 (0.3%)and minor discrepancies in 4/295 (1.4%) of the samples. The κvalue for this subset of samples was 0.976 (95% CI 0.950 to 1).

κ Analysis and discrepant diagnoses in thetwo study periodsInterobserver agreement increased during the study period,κ value: 0.890 (95% CI 0.835 to 0.945) in the first period, andκ value: 0.941 (95% CI 0.899 to 0.983) in the second period.In the first period of the study 5/226 (2.2%) major discrepanciesand 12/226 (5.3%) minor discrepancies were detected. Thenumber of major and minor discrepancies in the second periodwas 4/226 (1.80%) and 5/226 (2.2%), respectively. Interestingly,whereas the consensus gold standard diagnosis was in keepingwith the CLM diagnosis in 14/17 (82.4%) discrepanciesobserved in the initial period, the consensus was in keepingwith the WSI evaluation in 5/9 discrepancies (55.5%) observedin the second period (p=0.078).

DISCUSSIONThe results of our study show a high concordance between WSIand CLM evaluations (over 94%) in the diagnosis of a largeseries of routine gynaecological specimens. The κ value, consid-ered as a measure of the level of agreement among observerscorrected by chance, was at the almost perfect level (0. 914).Thus, our results confirm that WSI may safely be used for per-forming primary histological diagnoses of gynaecologicalspecimens.

The results of our study are comparable with other valid-ation studies conducted on skin,5 19 breast,14 26 prostate,17 27

urinary bladder,18 gastrointestinal8 20 and paediatric path-ology,28 which show a high rate of concordance between WSIand CLM-based diagnoses. However, no previous studies haveevaluated the accuracy of WSI diagnosis in the routine practiceof gynaecological pathology, and neither are there any dataabout intraobserver or interobserver agreement in the evalu-ation of routine gynaecological specimens using CLM. The rate


Original article


http://jcp.bmj.com/


of discrepancies observed in our study is within the range ofgenerally observed intraobserver variability in pathology.27 29

Interestingly, the final consensus diagnosis was in agreementwith the WSI evaluation in 22.2% of the major discrepanciesand in 35.3% of the minor discrepancies. None of the

discrepancies was related to the poor quality of the WSI imageor to insufficient magnification, but rather were mostly asso-ciated with different interpretations of difficult or borderlinecases or with the presence of small lesions overlooked in theevaluation.

Table 2 Absolute numbers and percentages (in parenthesis) of specimens showing normal/reactive lesions, benign tumours, low-gradepremalignant, high-grade premalignant and malignant tumours in each location

Location Normal/reactive Benign tumoursLow-gradepremalignant lesions

High-gradepremalignant lesions Malignant tumours

Vulva 7 (36.8) 6 (31.6) 0 (0) 5 (26.3) 1 (5.3)Vagina 8 (66.7) 0 (0) 0 (0) 0 (0) 4 (33.3)Uterine cervix 64 (51.2) 3 (2.4) 16 (12.8) 37 (29.6) 5 (4.0)Endocervix 30 (65.2) 9 (19.6) 2 (4.3) 3 (6.6) 2 (4.3)Endometrium 29 (48.3) 28 (46.7) 0 (0) 1 (1.7) 2 (3.3)Uterus 13 (18.8) 46 (66.7) 0 (0) 2 (2.9) 8 (11.6)Fallopian tube 7 (58.3) 5 (41.7) 0 (0) 0 (0) 0 (0)Ovary 8 (22.2) 24 (66.7) 0 (0) 0 (0) 4 (11.1)Lymph nodes 41 (93.2) 0 (0) 0 (0) 0 (0) 3 (6.8)Peritoneum 6 (42.9) 0 (0) 0 (0) 0 (0) 8 (57.1)Other * 5 (33.3) 9 (60.0) 0 (0) 0 (0) 1 (6.7)

*Includes biopsies of the abdominal wall (1), large bowel (9), urinary bladder (4) and ureter (1).

Figure 2 One of the major discrepancies identified in the study: a small area of high-grade squamous intraepithelial lesion (H-SIL) involving thesquamous epithelium of the uterine cervix diagnosed as reactive changes in the whole slide imaging (WSI) evaluation. In the consensus meeting ap16 staining was requested, which confirmed the diagnosis of H-SIL rendered by conventional light microscopy (A and C, H&E; B and D p16immunostaining; screenshots of the WSI image).


Original article


http://jcp.bmj.com/


Eight out of the nine discrepancies (88.9%) observed in thestudy were related to the diagnosis of H-SIL as L-SIL or as nega-tive or reactive changes in the uterine cervix (four cases each).Similarly, 13 out of the 17 minor discrepancies (76.5%) wererelated to discrepancies in the diagnosis of L-SIL lesions versusnormal/reactive cervical epithelium. Consequently, the κ valuefor the subset of cervical biopsies or excisions from patientsreferred because of abnormal Pap smear was 0.832, clearly lowerthan the general value. A number of studies using CLM haveshown that there is a substantial variation between andwithin-observers in the interpretation of squamous intraepitheliallesions on H&E-stained tissue sections. Indeed, κ coefficients aretypically found within the range of 0.45–0.50,29–36 indicatingmoderate agreement. Estimates for the reproducibility of histo-logical cervical specimen interpretations performed in the courseof the ASCUS/L-SIL Triage Study comparing the diagnosticresults of the original clinical centre pathologists with the resultsfrom a quality control review35 showed that the reproducibilityof histological interpretations of biopsy specimens was moderate(κ<0.5). In the latter study, the lack of reproducibility was sub-stantially higher for punch biopsy specimens than loop excisionprocedure specimens, and variability was more prominent in thelow-grade abnormalities, similar to what was observed in our

study. The p16 immunohistochemical stain, strongly expressed inalmost all H-SIL and not in reactive lesions,36 has recently beenrecommended by the College of American Pathologists to reduceinterobserver variability, particularly in cases of professional dis-agreement.37 This technique was used in our series to achieve thefinal consensus diagnosis in all disagreements between CLM andWSI detected in biopsies of the uterine cervix.

The pathologist who performed the WSI evaluation had pre-viously had very little experience in the use of WSI, althoughthis did not severely affect the reproducibility, even in the initialperiod of the study. Nevertheless, a clear increase in the rate ofreproducibility was observed during the study period. This sug-gests that minor difficulties may arise in the initial periods ofusing this new tool and that increasing experience with WSIincreases the accuracy of the diagnosis. Moreover, whereas theconsensus diagnosis reached in the discrepant cases was inkeeping with the CLM diagnosis in most discrepancies observedin the initial period (82.4%), in the second period there was atendency towards a more balanced situation and in 55.5% ofthe cases the gold standard diagnosis was in keeping with theWSI evaluation.

The pathologist working with WSI did not report difficultiesin rendering the diagnosis at the magnification of 200× applied,

Figure 3 Screenshots of the only major discrepancy observed in the study not related to cervical pathology; a small lymph node metastasis of anovarian carcinoma, which was missed in the whole slide imaging (WSI) evaluation. (A) Low power magnification showing several lymph nodesincluded in the slide; (B) higher magnification showing the two areas of interest located in the centre of the largest node (green squares); (C) highermagnification of the area identified with the large green square in B, showing a lipogranuloma with a large fatty vacuole and giant macrophages inthe periphery; (D) higher magnification of the area identified in the small green square in B, showing a small nest of metastatic carcinoma cells,which were missed in the WSI evaluation.


Original article


http://jcp.bmj.com/


indicating that a higher magnification does not seem to be rele-vant for most cases. This scanning strategy which is used inmost validation studies,15 17 19–21 23 28 significantly saves scan-ning time and storage requirements. However, when WSI is rou-tinely used for primary diagnosis it is very likely that thepathologist would require an increase in the scanning magnifica-tion in a small percentage of cases to safely establish the diagno-sis. Although no formal timing was performed, the pathologistdoing the WSI evaluation perceived the diagnostic process to bea little slower.

The main strength of our study is that it is the largest valid-ation study focused on gynaecological pathology that includes asufficient number of cases to allow robust statistical power.Indeed, only one published study has analysed the correlationbetween WSI and CLM in the evaluation of frozen section diag-nosis of 52 ovarian lesions, showing that, as observed in ourstudy, the correlation between CLM and WSI diagnoses is verygood.11 The second strength of our study is that the participat-ing pathologists were aware of the clinical information whichcould affect the diagnostic outcome.

The main limitation of this study is that intraobserver variabil-ity, which is considered the best design to confirm the reprodu-cibility of the results obtained with two different techniques,was not evaluated.38 However, the very good interobserverreproducibility observed in our series suggests that very similardata would be obtained when cases are evaluated by the sameobserver.

A potential advantage of WSI is that it allows performingimage analysis. This may assist in the objective evaluation of thesize of the tumours and the depth of the invasion, which arerelevant information for tumours of the vulva, cervix and endo-metrium. Eventually, this may even permit computer-assisteddiagnosis that could help improve the diagnosis and decreaseinterobserver variability. Several legal issues have arisen from theuse of WSI for primary diagnosis related to image quality, imagepresentation (monitor quality), storage space, adequate backup,document transfer, patient confidentiality and the confidence ofthe pathologist to sign out a pathology report depending onWSI. Most of these issues will probably be settled in the nearfuture. Several digital pathology vendors are currently seekingapproval from the US Food and Drug Administration to useWSI in primary diagnosis, which will definitely encourage thegeneral use of this technique.

In conclusion, the diagnosis of gynaecological specimensusing WSI shows a high concordance with the results of CLMevaluation. Our results confirm that WSI may safely be used forperforming primary histological diagnosis of gynaecologicalspecimens.

Acknowledgements The authors are grateful to Rosana Millan, MargaritaMainar, Berta Coloma, Ingrid Rubio, Olga Ten, Gemma Laguna, Silvia Moya, andArantxa Sánchez, members of the technical and administrative staff of thedepartment of Pathology, for their support in the scanning of the slides. The authorsthank Antonio Teruel, Anna Rubi, Jaume Barderi, Jose Antonio Collados and EnricVidal, members of the staff of Roche for their technical support. The authors thankDonna Pringle for English revision of the manuscript.

Contributors JO: conception, design and conduct of the study, analysis andwriting of the manuscript, approval of the final version of the manuscript. PC, AS,LR-C: conduct of the study, writing of the manuscript, approval of the final versionof the manuscript. OO: conception, design and writing of the manuscript, approvalof the final version of the manuscript. MdP: conception of the study, the statisticalanalysis and writing of the manuscript, approval of the final version of themanuscript. JR: design of the study, writing of the manuscript, approval of the finalversion of the manuscript.

Competing interests None.

Ethics approval Ethical committee of clinical research of Hospital Clinic ofBarcelona.

Provenance and peer review Not commissioned; externally peer reviewed.

REFERENCES1 Al-Janabi S, Huisman A, Van Diest PJ. Digital pathology: current status and future

perspectives. Histopathology 2012;61:1–9.2 Brachtel E, Yagi Y. Digital imaging in pathology—current applications and

challenges. J Biophotonics 2012;5:327–35.3 Pantanowitz L, Valenstein PN, Evans AJ, et al. Review of the current state of whole

slide imaging in pathology. J Pathol Inform 2011;2:36.4 Pantanowitz L, Szymas J, Yagi Y, et al. Whole slide imaging for educational

purposes. J Pathol Inform 2012;3:46.5 Brick KE, Comfere NI, Broeren MD, et al. The application of virtual microscopy in a

dermatopathology educational setting: assessment of attitudes amongdermatopathologists. Int J Dermatol 2014;53:224–7.

6 Carlson AM, McPhail ED, Rodriguez V, et al. A prospective, randomized crossoverstudy comparing direct inspection by light microscopy versus projected images forteaching of hematopathology to medical students. Anat Sci Educ 2014;7:130–4.

7 Leong FJ, McGee JO. Automated complete slide digitization: a medium forsimultaneous viewing by multiple pathologists. J Pathol 2001;195:508–14.

8 Singson RP, Natarajan S, Greenson JK, et al. Virtual microscopy and the Internet astelepathology consultation tools. A study of gastrointestinal biopsy specimens. Am JClin Pathol 1999;111:792–5.

9 Wilbur DC, Madi K, Colvin RB, et al. Whole-slide imaging digital pathology as aplatform for teleconsultation: a pilot study using paired subspecialist correlations.Arch Pathol Lab Med 2009;133:1949–53.

10 Winokur TS, McClellan S, Siegal GP, et al. A prospective trial of telepathology forintraoperative consultation (frozen sections). Hum Pathol 2000;31:781–5.

11 Fallon MA, Wilbur DC, Prasad M. Ovarian frozen section diagnosis: use ofwhole-slide imaging shows excellent correlation between virtual slide and originalinterpretations in a large series of cases. Arch Pathol Lab Med 2010;134:1020–3.

12 Harnden P, Coleman D, Moss S, et al. Evaluation of the use of digital images for anational prostate core external quality assurance scheme. Histopathology2011;59:703–9.

13 Ho J, Parwani AV, Jukic DM, et al. Use of whole slide imaging in surgical pathologyquality assurance: design and pilot validation studies. Hum Pathol 2006;37:322–31.

14 Reyes C, Ikpatt OF, Nadji M, et al. Intra-observer reproducibility of whole slideimaging for the primary diagnosis of breast needle biopsies. J Pathol Inform2014;5:5.

15 Bauer TW, Schoenfield L, Slaw RJ, et al. Validation of whole slide imaging forprimary diagnosis in surgical pathology. Arch Pathol Lab Med 2013;137:518–24.

16 Onega T, Weaver D, Geller B, et al. Digitized whole slides for breast pathologyinterpretation: current practices and perceptions. J Digit Imaging 2014;27:642–8.

17 Camparo P, Egevad L, Algaba F, et al. Utility of whole slide imaging and virtualmicroscopy in prostate pathology. APMIS 2012;120:298–304.

18 Comperat E, Egevad L, Lopez-Beltran A, et al. An interobserver reproducibility studyon invasiveness of bladder cancer using virtual microscopy and heatmaps.Histopathology 2013;63:756–66.

19 Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primary diagnosticsin dermatopathology: a feasibility study. J Clin Pathol 2012;65:152–8.

Take home messages

▸ Interobserver agreement between whole slide imaging (WSI)and conventional light microscopy (CLM) evaluations is verygood, with κ values of over 0.91.

▸ Although the accuracy of WSI diagnosis is good even inusers with limited experience, the reproducibility betweenWSI and CLM improves over time, indicating that increasingexperience with WSI increases the accuracy of the diagnosis.

▸ WSI may safely be used for performing primary histologicaldiagnosis of gynaecological specimens using currentscanning technology and viewing interfaces.

▸ Scanning slides at the magnification of 20× is sufficient toachieve a correct diagnosis in most gynaecological biopsies.This scanning strategy significantly saves scanning time andstorage requirements.


Original article


http://jcp.bmj.com/


20 Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primarydiagnostics of gastrointestinal tract pathology: a feasibility study. Hum Pathol2012;43:702–7.

21 Gilbertson JR, Ho J, Anthony L, et al. Primary histologic diagnosis using automatedwhole slide imaging: a validation study. BMC Clin Pathol 2006;6:4.

22 Cornish TC, Swapp RE, Kaplan KJ. Whole-slide imaging: routine pathologicdiagnosis. Adv Anat Pathol 2012;19:152–9.

23 Bauer TW, Slaw RJ. Validating Whole-Slide Imaging for Consultation Diagnoses inSurgical Pathology. Arch Pathol Lab Med Published Online First: 9 May 2014.

24 Ordi J, Sagasta A, Munmany M, et al. Usefulness of p16/Ki67 immunostainingin the triage of women referred to colposcopy. Cancer Cytopathol2014;122:227–35.

25 Landis JR, Koch GG. The measurement of observer agreement for categorical data.Biometrics 1977;33:159–74.

26 Krishnamurthy S, Mathews K, McClure S, et al. Multi-institutional comparison ofwhole slide digital imaging and optical microscopy for interpretation ofhematoxylin-eosin-stained breast tissue sections. Arch Pathol Lab Med2013;137:1733–9.

27 Rodriguez-Urrego PA, Cronin AM, Al-Ahmadie HA, et al. Interobserver andintraobserver reproducibility in digital and routine microscopic assessment ofprostate needle biopsies. Hum Pathol 2011;42:68–74.

28 Al-Janabi S, Huisman A, Nikkels PG, et al. Whole slide images for primarydiagnostics of paediatric pathology specimens: a feasibility study. J Clin Pathol2013;66:218–23.

29 Nelson D, Ziv A, Bandali KS. Going glass to digital: virtual microscopy as asimulation-based revolution in pathology and laboratory science. J Clin Pathol2012;65:877–81.

30 Creagh T, Bridger JE, Kupek E, et al. Pathologist variation in reporting cervicalborderline epithelial abnormalities and cervical intraepithelial neoplasia. J ClinPathol 1995;48:59–60.

31 de Vet HC, Koudstaal J, Kwee WS, et al. Efforts to improve interobserver agreementin histopathological grading. J Clin Epidemiol 1995;48:869–73.

32 McCluggage WG, Walsh MY, Thornton CM, et al. Inter- and intra-observer variationin the histopathological reporting of cervical squamous intraepithelial lesions usinga modified Bethesda grading system. Br J Obstet Gynaecol 1998;105:206–10.

33 McCluggage WG, Bharucha H, Caughley LM, et al. Interobserver variation in thereporting of cervical colposcopic biopsy specimens: comparison of grading systems.J Clin Pathol 1996;49:833–5.

34 Robertson AJ, Anderson JM, Beck JS, et al. Observer variability in histopathologicalreporting of cervical biopsy specimens. J Clin Pathol 1989;42:231–8.

35 Stoler MH, Schiffman M. Interobserver reproducibility of cervical cytologic andhistologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study.JAMA 2001;285:1500–5.

36 Bergeron C, Ordi J, Schmidt D, et al. Conjunctive p16INK4a testing significantlyincreases accuracy in diagnosing high-grade cervical intraepithelial neoplasia. Am JClin Pathol 2010;133:395–406.

37 Darragh TM, Colgan TJ, Thomas CJ, et al. The Lower Anogenital Squamous TerminologyStandardization project for HPV-associated lesions: background and consensusrecommendations from the College of American Pathologists and the American Societyfor Colposcopy and Cervical Pathology. Int J Gynecol Pathol 2013;32:76–115.

38 Pantanowitz L, Sinard JH, Henricks WH, et al. Validating whole slide imaging fordiagnostic purposes in pathology: guideline from the College of AmericanPathologists Pathology and Laboratory Quality Center. Arch Pathol Lab Med2013;137:1710–22.


Original article


http://jcp.bmj.com/



[72]


[73]

Estudio número 3

“Validation of Whole-Slide Imaging in the Primary Diagnosis of

Liver Biopsies in a University Hospital”

Adela Saco, Alba Diaz, Monica Hernandez, Daniel Martinez, Carla

Montironi, Paola Castillo, Natalia Rakislova, Marta del Pino, Antonio

Martinez, Jaume Ordi

Dig Liver Dis. 2017 Jul 19. pii: S1590-8658(17)30977-5.

doi: 10.1016/j.dld.2017.07.002. [Epub ahead of print]


Ranking (2016): 35/134, Segundo cuartil

https://www.ncbi.nlm.nih.gov/pubmed/28780052


[74]

Please cite this article in press as: Saco A, et al. Validation of whole-slide imaging in the primary diagnosis of liver biopsies in a UniversityHospital. Dig Liver Dis (2017), http://dx.doi.org/10.1016/j.dld.2017.07.002

ARTICLE IN PRESSG ModelYDLD-3490; No. of Pages 7

Digestive and Liver Disease xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Digestive and Liver Disease

journa l homepage: www.e lsev ier .com/ locate /d ld

Liver, Pancreas and Biliary Tract

Validation of whole-slide imaging in the primary diagnosis of liverbiopsies in a University Hospital

Adela Saco a,1, Alba Diaz a,1, Monica Hernandez a, Daniel Martinez a, Carla Montironi a,Paola Castillo a,b, Natalia Rakislova a, Marta del Pino b,d, Antonio Martinez a,b,Jaume Ordi a,b,c,∗

a Department of Pathology, Hospital Clínic, Barcelona, Spainb ISGlobal, Barcelona Ctr. Int. Health Res. (CRESIB), Hospital Clínic – Universitat de Barcelona, Barcelona, Spainc University of Barcelona, School of Medicine, Barcelona, Spaind Institute of Gynecology, Obstetrics and Neonatology, Hospital Clínic – Institut dıInvestigacions Biomèdiques August Pi I Sunyer (IDIBAPS), Faculty ofMedicine, University of Barcelona, Spain

a r t i c l e i n f o

Article history:Received 30 January 2017Received in revised form 11 June 2017Accepted 11 July 2017Available online xxx

Keywords:Digital pathologyHepatic needle biopsiesIntra-observer agreementLiver pathology

a b s t r a c t

Background: Experience in the use of whole slide imaging (WSI) for primary diagnosis is limited and thereare no comprehensive reports evaluating this technology in liver biopsy specimens.Aims: To determine the accuracy of interpretation of WSI compared with conventional light microscopy(CLM) in the diagnosis of needle liver biopsies.Methods: Two experienced liver pathologists blindly analyzed 176 consecutive biopsies from the Pathol-ogy Department at the Hospital Clinic of Barcelona. One of the observers performed the initial evaluationwith CLM, and the second evaluation with WSI, whereas the second observer performed the first eval-uation with WSI and the second with CLM. All slides were digitized in a Ventana iScan HT at 400× andevaluated with the Virtuoso viewer (Roche diagnostics). We used kappa statistics (�) for two observations.Results: Intra-observer agreement between WSI and CLM evaluations was almost perfect (96.6%, � = 0.9;95% confidence interval [95% CI]: 0.9–1 for observer 1, and 90.3%, � = 0.9; 95%CI: 0.8-0.9 for observer 2).Both native and transplantation biopsies showed an almost perfect concordance in the diagnosis.Conclusion: Diagnosis of needle liver biopsy specimens using WSI is accurate. This technology can reliablybe introduced in routine diagnosis.

© 2017 Published by Elsevier Ltd on behalf of Editrice Gastroenterologica Italiana S.r.l.

1. Introduction

Conventional light microscopy (CLM) has been the basic and,until recently, the only tool for the histological diagnosis of biopsyspecimens. The development of the whole-slide imaging (WSI)technology has started to change this picture in the last few years.

The basis of the WSI technology is the use of high throughputscanners able to create high quality digital reproductions of glassslides containing a complete histological section and WSI viewersthat allow navigation across the virtual slide. These tools enable the

∗ Corresponding author at: Department of Pathology, Hospital Clínic – ISGlobal,Barcelona Ctr. Int. Health Res. (CRESIB), University of Barcelona Faculty of Medicine,Barcelona, Spain, C/Villarroel 170, 08036, Barcelona, Spain.

E-mail address: [email protected] (J. Ordi).1 These authors equally contributed to the work, and should share co-primary

authorship.

use of the computer as a CLM. WSI has many practical applicationsthat include education and teleconsultation [1–4]. In the last fewyears the medical community has shown increasing interest in theuse of WSI for routine primary diagnosis [5–7].

Indeed, routine pathological diagnosis can benefit from theadvantages of this technology. The WSI workstations are moreergonomic and facilitate a more efficient sign-out process. WSIallows viewing several slides at the same time on the same screen,which is particularly helpful for the evaluation of immuno- or his-tochemically stained slides that can be analyzed together withhematoxylin-eosin (H&E) staining (Fig. 1). The digital viewersincorporate tools that enable making annotations, rotating theimages and making precise measurements [8]. WSI has a muchlarger field of vision than CLM and a wider range of magnifica-tions, including very low magnifications that are very useful forthe evaluation of surgical specimens. WSI facilitates sharing imagesand information with clinicians and other pathologists. This is not

http://dx.doi.org/10.1016/j.dld.2017.07.0021590-8658/© 2017 Published by Elsevier Ltd on behalf of Editrice Gastroenterologica Italiana S.r.l.

dx.doi.org/10.1016/j.dld.2017.07.002

dx.doi.org/10.1016/j.dld.2017.07.002

http://www.sciencedirect.com/science/journal/15908658

http://www.elsevier.com/locate/dld

mailto:[email protected]

dx.doi.org/10.1016/j.dld.2017.07.002



2 A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx

Fig. 1. The WSI viewer may simultaneously show and synchronously move several slides of a case. This is particularly helpful in the evaluation of liver biopsy specimenssince it allows the analysis of an H&E stained slide together with histochemically and/or immunohistochemically stained slides.

only extremely useful in tumors boards, but also allows expert tele-consultation of difficult cases and frozen section intra-operativebiopsies [9,10]. Finally, with WSI algorithms can be used for theevaluation and quantification of immuhistochemical stains, result-ing in a more objective evaluation [11–14]. This tool is likely tobecome essential to achieve standardized diagnoses in the nearfuture.

Although WSI is considered to be comparable to CLM, adequatecorrelation between WSI and CLM diagnoses should be confirmedbefore this technology is used for primary diagnosis. The num-ber of studies aimed at validating WSI in the routine diagnosis ofthe different areas of pathology is rapidly growing [15]. However,whereas relatively abundant information is available in some areasof pathology, validation studies are very scant or even absent inother areas, such as liver biopsy. Indeed, while a few studies haveused this tool in research and automated image analysis [16–24],there is a complete absence of studies validating the use of WSI inneedle liver biopsies, which may lead to reluctance in implement-ing this technology in routine diagnosis.

2. Materials and methods

2.1. Characteristics of the institution

This study was performed at the Department of Pathology inthe Hospital Clinic (Barcelona, Spain). This department is com-posed of 16 pathologists, 8 residents and a variable number offellows. The specimens are divided into 14 subspecialties, and thepathologists limit their practice to one or two areas. In 2015 theDepartment handled 43,678 specimens with 11,081 paraffin blocks.The number of liver needle biopsy specimens during this year was230. The study was approved by the institutional ethics reviewboard/HCB/2014/0514.

2.2. Sample size calculation

The highest rate of discrepancy between the original diagnosisby CLM and that by WSI was calculated to be 3%, with a non-inferiority margin for WSI review of 5%. A 1-sided binomial testwas used for comparison at a level of significance of .05. The powerto be achieved was 70%, and the level of significance was .05. Basedon these assumptions, it was calculated that 100 cases would needto be reviewed to establish non-inferiority [25].

2.3. Specimens included in the study

All consecutive needle liver biopsy specimens received at theDepartment of Pathology of the Hospital Clinic in a 9-monthperiod (February–October 2015) and assigned to the same expertpathologist were included in the study (n = 176). This represented76.5% of the total number of liver biopsies evaluated in 2015. Allcases had a single paraffin block, containing one to five speci-mens (median 1). All specimens were routinely stained with H&E,Masson’s trichrome and reticulin stain. Additionally, immunohis-tochemical stains were used for specific cases after the request ofthe pathologist. The total number of scanned slides was 1286. Thebiopsies included both native and transplanted livers (n = 112 andn = 64, respectively). The median age of the patients was 57 years(range 18–91).

2.4. Scanning process and characteristics of the WSI display

All the needle liver biopsies were scanned daily after CLM diag-nosis. The scanning of the histological slides was performed on aVentana iScan HT (Ventana Medical Systems, Tucson, AZ, USA) ata magnification of 400x. The scanning process run automatically,and includes the selection of the area that contains the tissue, thedetermination of the focus points, the calibration, and the scanning.

dx.doi.org/10.1016/j.dld.2017.07.002



A. Saco et al. / Digestive and Liver Disease xxx (2017) xxx–xxx 3

When more than one section are mounted on a single slide, the sys-tem scans all the sections. No specific quality control of the slidesscanned was made prior to evaluation by the pathologist. The WSIproduced are stored in a dedicated mass storage environment andlinked to the pathology report, based on the recognition of a QRcode on the slide label. Although WSI can be accessed through thepathology laboratory information system (LIS) software (Novopath,Vitrosoft SL, Sevilla, Spain), for the purposes of this study access tothe WSI was made through the viewer.

The images were viewed with the Virtuoso viewer (Ventana),which works as a web browser and simulates a CLM. The imagesare shown using the same structure provided by the LIS. No specificsoftware installation is required to visualize the WSI. The scannedimages can be viewed up to a real magnification of 400× and upto 600× with a digital zoom, are always in focus, and have anoptimized contrast and adjusted illumination. The viewer showsa thumbnail of the whole slide, which allows confirmation that allthe material present on the glass slide has been included in the dig-ital image and helps in the navigation through the slide. The WSIare displayed on a 30′′ Coronis fusion MDC4130 monitor which hasa resolution of 4 Megapixels (Barco Electronic Systems, Barcelona,Spain).

2.5. CLM and WSI diagnosis

Two experts in liver pathology analyzed all cases. The firstobserver performed the initial evaluation with CLM, which wasconsidered the reference for diagnostic attribution, and the secondobservation with WSI, whereas the second observer performed theinitial evaluation with WSI and the second with CLM. In order toavoid interference with the first diagnosis, the minimum washoutperiod between the two observations was 1.5 months (range 1.5–4months). All the histological slides of each case were scanned andevaluated. A summary with the basic original clinical informationwas provided to the pathologist in both evaluations in order toemulate the real clinical environment, and the pathologist couldalso request additional information. When performing the WSI, thepathologist was blinded to the original diagnostic report, as well asto the evaluation made by the other pathologist. In all cases, a maindiagnosis was rendered in both evaluations. Additionally, in somespecimens additional secondary diagnoses were also provided.

2.6. Concordance between CLM and WSI diagnosis and featuresevaluated

An independent pathologist not involved with the evaluationcompared the original CLM and the WSI-based evaluations andjudged the concordance of the two diagnoses. Concordance wasclassified as: a) complete agreement; b) minor discrepancy (slightdifferences which would not have any clinical or prognostic impli-cations); and c) major discrepancy (differences with clinical and/orprognostic implications for the patient).

Some histological features were routinely evaluated in all thespecimens: portal fibrosis (using a 0–4 scale), presence or absenceof Mallory-Denk bodies, steatosis (using a 0–3 scale), and liver cellballooning [26,27]. Portal inflammation, cholangitis and endothe-litis were estimated using a 0–3 scale in all the acute rejectionbiopsies [28]. In the cases of cirrhosis and chronic hepatitis,necro-inflammatory activity (portal/periportal and lobular) wasevaluated with a 0–3 scale [29–32]. Other characteristics wererecorded when present.

2.7. Statistical analysis

The SPSS (SPSS IncTM140, Version 18, Chicago, IL, USA) wasused for statistical analyses. The results for categorical variables

are expressed as absolute numbers and percentages and 95% con-fidence intervals (95% CI). The Chi-squared or the Fisher’s exacttests were used to compare the variables. The results were eval-uated by unweighted Kappa statistics for two observations. Thismeasure calculates the degree of agreement in classification overthat which would be expected by chance and is scored as a num-ber between 0 and 1. According to the Landis-defined categoriesthe strength of agreement of the Kappa values (�) is: 0 none,beyond chance; 0–0.20 slight; 0.21–0.40 fair; 0.41–0.60 moder-ate; 0.61–0.80 substantial; 0.81–1.00 almost perfect. For the kappavalue calculation, the main diagnoses were grouped into nine cat-egories for the native livers and six categories for the transplantedlivers. The diagnostic categories for the native livers included: a)slight changes (including isolated steatosis), b) venous congestion,c) autoimmune diseases (autoimmune hepatitis and primary bil-iary cirrhosis), d) steatohepatitis, e) acute hepatitis, f) chronic viralhepatitis g) cirrhosis, h) malignant tumors (primary or metastatic),and i) other diseases. The diagnostic categories for the transplantedlivers included: a) slight changes; b) autoimmune hepatitis, c)steatohepatitis, d) hepatitis C virus infection, e) acute cellular rejec-tion, and f) chronic rejection.

3. Results

3.1. Intra-observer and inter-observer agreement

The overall intra-observer agreement between the CLM and theWSI diagnoses was 96.6% (� = 0.9; 95% CI: 0.9–1) for the observer 1and 90.3% (� = 0.9; 95% CI: 0.8–0.9) for the observer 2. There werefour minor discrepancies between the CLM and the WSI diagnosesfor observer 1 and 14 for observer 2. None of the discrepancieswere related to a poor quality of the WSI image or to insufficientmagnification. The diagnoses of carcinoma showed 100% concor-dance in all four evaluations. The overall inter-observer agreementbetween the CLM diagnoses performed by observer 1 and 2 was92.6% (� = 0.9; 95% CI: 0.9–1) and 96.6% (� = 1; 95% CI: 0.9–1) forthe WSI diagnoses.

The intra-observer agreement between CLM and WSI for thenative liver biopsies (n = 112) was 95.5% (� = 0.9; 95% CI: 0.9–1) forobserver 1 and 90.2% (� = 0.9; 95% CI: 0.8–0.9) for observer 2. Theinter-observer agreement between the CLM diagnoses performedby observer 1 and 2 in this group of native liver biopsies was 92.9%(� = 0.9; 95% CI: 0.9–1) and 87.5% (� = 0.9; 95% CI: 0.8–0.9) for theWSI diagnoses. Table 1 shows the intra-observer (WSI vs. CLM)for the two observers and the inter-observer agreement for CLMand WSI for each specific diagnostic group for the 112 native liverspecimens.

Table 2 shows the intra-observer (WSI vs. CLM) for the twoobservers and the inter-observer agreement for CLM and WSI foreach specific diagnostic group for the 64 transplantation biopsies.The inter-observer agreement between the CLM diagnoses per-formed by observer 1 and 2 in this group of transplanted liverbiopsies was 89.1% (� = 0.9; 95% CI: 0.8–1) for observer 1 and 87.5%(� = 0.8; 95% CI: 0.7–0.9) for observer 2. The inter-observer agree-ment between the CLM diagnoses performed by observer 1 and 2was in this group of transplanted liver biopsies was 89.1% (� = 0.8;95% CI: 0.7–1) and 87.5% (� = 0.8; 95% CI: 0.7–0.9) for the WSI diag-noses.

3.2. Agreement between WSI and CLM in relevant liver changes

Table 3 shows the intra-observer (WSI vs. CLM) and the inter-observer (WSI vs. WSI and CLM vs. CLM) agreement in theevaluation of relevant histological features in the native livers(steatosis, liver cell ballooning, presence or absence or Mallory-

dx.doi.org/10.1016/j.dld.2017.07.002




Table 1Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the diagnosis ofnative liver specimens (n = 112).

Diagnosis n Intra-observeragreement

Kappa value(95% CI)

Intra-observeragreement

Kappa value(95% CI)

Inter-observeragreement CLM

Kappa value(95% CI)

Inter-observeragreement

Kappa value(95% CI)

Observer 1 Observer 2 WSI

Mild changesa 22 98.3 0.9 (0.9–1) 94.9 0.9 (0.8–0.9) 96.0 0.9 (0.8–1) 96.0 0.9 (0.8–1)Venouscongestion

4 100 1 (NA) 98.9 0.9 (0.5–1) 98.3 0.7 (0.4–1) 99.4 0.9 (0.7–1)

Autoimmunediseasesb

15 99.4 1 (0.9–1.0) 98.3 0.9 (0.8–1) 98.9 0.9 (0.8–1) 97.7 0.9 (0.7–1)

Steatohepatitis 12 98.3 0.9 (0.8–1) 97.7 0.8 (0.7–1) 97.7 0.8 (0.7–1) 96.0 0.7 (0.6–0.9)Acutehepatitisc

7 99.4 0.9 (0.8–1) 98.3 0.8 (0.6–1) 99.4 0.9 (0.8–1) 98.3 0.8 (0.6–1)

Chronichepatitisd

14 97.7 0.9 (0.9–1) 94.9 0.8 (0.7–0.9) 97.7 0.9 (0.9–1) 93.7 0.8 (0.7–0.9)

Cirrhosis 19 100 1 (NA) 100 1 (NA) 98.9 0.9 (0.8–1) 98.9 0.9 (0.8–1)Tumorse 14 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)Otherf 5 100 1 (NA) 98.9 0.8 (0.6–1) 99.4 0.9 (0.8–1) 99.4 0.9 (0.8–1)

95% CI: 95% confidence interval.NA: not applicable.

a Includes mild to moderate macrovesicular steatosis (n = 15) and mild non-specific changes (n = 7).b Includes autoimmune hepatitis (7 cases) and primary biliary cirrhosis (8 cases).c Includes four toxic acute hepatitis, two acute B hepatitis and one hepatitis of unknown origin.d Includes 10 chronic hepatitis C, two chronic hepatits B and two drug induced hepatitis.e Includes one cholangiocarcinoma, six metastatic carcinomas, six hepatocellular carcinomas and one hemangioma.f Includes one case each of schistosomiasis, cystic fibrosis, graft versus host disease, sclerosing cholangitis and nodular regenerative hyperplasia.

Table 2Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the diagnosis ofliver transplantation biopsies (n = 64).

Diagnosis n Intra-observeragreement

Kappa value(95% CI)


Kappa value(95% CI)


Kappa value(95% CI)


Kappa value(95% CI)

Observer 1 Observer 2 CLM WSI

Mild changes 23 95.3 0.9 (0.8–1) 92.2 0.8 (0.7–1) 93.7 0.9 (0.7–1) 90.6 0.8 (0.6–1)Autoimmunehepatitis

2 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)

Steatohepatitis 1 100 1 (NA) 100 1 (NA) 98.4 7 (0.0–1) 98.4 0.7 (0.0–1)Chronichepatitisa

20 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)

Acute cellularrejection

15 95.3 0.9 (0.7–1) 89.1 0.7 (0.5–0.9) 90.6 0.7 (0.6–0.9) 93.7 0.8 (0.7–1)

Chronicrejection

1 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)

Other lesionsb 2 100 1 (NA) 100 1 (NA) 100 1 (NA) 100 1 (NA)

95%CI: 95% confidence interval; NA: not applicable.a Hepatitis C virus reinfection.b Includes one case of preservation injury and one insufficient biopsy.

Denk bodies, portal/peri-portal inflammatory activity and necrosisand lobular necrosis and inflammatory activity). Table 4 shows theintra-observer (WSI vs. CLM) and the inter-observer (WSI vs. WSIand CLM vs. CLM) agreement in the evaluation of major histologi-cal features in transplanted livers (portal inflammation, cholangitisand endothelitis).

4. Discussion

This is the first study evaluating the accuracy of WSI diagnosis inthe routine practice of needle liver biopsies. Our results show a highintra-observer concordance between the CLM and the WSI evalua-tions (over 90% for both obseervers) in the diagnoses of a large seriesof routine needle liver specimens. The kappa value, considered asa measure of the level of intra- and inter-observer agreement cor-rected by chance, was almost perfect (0.9 for both observers) and0.9–1 for the inter-oberver comparisons of the CLM and the WSIevaluations. The percentage of discrepancies between the CLM andWSI diagnoses observed in our study was below 10%, and onlyminor discrepancies were identified. Neither had an impact onpatient management. More importantly, none of the discrepancies

was related to a poor quality of the WSI image or to insufficientmagnification. All the discrepancies observed were associated witheither the small size of the material or to the intrinsic difficulty ofthe case. Thus, our results confirm that WSI may confidently beused for primary histological diagnosis of liver biopsies.

A number of studies have shown that there is a substantialvariation between and within observers in the evaluation of liverbiopsy specimens. These studies are limited to specific diseasessuch as non-alcoholic steatohepatitis and chronic viral hepatitis[33–37]. In an intra-observer concordance study including 50 biop-sies oriented as non-alcoholic steatohepatitis Kleiner et al. reporteda kappa value of 0.61 [34]. Our study showed a higher rate of con-cordance in the evaluation of steatohepatitis with a kappa valueranging from 0.7 to 0.9 in the different comparisons, althoughthe number of cases with this diagnosis was much lower andincluded both alcoholic and non-alcoholic steatohepatitis. Threestudies have evaluated intra-observer concordance in the diagno-sis of chronic viral hepatitis. The evaluation of fibrosis grade andstage in these studies showed kappa values ranging from 0.72 and1 [35–37], which were comparable with the concordance ratesobserved in our study (0.7–0.9). These discrepancies have mainlybeen attributed to the inherent intra-observer variability in the

dx.doi.org/10.1016/j.dld.2017.07.002




Table 3Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the evaluationof major histological changes in the native livers (n = 112).

Histologicalfeature


Kappa value(95% CI)


Kappa value(95% CI)


Kappa value(95% CI)


Kappa value(95% CI)


Fibrosisa 90.9 0.8 (0.8–0.9) 92.6 0.9 (0.8–0.9) 82.9 0.7 (0.6–0.9) 90.3 0.8 (0.8–0.9)Steatosisb 91.1 0.9 (0.8–0.9) 99.1 1 (NA) 92.9 0.9 (0.8–1) 92.0 0.9 (0.8–0.9)Liver cellballooningc

98.2 0.9 (0.9–1) 100 1 (NA) 96.4 0.9 (0.8–1) 98.2 0.9 (0.9–1)

Mallory-Denkbodiesc

98.2 0.9 (0.8–1) 99.5 1.0 (0.9–1.0) 97.3 0.9 (0.7–1) 100 1 (NA)

Portal/peri-portalinflammatoryactivity andnecrosis c,d

92.9 0.8 (0.6–0.9) 92.0 0.8 (0.7–0.9) 78.6 0.5 (0.3–0.6) 84.8 0.6 (0.5–0.8)

Lobularnecrosis andinflammatoryactivity c,d

96.4 0.9 (0.8–1) 93.7 0.9 (0.7–1.0) 87.5 0.7 (0.5–0.8) 86.6 0.7 (0.5–0.8)

95% CI: 95% confidence interval.a Graded on a 0–4 scale.b Graded on a 0–3 scale.c Evaluated as absent or present.d Portal/peri-portal inflammatory activity and necrosis and lobular necrosis and inflammatory activity were evaluated only in the cases with a diagnosis of cirrhosis (n = 19)

and chronic hepatitis (14 in native livers and 20 in transplanted livers).

Table 4Intra-observer (whole slide imaging [WSI] vs. conventional light microscopy [CLM]) for the two observers and Inter-observer agreement for CLM and WSI in the evaluationof the main histological features in the transplanted livers (n = 64).

Histologicalfeature


Kappa value(95% CI)


Kappa value(95% CI)


Kappa value(95% CI)


Kappa value(95% CI)


Portalinflammation

100 1 (NA) 90.6 0.8 (0.6–0.9) 85.9 0.7 (0.5–0.9) 95.3 0.9 (0.8–1)

Cholangitis 96.9 0.9 (0.8–1) 98.4 1 (0.9–1) 95.3 0.9 (0.7–1) 92.2 0.8 (0.6–1)Endothelitis 96.9 0.9 (0.8–1) 93.7 0.8 (0.8–1) 93.7 0.8 (0.7–1) 95.3 0.9 (0.7–1)

95% CI: 95% confidence interval; All features were graded on a 0–3 scale.

diagnosis of needle liver biopsy specimens. Interestingly, some ofthese studies analyzed a number of histological features separately,showing high concordance rates for steatosis (� = 0.79), periportalnecrosis (� = 0.74) and fibrosis (� = 0.86) and lower values for lobu-lar necrosis (� = 042) [33,34]. In the present study, the concordanceobserved for all these histological findings showed even betterresults. A possible limitation of our study is the lower number ofcases of each particular disease included in the analysis comparedto previous reports [33–37]. However, our study was designed toevaluate the reliability of the WSI tool for the diagnosis of any liverlesion and not specifically for a single disease.

Interestingly, the specific analysis of liver transplantation spec-imens (n = 64) showed a high intra-observer concordance thatremained almost perfect (93.7%; � = 0.9 for observer 1, 87.5, � = 0.8for observer 2). There were no differences in the diagnosis of rejec-tion.

The results obtained in our study with the liver biopsies arecomparable to other validation studies conducted in other areas ofpathology, such as breast [38], skin [39], gastrointestinal [40,41],prostate [42–46], gynecological [25], renal [46,47] or pediatricpathology [48,49] which show similar high rates of concordancebetween CLM and WSI diagnoses. Thus, the results of all thesestudies indicate that WSI should be considered as a validated tool,almost equivalent to the CLM. In keeping with this assumption,the guidelines and recommendations of the College of Ameri-can Pathologists, the Canadian Association of Pathologists and theAmerican Telemedicine Association for adequate validation of WSIbefore its use in routine diagnosis do not require a validation foreach specific area [1,4]. These recommendations indicate that only

60 samples per pathologist should be evaluated in order to ensurethe familiarity of the pathologist with the new tool. Indeed, as withany other tool, there is a learning curve for WSI [25,50–54].

Remarkably, the pathologist did not report any difficulty in ren-dering the diagnosis at the magnification used in this study (400×).A 200× magnification is considered as appropriate to achieve acorrect diagnosis in most previously published studies evaluatingother areas of pathology [25,42,48,50,55–57]. However, this scan-ning magnification may not be sufficient for some areas, such asthe liver due to the small size of the specimens and the need toevaluate subtle changes that frequently require the use of highmagnification.

The introduction of the WSI technology may significantlyimprove the diagnosis of routine needle liver biopsy specimens tak-ing into account the advantage of the possibility of viewing multipleslides at the same time with this technique. Indeed, this advan-tage can be very useful in liver pathology since several stains areoften used and WSI facilitates tele-consultation. Finally, the futuredevelopment of computer-assisted diagnostic algorithms is likelyto help reduce intra- and inter-observer variability. However, manyissues should be addressed to make this implementation feasibleand cost-efficient, such as the cost of the scanners [4,8,58–61], thecosts associated with the maintenance of the system and the stor-age of the images and legal issues related to the use of WSI forprimary diagnosis, including image storage and patient confiden-tiality. Approval is currently being sought from the US Food andDrug Administration (FDA) for the use of WSI in primary diagnosis.In the meantime, WSI is being increasingly used in several centersaround the world.

dx.doi.org/10.1016/j.dld.2017.07.002




In conclusion, the diagnosis of needle liver biopsies using WSIhas high intra-observer concordance with the results of CLM eval-uation. Our results confirm that WSI can be safely used for primaryhistological diagnosis of liver biopsies, including native and trans-plantation specimens.

Conflict of interest statementAll the authors have read and approved the manuscript, take publicresponsibility for its content, and consent the title, the authorship,and the contents of the paper. Neither the article nor any part of ithas been published or submitted elsewhere. There are no conflict ofinterest or commercial interests with this work. Should the paperbe accepted for publication, we authorize transferring the copyrightto Digestive and Liver Disease.

References

[1] Pantanowitz L, Sinard JH, Henricks WH, et al. Validating whole slide imagingfor diagnostic purposes in pathology: guideline from the College of AmericanPathologists Pathology and Laboratory Quality Center. Arch Pathol Lab Med2013;137:1710–22.

[2] Wilbur DC, Madi K, Colvin RB, et al. Whole-slide imaging digital pathology as aplatform for teleconsultation: a pilot study using paired subspecialist correla-tions. Arch Pathol Lab Med 2009;133:1949–53.

[3] Saco A, Bombi JA, Garcia A, et al. Current status of whole-slide imaging ineducation. Pathobiology 2016;83:79–88.

[4] Bernard C, Chandrakanth SA, Cornell IS, et al. Guidelines from the CanadianAssociation of Pathologists for establishing a telepathology service for anatomicpathology using whole-slide imaging. J Pathol Inform 2014;5:15.

[5] Al-Janabi S, Huisman A, Van Diest PJ. Digital pathology: current status andfuture perspectives. Histopathology 2012;61:1–9.

[6] Brachtel E, Yagi Y. Digital imaging in pathology–current applications and chal-lenges. J Biophotonics 2012;5:327–35.

[7] Pantanowitz L, Valenstein PN, Evans AJ, et al. Review of the current state ofwhole slide imaging in pathology. J Pathol Inform 2011;2:36.

[8] Thorstenson S, Molin J, Lundstrom C. Implementation of large-scale routinediagnostics using whole slide imaging in Sweden: digital pathology experi-ences 2006-2013. J Pathol Inform 2014;5:14.

[9] Hartman DJ, Parwani AV, Cable B, et al. Pocket pathologist: a mobile applicationfor rapid diagnostic surgical pathology consultation. J Pathol Inform 2014;5:10.

[10] Speiser JJ, Hughes I, Mehta V, et al. Mobile teledermatopathology: using a tabletPC as a novel and cost-efficient method to remotely diagnose dermatopathol-ogy cases. Am J Dermatopathol 2014;36:54–7.

[11] Gavrielides MA, Conway C, O’Flaherty N, et al. Observer performance in theuse of digital and optical microscopy for the interpretation of tissue-basedbiomarkers. Anal Cell Pathol (Amst) 2014;2014:157308.

[12] Nassar A, Cohen C, Agersborg SS, et al. A multisite performance study comparingthe reading of immunohistochemical slides on a computer monitor with con-ventional manual microscopy for estrogen and progesterone receptor analysis.Am J Clin Pathol 2011;135:461–7.

[13] Micsik T, Kiszler G, Szabo D, et al. Computer aided semi-automated evaluationof HER2 immunodetection–a robust solution for supporting the accuracy ofanti HER2 therapy. Pathol Oncol Res 2015;21:1005–11.

[14] Krenacs T, Zsakovics I, Diczhazi C, et al. The potential of digital microscopy inbreast pathology. Pathol Oncol Res 2009;15:55–8.

[15] Saco A, Ramirez J, Rakislova N, et al. Validation of whole-slide imaging forhistolopathogical diagnosis: current state. Pathobiology 2016;83:89–98.

[16] Atupelage C, Nagahashi H, Kimura F, et al. Computational hepatocellular car-cinoma tumor grading based on cell nuclei classification. J Med Imaging(Bellingham, Wash) 2014;1:34501.

[17] Bejnordi BE, Litjens G, Timofeeva N, et al. Stain specific standardiza-tion of whole-slide histopathological images. IEEE Trans Med Imaging2016;35:404–15.

[18] Isse K, Grama K, Abbott IM, et al. Adding value to liver (and allograft) biopsyevaluation using a combination of multiplex quantum dot immunostaining,high-resolution whole-slide digital imaging, and automated image analysis.Clin Liver Dis 2010;14:669–85.

[19] Abe T, Hashiguchi A, Yamazaki K, et al. Quantification of collagen and elas-tic fibers using whole-slide images of liver biopsy specimens. Pathol Int2013;63:305–10.

[20] Liang Y, Wang F, Treanor D, et al. A framework for 3D vessel analysis using wholeslide images of liver tissue sections. Int J Comput Biol Drug Des 2016;9:102–19.

[21] Liang Y, Wang F, Treanor D, et al. Liver whole slide image analysis for 3D vesselreconstruction. Proc IEEE Int Symp Biomed Imaging 2015;2015:182–5.

[22] Nagase A, Takahashi M, Nakano M. Automatic calculation and visualization ofnuclear density in whole slide images of hepatic histological sections. BiomedMater Eng 2015;26(Suppl. 1):S1335–44.

[23] Rawlins SR, El-Zammar O, Zinkievich JM, et al. Digital quantification is moreprecise than traditional semiquantitation of hepatic steatosis: correlation withfibrosis in 220 treatment-naive patients with chronic hepatitis C. Dig Dis Sci2010;55:2049–57.

[24] Hall AR, Dhillon AP, Green AC, et al. Hepatic steatosis estimated microscopicallyversus digital image analysis. Liver Int 2013;33:926–35.

[25] Ordi J, Castillo P, Saco A, et al. Validation of whole slide imaging in the primarydiagnosis of gynaecological pathology in a University Hospital. J Clin Pathol2015;68:33–9.

[26] Scheuer PJ. Classification of chronic viral hepatitis: a need for reassessment. JHepatol 1991;13:372–4.

[27] Brunt EM, Janney CG, Di Bisceglie AM, et al. Nonalcoholic steatohepatitis: aproposal for grading and staging the histological lesions. Am J Gastroenterol1999;94:2467–74.

[28] Demetris AJ, Batts KP, Dhilion AP, et al. Banff schema for grading liver allograftrejection: an international consensus document. Hepatology 1997;25:658–63.

[29] Brunt EM. Grading and staging the histopathological lesions of chronichepatitis: the Knodell histology activity index and beyond. Hepatology2000;31:241–6.

[30] Batts KP, Ludwig J. Chronic hepatitis. An update on terminology and reporting.Am J Surg Pathol 1995;19:1409–17.

[31] Desmet VJ, Knodell RG, Ishak KG, et al. Formulation and application of a numer-ical scoring system for assessing histological activity in asymptomatic chronicactive hepatitis [Hepatology 1981;1:431-435]. J Hepatol 2003;38:382–6.

[32] Hubscher SG. Histological grading and staging in chronic hepatitis: clinicalapplications and problems. J Hepatol 1998;29:1015–22.

[33] Rousselet M-C, Michalak S, Dupre F, et al. Sources of variability in histologicalscoring of chronic viral hepatitis. Hepatology 2005;41:257–64.

[34] Kleiner DE, Brunt EM, Van Natta M, et al. Design and validation of a his-tological scoring system for nonalcoholic fatty liver disease. Hepatology2005;41:1313–21.

[35] Regev A, Berho M, Jeffers LJ, et al. Sampling error and intraobserver varia-tion in liver biopsy in patients with chronic HCV infection. Am J Gastroenterol2002;97:2614–8.

[36] Robert M, Sofair AN, Thomas A, et al. A comparison of hepatopathologists’ andcommunity pathologists’ review of liver biopsy specimens from patients withhepatitis C. Clin Gastroenterol Hepatol 2009;7:335–8.

[37] Skripenova S, Trainer TD, Krawitt EL, et al. Variability of grade and stage insimultaneous paired liver biopsies in patients with hepatitis C. J Clin Pathol2007;60:321–4.

[38] Al-Janabi S, Huisman A, Willems SM, et al. Digital slide images for primary diag-nostics in breast pathology: a feasibility study. Hum Pathol 2012;43:2318–25.

[39] Al Habeeb A, Evans A, Ghazarian D. Virtual microscopy using whole-slide imag-ing as an enabler for teledermatopathology: a paired consultant validationstudy. J Pathol Inform 2012;3:2.

[40] Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primary diag-nostics of gastrointestinal tract pathology: a feasibility study. Hum Pathol2012;43:702–7.

[41] Molnar B, Berczi L, Diczhazy C, et al. Digital slide and virtual microscopy basedroutine and telepathology evaluation of routine gastrointestinal biopsy speci-mens. J Clin Pathol 2003;56:433–8.

[42] Camparo P, Egevad L, Algaba F, et al. Utility of whole slide imaging and virtualmicroscopy in prostate pathology. APMIS 2012;120:298–304.

[43] Chargari C, Comperat E, Magne N, et al. Prostate needle biopsy examination bymeans of virtual microscopy. Pathol Res Pract 2011;207:366–9.

[44] Fine JL, Grzybicki DM, Silowash R, et al. Evaluation of whole slide imageimmunohistochemistry interpretation in challenging prostate needle biopsies.Hum Pathol 2008;39:564–72.

[45] Helin H, Lundin M, Lundin J, et al. Web-based virtual microscopy in teachingand standardizing Gleason grading. Hum Pathol 2005;36:381–6.

[46] Al-Janabi S, Huisman A, Jonges GN, et al. Whole slide images for primarydiagnostics of urinary system pathology: a feasibility study. J Ren Inj Prev2014;3:91–6.

[47] Furness P. A randomized controlled trial of the diagnostic accuracy of internet-based telepathology compared with conventional microscopy. Histopathology2007;50:266–73.

[48] Al-Janabi S, Huisman A, Nikkels PGJ, et al. Whole slide images for primarydiagnostics of paediatric pathology specimens: a feasibility study. J Clin Pathol2013;66:218–23.

[49] Arnold MA, Chenever E, Baker PB, et al. The College of American Patholo-gists guidelines for whole slide imaging validation are feasible for pediatricpathology: a pediatric pathology practice experience. Pediatr Dev Pathol2015;18:109–16.

[50] Al-Janabi S, Huisman A, Vink A, et al. Whole slide images for primary diagnosticsin dermatopathology: a feasibility study. J Clin Pathol 2012;65:152–8.

[51] Randell R, Ruddle RA, Mello-Thoms C, et al. Virtual reality microscope versusconventional microscope regarding time to diagnosis: an experimental study.Histopathology 2013;62:351–8.

[52] Krishnamurthy S, Mathews K, McClure S, et al. Multi-institutional compari-son of whole slide digital imaging and optical microscopy for interpretationof hematoxylin-eosin-stained breast tissue sections. Arch Pathol Lab Med2013;137:1733–9.

[53] Houghton JP, Ervine AJ, Kenny SL, et al. Concordance between digital pathologyand light microscopy in general surgical pathology: a pilot study of 100 cases.J Clin Pathol 2014;67:1052–5.

[54] Randell R, Ruddle RA, Thomas RG, et al. Diagnosis of major cancer resectionspecimens with virtual slides: impact of a novel digital pathology workstation.Hum Pathol 2014;45:2101–6.

[55] Bauer TW, Schoenfield L, Slaw RJ, et al. Validation of whole slide imaging for pri-mary diagnosis in surgical pathology. Arch Pathol Lab Med 2013;137:518–24.

dx.doi.org/10.1016/j.dld.2017.07.002

http://refhub.elsevier.com/S1590-8658(17)30977-5/sbref0005
































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































[56] Gilbertson JR, Ho J, Anthony L, et al. Primary histologic diagnosis using auto-mated whole slide imaging: a validation study. BMC Clin Pathol 2006;6:4.

[57] Bauer TW, Slaw RJ. Validating whole-slide imaging for consultation diagnosesin surgical pathology. Arch Pathol Lab Med 2014;138:1459–65.

[58] Hedvat CV. Digital microscopy: past, present, and future. Arch Pathol Lab Med2010;134:1666–70.

[59] Ho J, Ahlers SM, Stratman C, et al. Can digital pathology result in cost savings? Afinancial projection for digital pathology implementation at a large integratedhealth care organization. J Pathol Inform 2014;5:33.

[60] Isaacs M, Lennerz JK, Yates S, et al. Implementation of whole slide imaging insurgical pathology: a value added approach. J Pathol Inform 2011;2:39.

[61] Pantanowitz L. Digital images and the future of digital pathology. J PatholInform 2010;1.

dx.doi.org/10.1016/j.dld.2017.07.002




































































































































[82]


[83]

Estudio número 4

“Current Status of Whole-Slide Imaging in Education”

Adela Saco, Josep Antoni Bombi, Adriana Garcia, José Ramirez, Jaume Ordi





[84]

E-Mail [email protected]

Original Paper

Pathobiology 2016;83:79–88 DOI: 10.1159/000442391

Current Status of Whole-Slide Imaging in Education

Adela Saco a Jose Antoni Bombi a Adriana Garcia a Jose Ramírez a

Jaume Ordi a, b

a Department of Pathology, Hospital Clínic, University of Barcelona School of Medicine, and b ISGlobal, Barcelona Center for International Health Research (CRESIB), Barcelona , Spain

shown to be an extremely useful tool for undergraduate ed-ucation (medical, dental and veterinary schools), for the training of residents of pathology, tele-education and in tu-mor boards. © 2016 S. Karger AG, Basel

Introduction and Historical Perspective

Basic skills in histology and pathology are an essential component of the education of undergraduate students, not only at medical schools, but also at schools of den-tistry, veterinary medicine and biology. Residents and fel-lows in pathology demand more advanced training, and certified pathologists require continuing education. This training has to be provided more and more often to par-ticipants located in distant sites. Until a few years ago, the only tool that fulfilled these needs was conventional light microscopy (CLM). CLM was introduced as a diagnostic and teaching tool nearly two centuries ago and became the basis for teaching histology and pathology [1, 2] . However, CLM has many limitations, not only in con-ducting the classes but also in assessing the skills of the students. One of the main disadvantages of using CLM is that it does not allow simultaneous viewing of the slides

Key Words

Medical education · Teaching · Virtual microscopy · Whole-slide imaging

Abstract

Conventional light microscopy (CLM) has classically been the basic tool to teach histology and pathology. In recent years, whole-slide imaging (WSI), which consists of generat-ing a high-magnification digital image of an entire histolog-ical glass slide, has emerged as a useful alternative to CLM offering a myriad of opportunities for education. Navigation through the digitized slides closely simulates viewing glass slides with a microscope and is also referred to as virtual mi-croscopy. WSI has many advantages for education. Students feel more comfortable with its use, and it can be used in any classroom as it only requires a computer with Internet access and it allows remote access from anywhere and from any device. WSI can be used simultaneously by a large number of people, stimulating cooperation between students and improving the interaction with the teachers. It allows mak-ing marks and annotations on specific fields, which enable specific directed questions to the teacher. Finally, WSI sup-ports are cost-effective compared with CLM. Consequently, WSI has begun to replace CLM in many institutions. WSI has

Published online: April 26, 2016

Jaume Ordi Department of Pathology, Hospital Clínic University of Barcelona, C/Villarroel 170 ES–08036 Barcelona (Spain) E-Mail jordi @ clinic.ub.es

© 2016 S. Karger AG, Basel1015–2008/16/0833–0079$39.50/0

www.karger.com/pat

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M

Saco/Bombi/Garcia/Ramírez/Ordi


80

by multiple students. Secondly, rooms with multiple mi-croscopes are expensive and require continuous mainte-nance. A third important disadvantage of CLM is the need to generate glass slides and store them, which entails significant economic costs and the loss of specimen mate-rial.

In the early 20th century, the emergence of projectors facilitated the training of students because it allowed dis-playing the histological images to a number of partici-pants at the same time. The introduction and develop-ment of initially analogic and later on digital video cam-eras boosted the use of CLM in multiple teaching settings. However, these devices allowed moving the slide only to one person and, consequently, served as a teaching sup-plement but did not allow the replacement of CLM.

In the 1980s, the first digital images were generated from histological slides, although it was not until later, with the advent of personal computers with sufficient memory capacity, that digital microscopy progressed rapidly to the technology we know today. A few years lat-er, imaging converter programs and servers appeared which allowed uploading the virtual slides to the web and permitted viewing the images and zooming [1] .

Whole slide imaging (WSI), also referred to as virtual or digital pathology, consists of generating a virtual image of the entire histological glass slide. The process is per-formed by WSI scanners, robotic microscopes capable to automatically generate digital images from the glass slide. Current scanners are able to use different optical objec-tives (×10, ×20, ×40 and ×60) depending on the specific needs and allow scanning the whole slide or a particular portion of the tissue. Specific software allows viewing dig-ital slides online from computers without the need for a CLM. At present, there are numerous systems prepared to generate good-quality virtual images of histological sections obtained from paraffin blocks or frozen tissue as well as from cytological smears. Image browsing is per-formed with a mouse or a joystick, which allows moving through the different areas of the slide and permits image zooming, thus simulating the optical objectives of a CLM

[3] . Software viewers have multiple tools to make mea-surements and annotations on the images, and may con-tain additional patient information, thus providing a complete view of the cases [4] . The information is ade-quately protected because, although students and resi-dents can access the virtual slides from any computer, ac-cess is controlled by the use of a specific password and registration in the system.

WSI is an extremely useful teaching tool because it al-lows displaying the histological slides on a computer monitor and solves many of the problems of CLM. In-deed, WSI has been used for many different educational activities ( table 1 ). This technology is rapidly expanding as a teaching tool, either as a complement to or a substi-tute of CLM. In many centers, the transition from CLM to WSI has already occurred either gradually or suddenly. This review focuses on the applications of WSI (or vir-tual microscopy) in education.

Advantages of WSI

WSI has brought about an important change in the way of understanding teaching in histology and pathol-ogy, allowing the introduction of some actions and ca-pacities that were previously not possible. The main ad-vantages of WSI are summarized in table 2 .

Some of the advantages of WSI are related to the change in the physical tool itself. WSI allows any com-puter to work as a CLM, and consequently reduces or eliminates the need for conventional microscopes. It has been clearly shown that students feel more comfortable with the use of WSI because they all have prior knowl-edge of computers (almost all students are currently dig-ital natives) [5–10] . An additional advantage of WSI is that the slides are always in focus. Thus, the students get used to the tool very quickly and can immediately con-centrate on the histological features of the slides and do not have to become acquainted with the microscope [6, 9, 10] . WSI can be used in any classroom, because it only requires a computer with Internet access [1, 11] . More-over, WSI allows access from any device, either in or outside the facilities of the institution. This implies that the students can review the histological slides at any time and from anywhere, thereby facilitating their study and eliminating the restrictions of access to the labora-tory of microscopy after class hours [1, 6, 10, 12–14] . As a consequence, laboratories of microscopy, which are very expensive and costly in maintenance, become un-necessary, leading to a significant reduction in costs [9,

Table 1. Major educational uses of WSI

Undergraduate teaching (medicine, veterinary medicine, biologyand dentistry)

Pathology training (residency and fellowship programs)Schools of cytotechnologyTumor boardsTele-education, e-learning and virtual workshops

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M

WSI in Education Pathobiology 2016;83:79–88DOI: 10.1159/000442391

81

15] . With an appropriate bandwidth, virtual micro-scopes can be used simultaneously by a large number of people, thereby surpassing the restrictions related to a limited number of CLMs in a laboratory [4, 9, 12, 16, 17] and stimulating cooperation between students [9, 12, 15, 17–19] . Moreover, the interaction between teachers and students is improved by viewing the same image at the same time in the classroom, which allows bringing up questions easily and improves the learning curve [15, 18, 19] .

The second group of advantages of WSI is related to the characteristics of the viewer. The presence of a thumb-nail indicating the area shown on the screen ( fig. 1 ) pro-motes better orientation when the student browses the slide [12] . Digital viewers allow seeing the slide at a very low magnification, which also helps the student to be ori-entated in the tissue. Teaching tools based on WSI allow completing the information of the histological slides with clinical data, imaging studies (conventional radiology, CT, ultrasound and MRI), macroscopic images, as well as

Table 2. Advantages of WSI for teaching purposes

Related to the equipmentStudents feel more comfortable with the use of the virtual microscope because they have prior computer knowledgeIt can be used in any classroom: it only requires a computer with Internet accessIt allows remote access anywhere and from any deviceIt can be simultaneously used by an unlimited number of peopleSeveral students can use the same computer at the same time, stimulating cooperation between studentsIt improves the interaction between teachers and students by viewing the same image at the same timeEliminates the need for investing in the creation and maintenance of microscopic laboratories

Related to the viewerThe thumbnail in the viewer facilitates better orientation when the students browse through the slideThe information of the histological slide can be completed with macroscopic images, immunohistochemical stains, radiologicalimages and clinical dataSeveral slides can be displayed simultaneously on the same screen, facilitating the interpretation of immunohistochemical techniquesMarks and annotations can be made on specific fields, which facilitate specific directed questions to the teacher

Related to the digital slidesNo deterioration in digital slides with time and slides do not have to be replacedHomogeneity in the quality of images available to studentsAdditional histological sections are not neededCases with scant tissue or consultation material can also be studiedFISH or immunofluorescence digital images do not lose fluorescenceOld cases are immediately accessible, without the help of technical staff searching for the files in the archive

Fig. 1. The viewer displays a thumbnail that indicates the area shown on the screen. This promotes better orientation when the student browses the slide.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M



82

histochemical or immunohistochemical stains ( fig. 2 ), which, in turn, allow a more complete picture of the study cases [16, 17] . Many viewers allow simultaneously dis-playing several slides on the screen, thus facilitating the interpretation of immunohistochemical techniques by comparing them with the conventional hematoxylin-eo-sin stain ( fig. 3 ) [1] . Another feature of WSI viewers that is particularly helpful for teaching purposes is the possi-bility to make marks and annotations on specific fields of the digital slides to show the key features that allow the recognition of the lesions. This tool facilitates solving spe-cific doubts and improves the interaction between the teacher and the student. With the annotations made by the teacher, the students can easily identify the key areas on a slide, allowing them to focus on the diagnostic clues of a specific lesion. Annotations and marks made by the students allow them to ask the teacher about specific doubts. Indeed, annotations seem to improve the final results of the students, and several studies have shown

that the students who had notes on the slides have better scores than those who do not make any annotation [1, 12, 14, 20–22] .

The third set of advantages of WSI is related to the his-tological slides. Digital slides always have the same qual-ity. They never break, get lost or deteriorate with time. Thus, it is not necessary to replace them by recutting and staining new slides every certain period of time, with the subsequent impact on costs and preservation of the fre-quently highly valuable tissue from the paraffin block [17, 20] . WSI homogenizes the material available to students. All students have exactly the same digital slide, which eliminates the variability in quality among different glass slides. Creating selections of interesting cases to students becomes extremely easy because there is no need for ad-ditional histological sections [10, 17, 22] . The easiness to prepare collections of cases with a high number of slides from each organ or pathology allows randomly showing many different cases to the students. This makes the stu-

Fig. 2. Teaching tools based on WSI allow completing the information of the histological slides with clinical in-formation, imaging studies (conventional radiology, CT, ultrasound and MRI) macroscopic images, as well as histochemical or immunohistochemical stains.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M


83

dent learn to recognize the histological characteristics of the organ or the lesion, and not to remember the slide because of other features, such as the size or shape of the tissue [17] . These features enhance learning and are espe-cially useful during the evaluation test because the results are totally dependent on the knowledge of the student [23] .

Using WSI, cases with very scant material and consul-tation cases can be investigated, even when the material has to be returned to the referral center. WSI enables im-mediate accessibility to old cases, without the need for technical staff to retrieve the slides or paraffin blocks from the archive. WSI also allows using cytological im-ages while preserving the original slide [22] . Finally, WSI permits introducing FISH or immunofluorescence slides, which previously had to be shown on static images be-cause of the deterioration and loss of fluorescence.

Weaknesses of WSI

The main disadvantage of WSI is the initial economic investment for the acquisition of the scanner and the complete WSI system [1, 12] . The high-resolution WSI images are associated with files of very large size, which means significant needs in terms of disk memory for their storage. Thus, high-capacity servers to store and distrib-ute the information are needed, and regular maintenance of the servers is required. Proper functioning of the sys-tem requires high-speed Internet connection [4] .

However, the tool can be used for different subjects and, thus, be shared by different departments at the school (histology, pathology or medical specialties). A possible

solution to this problem is the use of scanners already working at the institution for pathological diagnosis and using free or low-cost software to view the virtual slides [1] . Several vendors offer renting systems that include the digitalization of a set (or a few sets) of slides, the use of teaching software and a number of terabytes in a server. This option eliminates the investment in the equipment and reduces the costs associated with servers. In any case, computer classrooms are much more versatile than labo-ratories of microscopy, and all the initial economic disad-vantages of WSI are rewarded with time, because the in-vestment in computer equipment and maintenance is much less than that of the costs associated with a labora-tory with optical microscopes [9, 15] .

Another possible disadvantage is that in the centers that only use WSI, the students do not learn how to use a conventional light microscope. Nonetheless, this is a mi-nor problem since most medical students almost never use the microscope after finishing their training. Indeed, it is more important for them to recognize histological patterns and lesions than to learn how to use the tool [8, 15, 24] .

Finally, the use of these computer-based tools may re-sult in a dramatic reduction in the personal contact be-tween teachers and students. Actions promoting face-to-face meetings should be considered to avoid the deperson-alization associated with the spread in the use of computers. On the other hand, periodic quality control of WSI is high-ly recommended. This should include a comparison be-tween virtual and glass slides, an evaluation of the use of the system outside the course schedule and the registration of what virtual images and which areas in these virtual im-ages are most often seen by the students [1] .

Fig. 3. The viewer allows displaying simul-taneously several slides on the screen, thus facilitating the analysis and interpretation of immunohistochemical techniques.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M



84

WSI in Undergraduate Teaching

The number of centers using WSI in undergraduate teaching, either as a complement to or as a substitute of CLM, has markedly increased over the last few years. Many studies have shown excellent results with WSI, not only for medical students but also for students of biology, dentistry, parasitology and veterinary medicine [2, 7, 8, 23, 25–30] . Many of these centers have evaluated the

opinion of students after the introduction of WSI, and all have reported very positive feedback [9, 26, 30, 31] . One of the most valued advantages is the improved accessibil-ity to the slides with WSI, allowing the student to access the slides at any time and from any place [10, 14, 19] . Sev-eral studies have shown that this is one of the most ap-preciated features of WSI, and this was the most prized feature in a study conducted at our institution [10] . In-deed, data obtained from the audit of accesses to the nav-

400

350

300

250

200

150

100

50

0

Acce

sses

(n)

140

120

100

80

60

40

20

00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Acce

sses

(n)

–30 –28 –19 –15 –13 –11 –9 –7 –5 –3 –1 ExamDay

Time of day (h)

a

b

Fig. 4. a Number of accesses to the virtual slides from the opening of the website to the day of the examination. The x-axis shows the day slides were accessed in rela-tion to the examination (day 0). The y-axis shows the absolute number of accesses. Red columns identify accesses on week-ends/holidays (Saturday, Sunday or other holidays). Blue columns indicate work-days. b Time virtual slides were accessed during the day. The x-axis shows the time of the day and the y-axis shows the absolute number of accesses.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M


85

igator at our institution showed that over half of the ac-cesses are made on holidays and over one third after working hours ( fig. 4 ).

Students consider the use of WSI to be easier than that of CLM. WSI allows them to concentrate on the tissues and lesions and not on handling the microscope [8, 15, 32] . All studies agree that WSI improves collaboration between students and self-learning [6, 9, 16, 24, 33, 34] . Teachers also positively evaluate WSI because, although there is initially a significant increase in the time related to the preparation of the material, in the end it results in a significant reduction in the time spent in the prepara-tion of the lessons [19, 35] . WSI provides a more complete approach to the cases by adding clinical information, ra-diological imaging, macroscopic images, immunohisto-chemical stains and molecular data to the virtual histo-logical images [33, 36] . This can more accurately simulate the actual diagnostic practice and seems to be associated with better final results. Most studies agree that the image quality is better with WSI than with CLM, because the microscopes used for undergraduate teaching are, in gen-eral, of poor quality [20] . Images generated at a ×200 magnification are of sufficient quality for undergraduate teaching.

The negative evaluations are basically related to tech-nical aspects, such as the speed in loading the images or compatibility problems with some computer models. A high-speed Internet connection and a server with suffi-cient capacity are required, especially when many stu-dents view slides at the same time [12] . Some centers have solved the problem of storage space by giving the virtual images on a DVD support or scanning only the portion of tissue they want to view, thus saving space in relation with the whole slide [37] . There is still some controversy about whether WSI can totally replace CLM, because stu-dents using WSI do not learn how to handle the CLM. The students’ opinions about this issue are divided. Most studies reveal that the teachers think that it is more im-portant for the students to use their time in learning his-tology or pathology than in getting familiar with the tech-nical aspects of the management of the microscope [2] .

Finally, the evaluation process using WSI offers choos-ing between numerous slides and allows the homogeniza-tion of the test, because all the students can visualize the same slide. In addition, in contrast to CLM, only the his-topathological knowledge of the student is evaluated us-ing WSI, obviating the influence of the students’ ability to manage the microscope [32] . Students positively evaluate the use of WSI in the tests as long as the practices are car-ried out in the same way [30] . Indeed, it is strongly rec-

ommended that the same system be used to perform the practical lessons and the examinations. There is major reluctance to completely abandon CLM when lessons are taken with WSI and the tests are performed with glass slides. In contrast, if WSI is also used for the evaluation, the students do not feel the need to use CLM during the course [30] . Comparing the final results of students using WSI with those using CLM (either in the same year or in previous years), either no differences have been observed or results were better among those using WSI. Moreover, students using WSI seem to recognize histological pat-terns better [5, 31, 38, 39] .

In summary, studies assessing the use of WSI by stu-dents have shown very positive results. WSI does not seem to affect the final knowledge or may even improve it [10, 40–42] .

WSI in Postgraduate Training (Residency and

Fellowship Programs)

One of the first uses of WSI was to create series of in-teresting cases for residents or fellows in pathology and other specialties that require histological recognition of normal and pathological tissues [17] . Preparing virtual slides prevents the loss of biopsy material in making sec-tions for teaching, thus making a collection of interesting cases much easier. Many residents can view the same im-age at the same time, with the possibility of working out-side the center facilities and at any hour. These features provide an optimization of time and generate more com-fort for residents and fellows.

The presence of annotations on the virtual slides also represents a major improvement because it facilitates learning. In a study comparing results between residents of dermatology and pathology who visualized the same virtual images supplemented with or without annotations made by teachers, residents whose preparations had an-notations showed a better score, because the learning was more directed towards the most characteristic histologi-cal changes that led to the correct diagnosis [21] .

WSI has also successfully been used to measure the learning level reached by residents. Several initiatives, such as the European Association of Pathology Chairs and Residency Program Directors, aim to homogenize the knowledge of residents all over Europe evaluating them by performing a test containing virtual slides [17] . WSI is well evaluated by residents in the studies, although the final results comparing abilities reached with WSI and CLM differ from one study to another [43–45] . The

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M



86

presence of clinical and radiological information and previous practice using WSI seem to influence the results [36] .

Some studies point out the relevance of establishing the steps involved in the diagnostic process, stressing that residents should learn not only what the correct diagnosis is, but also what the logical sequence to achieve this diag-nosis is. These studies use eye-tracking cameras that pick up eye movement during the diagnostic process. The dif-ference between senior pathologists and residents are es-tablished by examining which fields are more often viewed, the time spent in these fields and the order in which the slide is observed. Residents spend much more time evaluating the slides than experienced pathologists. However, experienced pathologists are slower in choos-ing the field in where they will zoom and they do this more frequently in areas outside central (foveal) vision than residents. Thus, there are two kinds of slide evalua-tion: one more dispersed and time-consuming and an-other more targeted and effective [17, 46] .

Although WSI is currently mainly a complement to CLM, in the future it is expected to almost completely re-place CLM as the use of WSI in routine diagnosis expands in laboratories of pathology. Studies to determine the im-pact of training pathologists without exposure to actual glass slides are warranted.

WSI in Schools of Cytopathology

The homogeneity in the quality of the images dis-played to the students is a highly appreciated feature of WSI and represents a major advantage in cytopathology, because it allows many students to use the same slide si-multaneously without the risk of breaking or losing the slide. In addition, annotated digitized slides are very use-ful for teaching cytomorphology to cytotechnologists. The use of annotations for evaluation has excellent re-sults, allowing comparison between expert cytologists and students [47] .

However, the implications of the relatively low resolu-tion of some WSI systems at low (screening) magnifica-tion still need to be solved. Moreover, viewing through a CLM may provide a different perceived field width than what is seen on a monitor. WSI has more difficulty in gen-erating appropriate images due to the difficulties in focus-ing the images at different levels [48] . This problem can be solved with the use of software packages that allow focusing the different levels of the virtual slide thereby more closely simulating the daily practice of cytology.

In conclusion, although WSI has significant educa-tional advantages, a number of technical problems should be solved before it can be confidently used to teach cytol-ogy.

Tumor Boards

Many hospitals have tumor boards where clinicians meet for multidisciplinary case presentations. Patholo-gists are often required to present the pathology findings at the board presentations. WSI is currently successfully used for this purpose at several institutional tumor boards [49, 50] . The use of WSI in tumor boards and interdisci-plinary sessions also helps to bring histology knowledge to residents of medical or surgical specialties who do not have much contact with the laboratory [22] .

Tele-Education and e-Learning

The use of e-learning and tele-education is expanding extraordinarily because it allows providing continuous medical education in a practical and convenient way, markedly reducing costs. Virtual workshops avoid the need to travel to meetings, providing significant econom-ic savings and more flexibility. WSI allows easy visualiza-tion of the cases and eliminates the requirement of send-ing the glass slides, saving time and the costs of courier service. Moreover, the system avoids the risks of loss of or damage to the glass slides [22] . Different online services permit the access to online teaching with the objective of sharing virtual slides (e.g. PathXchange, vMic Pathorama or Slide2Go) [51] .

The use of this technology is markedly spreading, and numerous institutions promote web-based learning. The United States and Canadian Academy of Pathology (USCAP), the American Society of Cytopathology and the International Academy of Cytopathology, among others, have virtual atlases that include numerous cases with educational purposes [52] . Currently, there are sev-eral collections of cases aimed at teaching residents such as the Dr. Juan Rosai collection (www.rosaicollection.org; approx. 20,000 cases from 1945 onwards), the col-lection of the Pathological Society of Great Britain (www.pathsoc.org) or other collections intended for the general public (www.virtualpathology.leeds.ac.uk) [17, 22, 53] . Virtual slides and seminars are offered by the USCAP online academy (>100 virtual slides from differ-ent organ systems). The number of journals that allow

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M


87

access to WSI examples to illustrate the articles, thus improving the content of the publication, is also on the rise.

Perspectives for the Future

CLM has been used in pathology departments, as well as in the laboratories of the schools of medicine, biology, dentistry and veterinary medicine, for decades. Its re-placement by new technologies represents a major chal-lenge. However, increasing needs of education and grow-ing evidence indicating the very positive effect of WSI on teaching undergraduates and pathologists are resulting in a significant expansion of this tool. Comfort in its use and

the ability to display the same slide anywhere and at any time by several people simultaneously makes this tech-nology much more convenient than CLM. However, the high maintenance costs of the laboratories of microscopy and optimization of the time of teachers and students with the use of WSI is likely to push most training centers towards the replacement of CLM by WSI. WSI is also ex-traordinarily expanding as a tool for tele-education and e-learning. Further investigation is necessary to improve the existing WSI applications specifically designed for ed-ucation and to develop ergonomic tools that improve the navigation of virtual slides. It is possible that in the near future textbooks will have the option to visualize WSI, which will be facilitated with the use of portable devices like tablets or smartphones [22] .

References

1 Paulsen FP, Eichhorn M, Brauer L: Virtual microscopy – the future of teaching histology in the medical curriculum? Ann Anat 2010; 192: 378–382.

2 Blake CA, Lavoie HA, Millette CF: Teaching medical histology at the University of South Carolina School of Medicine: transition to virtual slides and virtual microscopes. Anat Rec B New Anat 2003; 275: 196–206.

3 Pantanowitz L, Valenstein PN, Evans AJ,Kaplan KJ, Pfeifer JD, Wilbur DC, Collins LC, Colgan TJ: Review of the current state of whole slide imaging in pathology. J Pathol In-form 2011; 2: 36.

4 Al-Janabi S, Huisman A, Van Diest PJ: Digital pathology: current status and future perspec-tives. Histopathology 2012; 61: 1–9.

5 Anyanwu GE, Agu AU, Anyaehie UB: En-hancing learning objectives by use of simple virtual microscopic slides in cellular physiol-ogy and histology: impact and attitudes. Adv Physiol Educ 2012; 36: 158–163.

6 Husmann PR, O’Loughlin VD, Braun MW: Quantitative and qualitative changes in teach-ing histology by means of virtual microscopy in an introductory course in human anatomy. Anat Sci Educ 2009; 2: 218–226.

7 Farah CS, Maybury TS: The e-evolution of microscopy in dental education. J Dent Educ 2009; 73: 942–949.

8 Fonseca FP, Santos-Silva AR, Lopes MA, Al-meida OP, Vargas PA: Transition from glass to digital slide microscopy in the teaching of oral pathology in a Brazilian dental school. Med Oral Patol Oral Cir Bucal 2015; 20:e17–e22.

9 Boutonnat J, Paulin C, Faure C, Colle PE, Ronot X, Seigneurin D: A pilot study in two French medical schools for teaching histology using virtual microscopy. Morphologie 2006; 90: 21–25.

10 Ordi O, Bombi JA, Martinez A, Ramirez J, Alos L, Saco A, Ribalta T, Fernandez PL, Campo E, Ordi J: Virtual microscopy in the undergraduate teaching of pathology. J Pathol Inform 2015; 6: 1.

11 Romer DJ, Suster S: Use of virtual microscopy for didactic live-audience presentation in an-atomic pathology. Ann Diagn Pathol 2003; 7: 67–72.

12 Foster K: Medical education in the digital age: digital whole slide imaging as an e-learning tool. J Pathol Inform 2010; 1: 14.

13 Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F: Comparison of a virtual mi-croscope laboratory to a regular microscope laboratory for teaching histology. Anat Rec 2001; 265: 10–14.

14 Merk M, Knuechel R, Perez-Bouza A: Web-based virtual microscopy at the RWTH Aachen University: didactic concept, meth-ods and analysis of acceptance by the stu-dents. Ann Anat 2010; 192: 383–387.

15 Braun MW, Kearns KD: Improved learning efficiency and increased student collabora-tion through use of virtual microscopy in the teaching of human pathology. Anat Sci Educ 2008; 1: 240–246.

16 Craig FE, McGee JB, Mahoney JF, Roth CG: The Virtual Pathology Instructor: a medical student teaching tool developed using patient simulator software. Hum Pathol 2014; 45: 1985–1994.

17 Hamilton PW, Wang Y, McCullough SJ: Vir-tual microscopy and digital pathology in training and education. APMIS 2012; 120: 305–315.

18 Collier L, Dunham S, Braun MW, O’Loughlin VD: Optical versus virtual: teaching assistant perceptions of the use of virtual microscopy in an undergraduate human anatomy course. Anat Sci Educ 2012; 5: 10–19.

19 Szymas J, Lundin M: Five years of experience teaching pathology to dental students using the WebMicroscope. Diagn Pathol 2011; 6(suppl 1):S13.

20 Helin H, Lundin M, Lundin J, Martikainen P, Tammela T, Helin H, van der Kwast T, Isola J: Web-based virtual microscopy in teaching and standardizing Gleason grading. Hum Pathol 2005; 36: 381–386.

21 Marsch AF, Espiritu B, Groth J, Hutchens KA: The effectiveness of annotated (vs. non-anno-tated) digital pathology slides as a teaching tool during dermatology and pathology resi-dencies. J Cutan Pathol 2014; 41: 513–518.

22 Pantanowitz L, Szymas J, Yagi Y, Wilbur D: Whole slide imaging for educational purpos-es. J Pathol Inform 2012; 3: 46.

23 Linder E, Lundin M, Thors C, Lebbad M, Winiecka-Krusnell J, Helin H, Leiva B, Isola J, Lundin J: Web-based virtual microscopy for parasitology: a novel tool for education and quality assurance. PLoS Negl Trop Dis 2008; 2:e315.

24 Kumar RK, Velan GM, Korell SO, Kandara M, Dee FR, Wakefield D: Virtual microscopy for learning and assessment in pathology. J Pathol 2004; 204: 613–618.

25 Chen YK, Hsue SS, Lin DC, Wang WC, Chen JY, Lin CC, Lin LM: An application of virtual microscopy in the teaching of an oral and maxillofacial pathology laboratory course. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008; 105: 342–347.

26 Diaz-Perez JA, Raju S, Echeverri JH: Evalua-tion of a teaching strategy based on integra-tion of clinical subjects, virtual autopsy, pa-thology museum, and digital microscopy for medical students. J Pathol Inform 2014; 5: 25.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M



88

27 Gatumu MK, MacMillan FM, Langton PD, Headley PM, Harris JR: Evaluation of usage of virtual microscopy for the study of histology in the medical, dental, and veterinary under-graduate programs of a UK University. Anat Sci Educ 2014; 7: 389–398.

28 Helle L, Nivala M, Kronqvist P, Gegenfurtner A, Bjork P, Saljo R: Traditional microscopy instruction versus process-oriented virtual microscopy instruction: a naturalistic experi-ment with control group. Diagn Pathol 2011; 6(suppl 1):S8.

29 McCready ZR, Jham BC: Dental students’ perceptions of the use of digital microscopy as part of an oral pathology curriculum. J Dent Educ 2013; 77: 1624–1628.

30 Weaker FJ, Herbert DC: Transition of a dental histology course from light to virtual micros-copy. J Dent Educ 2009; 73: 1213–1221.

31 Krippendorf BB, Lough J: Complete and rap-id switch from light microscopy to virtual mi-croscopy for teaching medical histology. Anat Rec B New Anat 2005; 285: 19–25.

32 Kumar RK, Freeman B, Velan GM, De Per-mentier PJ: Integrating histology and histopa-thology teaching in practical classes using vir-tual slides. Anat Rec B New Anat 2006; 289: 128–133.

33 Durosaro O, Lachman N, Pawlina W: Use of knowledge-sharing web-based portal in gross and microscopic anatomy. Ann Acad Med Singapore 2008; 37: 998–1001.

34 Goldberg HR, Dintzis R: The positive impact of team-based virtual microscopy on student learning in physiology and histology. Adv Physiol Educ 2007; 31: 261–265.

35 Triola MM, Holloway WJ: Enhanced virtual microscopy for collaborative education. BMC Med Educ 2011; 11: 4.

36 Bruch LA, De Young BR, Kreiter CD, Haugen TH, Leaven TC, Dee FR: Competency assess-ment of residents in surgical pathology using virtual microscopy. Hum Pathol 2009; 40: 1122–1128.

37 Gongora JH, Barcelo HA: Telepathology and continuous education: important tools for pa-thologists of developing countries. Diagn Pathol 2008; 3(suppl 1):S24.

38 Scoville SA, Buskirk TD: Traditional and vir-tual microscopy compared experimentally in a classroom setting. Clin Anat 2007; 20: 565–570.

39 Sivamalai S, Murthy SV, Gupta TS, Woolley T: Teaching pathology via online digital mi-croscopy: positive learning outcomes for ru-rally based medical students. Aust J Rural Health 2011; 19: 45–51.

40 Inuwa IM, Taranikanti V, Al-Rawahy M, Habbal O: Anatomy practical examinations: how does student performance on computer-ized evaluation compare with the traditional format? Anat Sci Educ 2012; 5: 27–32.

41 Mione S, Valcke M, Cornelissen M: Evalua-tion of virtual microscopy in medical histol-ogy teaching. Anat Sci Educ 2013; 6: 307–315.

42 Tian Y, Xiao W, Li C, Liu Y, Qin M, Wu Y, Xiao L, Li H: Virtual microscopy system at Chinese medical university: an assisted teaching plat-form for promoting active learning and prob-lem-solving skills. BMC Med Educ 2014; 14: 74.

43 Brick KE, Sluzevich JC, Cappel MA, DiCaudo DJ, Comfere NI, Wieland CN: Comparison of virtual microscopy and glass slide microscopy among dermatology residents during a simu-lated in-training examination. J Cutan Pathol 2013; 40: 807–811.

44 Brick KE, Comfere NI, Broeren MD, Gibson LE, Wieland CN: The application of virtual microscopy in a dermatopathology educa-

tional setting: assessment of attitudes among dermatopathologists. Int J Dermatol 2014; 53: 224–227.

45 Koch LH, Lampros JN, Delong LK, Chen SC, Woosley JT, Hood AF: Randomized compar-ison of virtual microscopy and traditional glass microscopy in diagnostic accuracy among dermatology and pathology residents. Hum Pathol 2009; 40: 662–667.

46 Krupinski EA, Tillack AA, Richter L, Hender-son JT, Bhattacharyya AK, Scott KM, Graham AR, Descour MR, Davis JR, Weinstein RS: Eye-movement study and human performance us-ing telepathology virtual slides: implications for medical education and differences with ex-perience. Hum Pathol 2006; 37: 1543–1556.

47 Stewart J III, Bevans-Wilkins K, Bhattacharya A, Ye C, Miyazaki K, Kurtycz DF: Virtual mi-croscopy: an educator’s tool for the enhance-ment of cytotechnology students’ locator skills. Diagn Cytopathol 2008; 36: 363–368.

48 Donnelly AD, Mukherjee MS, Lyden ER, Ra-dio SJ: Virtual microscopy in cytotechnology education: application of knowledge from vir-tual to glass. Cytojournal 2012; 9: 12.

49 Heffner S: Streamlining tumor board reviews. Adv Lab 2008; 17: 20.

50 Spinosa J: Scripp’s tumor board finds value in digital imaging of slides. Dark Rep 2009; 12: 10–15.

51 Conran R, Fontelo P, Liu F, Fontelo M, White E: Slide2Go: a virtual slide collection for pa-thology education. AMIA Annu Symp Proc 2007, p 918.

52 Khalbuss WE, Pantanowitz L, Parwani AV: Digital imaging in cytopathology. Patholog Res Int 2011; 2011: 264683.

53 Rosai J: Digital images of case reports and oth-er articles. Int J Surg Pathol 2007; 15: 5.

Dow

nloa

ded

by:

Uni

vers

itat d

e B

arce

lona

16

1.11

6.10

0.92

- 4

/25/

2016

11:

51:1

8 A

M


[95]

Estudio número 5

“Virtual Microscopy in the Undergraduate Teaching of

Pathology”

Oriol Ordi, Josep Antoni Bombí, Antonio Martínez, Josep Ramírez, Llúcia

Alòs, Adela Saco, Teresa Ribalta, Pedro L. Fernández, Elias Campo, Jaume

Ordi

Journal of Pathology Informatics 2015 Jan 29; 6:1

Factor de impacto (2016): 0


[96]

J Pathol Inform Editor-in-Chief: Anil V. Parwani , Liron Pantanowitz, Pittsburgh, PA, USA Pittsburgh, PA, USA

OPEN ACCESS HTML format

For entire Editorial Board visit : www.jpathinformatics.org/editorialboard.asp

Research Article

Virtual microscopy in the undergraduate teaching of pathology

Oriol Ordi1, Josep Antoni Bombí1,2, Antonio Martínez1,2, Josep Ramírez1,2, Llúcia Alòs1,2, Adela Saco1, Teresa Ribalta1,2, Pedro L. Fernández1,2, Elias Campo1,2, Jaume Ordi1,2,3

1Department of Pathology, School of Medicine, University of Barcelona, 2Department of Pathology, Hospital Clínic, 3 ISGlobal, Barcelona Center for International Health Research (CRESIB), Barcelona, Spain

E-mail: *Prof. Dr. Jaume Ordi - [email protected] *Corresponding author

Received: 28 August 2014 Accepted: 24 November 2014 Published: 29 January 15

INTRODUCTION

In the last 20 years, web‑based resources developed to supplement or replace the traditional methodologies have expanded dramatically. These resources have shown clear benefits, as classes can be delivered to many students simultaneously, and this has helped medical schools to train in a more cost‑effective way.[1]

Histology and pathology play an essential role in education in undergraduate courses in medicine. The practical knowledge of these disciplines has classically been delivered using glass slides and conventional microscopes (CM), as web‑based resources were limited to static images, which were very different from real practice. Virtual microscopy (VM), also referred to as whole slide imaging, has recently started to change the

Copyright: © 2015 Ordi O. This is an open‑access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

This article may be cited as:Ordi O, Bombí JA, Martínez A, Ramírez J, Alòs L, Saco A, et al. Virtual microscopy in the undergraduate teaching of pathology. J Pathol Inform 2015;6:1.

Available FREE in open access from: http://www.jpathinformatics.org/text.asp?2015/6/1/1/150246

Abstract

Background: Little evidence is available concerning the impact of virtual microscopy (VM) in the undergraduate teaching of pathology. We aimed: (1) to determine the impact in student scores when moving from conventional microscopy (CM) to VM; (2) to assess the students’ impressions and changes in study habits regarding the impact of this tool. Methods: We evaluated two groups taking the discipline of pathology in the same course, one using CM and the other VM. The same set of slides used in the CM classes was digitized in a VENTANA iScan HT (Roche Diagnostics, Sant Cugat, Spain) at ×20 and observed by the students using the Virtuoso viewer (Roche Diagnostics). We evaluated the skill level reached by the students with an online test. A voluntary survey was undertaken by the VM group to assess the students’ impressions regarding the resource. The day and time of any accession to the viewer were registered. Results: There were no differences between the two groups in their marks in the online test (mean marks for the CM and the VM groups: 9.87 ± 0.34 and 9.86 ± 0.53, respectively; P = 0.880). 86.6% of the students found the software friendly, easy‑to‑use and effective. 71.6% of the students considered navigation easier with VM than with CM. The most appreciated feature of VM was the possibility to access the images anywhere and at any time (93.3%). 57.5% of the accesses were made on holidays and 41.9% later than 6:00 pm. Conclusions: Virtual microscopy can effectively replace the traditional methods of learning pathology, providing mobility and convenience to medical students.Key words: Pathology, undergraduate teaching, whole slide imaging

Access this article onlineWebsite: www.jpathinformatics.org

DOI: 10.4103/2153-3539.150246

Quick Response Code:

[Downloaded free from http://www.jpathinformatics.org on Thursday, May 12, 2016, IP: 82.159.201.132]

J Pathol Inform 2015, 1:1 http://www.jpathinformatics.org/content/6/1/1

way in which these disciplines can be delivered online by providing the ability to scan entire glass slides at diagnostic resolution. A digital slide of the tissue section is created and with the use of specific software can be viewed and magnified in real‑time across the web very much like using a CM. Currently, several commercially available systems can digitize glass slides containing tissue sections and produce virtual slides of excellent quality. The rapid progress of this technology and its many potential benefits will probably result in a progressive shift from conventional to VM in routine diagnostic pathology.[2‑4]

Several studies have documented the success of VM in graduate education in medical,[5‑9] dental,[10,11] and veterinary schools.[12‑14] However, although VM has been used for many years in the US, reports on the experience with this technology in the undergraduate teaching of pathology are still limited and there is little evidence about its true impact on students’ knowledge and study habits.

In this study, we report the experience in the transition of a General Pathology course in the medical school of the University of Barcelona from practical teaching based on CM to a new format in which this tool has been totally eliminated, and all microscopy work is conducted on the computer using VM. We specifically aimed: (1) To assess the students’ impressions regarding the impact of VM on their learning and to objectively assess the success of implementing this new technology in their curriculum; (2) to determine whether moving from glass to virtual slides has an impact on student scores in practical exams of pathology; and (3) to evaluate the changes in study habits associated with the introduction of this tool in undergraduate teaching.

METHODS

The study was conducted in the Department of Pathology of the School of Medicine in the University of Barcelona, Spain. This university implemented the open‑source course management system modular object‑oriented dynamic learning environment (Moodle) in 2004, and all the supporting information and most of the activities of each discipline are available to the students in this platform.

The action was conducted in General Pathology, which is delivered in the 3rd year as a 4‑month course and has 6 European Credit Transfer System credits.[15] Two different groups take the discipline each year, the first group from September to January and the second from February to May. All the students in both groups had previously had a whole year of experience using CM in the course of histology, but none had had any previous experience with VM. During the 2013–2014 course, one group studied anatomical pathology using CM whereas the second group used VM.

Characteristics of the Conventional Microscopy CourseThis group took the discipline from September 2013 to January 2014 using CM, following the same rules established in the previous 5 years. The group was divided into six smaller groups each composed of 15 students. Two practical classes were scheduled during the course, each of which included 16 histological slides stained with hematoxylin and eosin, being representative of basic pathological lesions. Eighteen sets of slides including consecutive sections of the 16 cases were available for the use of the students in each practical class. Practical classes of 2 h in duration were delivered in the microscopy room of the medical school and were conducted by a professor of pathology. The professor briefly showed the most relevant features of each particular slide with a microscope connected to a video camera and several screens. Thereafter, each student had a set of the slides for his/her use and had 90 min to observe the slides on his/her own single‑headed microscope with the support of the faculty member, who solved all the questions and problems brought up by the students.

Characteristics of the Virtual Microscopy CourseThis group took the discipline from February to May 2014 using only VM. The practical course included the same 32 cases used in the previously described group. The group was divided into six smaller groups each composed of 15–16 students. A single practical class was scheduled at the beginning of the course (February). The class was delivered in a room equipped with a computer with internet connection and a 52‑inch screen and was conducted by a faculty member, being of 1/2 h in duration. The professor briefly showed how to accede to the website, the general characteristics of the navigator and how to retrieve the supporting information. After this initial session, all the students were allowed to access the virtual slides any day and at any time from any computer connected to Internet. The students were given the opportunity to contact their tutors for any problem or doubt encountered when observing the slides on their computers.

Virtual Slides, Navigation and Supporting FilesAll the cases were digitized in a VENTANA iScan HT (Roche‑Ventana Medical Systems, Tucson, AZ, USA) at a magnification of ×20. The system creates high‑resolution digital images of tissue sections. All files were stored on a server hosted at the Spanish Division of Roche Diagnostics. The students access the virtual slides through a hyperlink on the Moodle platform, using their own computers as virtual microscopes. The images are viewed in the Virtuoso viewer (Roche‑Ventana Medical Systems, Tucson, AZ, USA), which works as a web browser and simulates a CM [Figure 1]. Virtuoso is designed to organize the images into different cases and the cases into groups. No specific



software installation is required to visualize the virtual slides.

Two supporting pdf files were posted on the main page of the discipline in the Moodle platform. One included general information to guide student accession to the viewer and the username and password necessary to accede to the website. The second pdf file included educational text discussions for each particular case.

Online Evaluation of the Skill LevelAn online quiz was performed to evaluate the skill level reached by the students in the evaluation of the microscopic lesions. A question bank containing 200 multiple choice questions was created in the Moodle platform.[16] All questions were based on static microphotographs prepared by the faculty by selection of specific areas from the same 32 glass slides included in both practical classes and had 5 possible answers. All the questions only had one correct answer qualified with a mark of +1. Each wrong answer was qualified with a mark of –0.25. For the examination, 40 questions were randomly selected from the pool. The test was available on the Moodle platform during a 24 h period. Both the order of the questions, as well as that of the answers was automatically distributed randomly for each student, thus, questions and answers were presented in a different order to every student. There was a 20 min time limit to complete the exam.

Audit of Student Access to the Virtual Slide ViewerFor the VM course, the viewer registered any access to the virtual slides by any of the students. This registration was performed anonymously, as all the students logged in using the same login and password. The day and time of any single accession, as well as the time spent by the student on each slide, were registered, and a pdf file was created with all the information.

Voluntary Student SurveyAt the conclusion of the VM course, a voluntary survey was undertaken by the students to assess the students’

impressions regarding the impact of VM on their learning and the success of implementing this new technology in their curriculum. The survey was designed using the free website https://www.surveymonkey.com/(SurveyMonkey®, Menlo Park, CA, USA) and was posted as an online hyperlink on the Moodle platform, which remained open for a whole week after the online exam was completed. Questions were related to the quality and easiness‑of‑use of the software and navigation, VM versus CM, supporting information, introduction of the activity by the professor and an online quiz [Table 1]. The student survey was designed as a Likert‑scale questionnaire with a five‑point scale with the following options: Strongly agree, agree, undecided, disagree, and strongly disagree. Ethical clearance was granted with participation in evaluations being entirely voluntary and completely anonymous.

Data AnalysisStatistical analysis was performed using the SPSS (version 18.0; SPSS, Inc., Chicago, IL). The results are presented as absolute numbers and percentages or mean and standard deviation. The analysis was mostly descriptive and included Chi‑square tests.

RESULTS

Characteristics of the GroupsThe CM course had 88 students, 67.0% (59/88) females and 33.0% (29/88) males, with a mean age of 20.6 ± 1.4. The VM course had 93 students, 68.8% (64/93) females and 31.2% (29/93) males, with a mean age of 20.8 ± 1.3. No differences were observed between the two groups.

Characteristics of the Virtual SlidesThe size of the files ranged from 149,321 to 1,851,049 Kb (mean 751,562.7 ± 413,330.2 Kb). The total weight of the 32 files was 24,050,005 Kb. The scanned images can be viewed up to a magnification of ×400 and are always in focus, with optimized contrast and adjusted illumination. At high magnifications it is easy for the student to maintain orientation with respect to the entire section, because the system indicates the position of the slide on a thumbnail showing a small representation of the section [Figure 1].

Online Evaluation of the Skill LevelThe mean mark in the online test in the CM course was 9.87 ± 0.34 (range: 8.3–10), with all the students passing the exam. Seventy‑six out of 88 (86.4%) answered all the questions correctly. The mean mark in the VM course was 9.86 ± 0.53 (range: 6.7–10) with 91/93 students (97.8%) passing the online test. Eighty‑five out of 93 (91.4%) answered all the questions correctly. No differences were observed between the two groups (P = 0.880, Student’s t‑test).

Figure 1: Screen shot of the virtual microscope display. A thumbnail showing a small representation of the whole section makes easier for the student to maintain orientation with respect to the entire section



Audit of Student Accesses to the ViewerThe number of visits to the VM from the opening of the website to the day of the exam is shown in Figure 2; about 80.3% of the accesses (862/1073) were done in the week prior to the examination; 57.5% of the visits were made on holidays and 42.5% on working days. The times of access during the day are shown in Figure 3. The earliest access was at 8:33 am and the latest at 00:55 am, with 58.1% of the visits being observed between 8:00 am and 6:00 pm and 41.9% later than 6:00 pm. The mean length of the students’ accesses was 4 min 2 s. Thus, the overall time spent by the students on studying the 32 slides included in the course (including the 30‑min time of the initial presentation of the website) was 2 h 50 min.

Student SurveySixty‑one out of 93 (65.6%) students participated in the survey. Table 1 shows the questions included in the survey and the students’ answers, and Figure 4 shows the mean students’ ratings for the main items of the survey concerning the use of VM; 86.6% of the students found the software friendly, easy‑to‑use and effective for the purposes of the course. The most appreciated feature of VM was the possibility to access the images anywhere and at any time (93.3%), and 71.6% of the students thought that navigation with the virtual was easier than with glass slides. Although a significant percentage of the students were neutral to both methods, most (50.8%) preferred VM to CM.

Consequences on the Workload for the Faculty StaffThe workload in terms of time in classroom teaching for the faculty staff was reduced from 20 h to 2.5. None of the students required faculty assistance. No questions to the tutors were registered in relation to problems or

doubts encountered when observing the slides on their computers.

CONCLUSIONS

The results of this study confirm that VM can effectively replace CM to teach pathology in undergraduate courses in medical schools and show that the microscopic skills acquired with VM are comparable to those acquired with CM, the classical tool for teaching pathology. The overall feedback from the students was highly positive. Students complemented the ease of use of the software. Students felt they worked faster with VM, and over 70% thought that the navigation with the VM was easier than with the CM. The most appreciated feature of VM was the possibility to access the images anywhere and at any

Table 1: Student’s responses to the survey regarding the use of virtual and the conventional microscope

Questions included in the student’s survey Strongly agree

Agree Neutral Disagree Strongly disagree

The software is friendly, easy‑to‑use and effective for the purposes of the course 43.3 43.3 10.0 3.4 0The access to the virtual slides is quick 21.7 53.3 13.4 8.3 3.3I liked the possibility to access the images anywhere and at any time 65.0 28.3 5.0 1.7 0The quality of the image of virtual slides is adequate 26.7 38.3 21.7 8.3 5.0VM allows time saving 31.2 44.3 21.3 1.6 1.6The identification of cells and structures with VM is easy 6.7 41.7 38.3 10.0 3.3Navigation with the VM viewer is easier than with glass slides 33.3 38.3 26.7 1.7 0I had problems with the navigation 3.3 13.3 15.0 31.7 36.7The presentation of the virtual viewer and the slides by the professor is useful 6.7 26.7 33.3 16.7 16.6The presentation of the virtual viewer by the professor is unnecessary 21.7 6.7 35.0 18.3 18.3The supporting material (pdf document) is useful and adequate 73.8 21.3 4.9 0 0The test based on image captures is a good way to evaluate the knowledge acquired with VM

14.7 57.4 18.1 4.9 4.9

I prefer virtual to CM 21.3 29.5 34.4 11.5 3.3I prefer conventional to VM 1.6 13.1 42.6 23.0 19.7

The figures indicate percentages, VM: Virtual microscopy, CM: Conventional microscopy

Figure 2: Number of accesses to the virtual slides from the opening of the website to the day of the exam. The X axis shows the day of accession in relation to the exam (day 0). The Y axis shows the absolute number of accessions. Red marks identify accessions on weekends/holidays (Saturday, Sunday or other holidays). Gray marks indicate work days



time. This finding was highlighted by data obtained from the audit of accesses to the navigator, showing that over half of the accesses were made on holidays, and over one‑third were made after working hours. Finally, the introduction of VM resulted in a significant reduction of the workload for the students and for the faculty staff of pathology in terms of classroom teaching, although there was a significant increase in the time related to the preparation of the material. Interestingly, the transition from conventional to VM was not a gradual process, but a sudden change, showing that positive results may be immediate.

Curriculum reform in medical schools worldwide has focused on a reduction in contact hours to decompress crowded programs, an emphasis on independent learning, and on the development of interpersonal skills and problem‑solving abilities.[11] Achieving this objective has inevitably meant that time has been reallocated from traditional areas to new educational activities deemed to be more important. In some medical schools, this has led to curricula that offer diminished opportunities for students to learn the basic medical sciences.[1,5,11,17] In this context of standardized curricula and the growing number of medical students, new strategies have been employed to improve the student experience of learning pathology. The introduction of VM, an adequate alternative to the traditional methods of teaching pathology, can help the students to achieve a satisfactory knowledge of these basic disciplines in these newly reformed curricula. Our results showed no differences in the exams between the VM and the CM groups, plus the overall favorable feelings of the students about VM are in keeping with the adequacy of this method. For the medical staff, using this tool not only results in a reduction of the work load but also allows obtaining more information about how students learn pathology, how instructive the laboratory classes are, and which slides are of real didactic value for the students.

Virtual microscopy imitates the use of a traditional microscope and glass slide. One of the main advantages of

this tool for students is that the slides are always in focus, with optimized contrast and adjusted illumination. Indeed, over 70% of the students thought that the navigation with the VM viewer was easier than with glass slides.[2‑4]

The anonymous survey showed that the students found VM useful. Our results are in keeping with previous reports showing that the students’ experience with VM is very favorable.[11] This provides clear evidence of the learning benefits derived from using this tool. VM allows students to independently explore the entire histological slide, as well as control the content and its rhythm of delivery. As observed in previous studies,[5,6,11,18,19] this interactive technology makes microscopic laboratory studies in pathology more efficient and teaching resources more portable and independent of class schedules. As shown in our study, according to the students, the most appreciated feature of VM was the possibility to accede to the images anywhere and at any time. Indeed, data obtained from the audit of accesses to the navigator showed that over half of the accesses were made on holidays and over one‑third were made after working hours. The adoption of electronic course materials, along with almost universal use of personal and laptop computers by the medical school students facilitates the introduction of VM.[11]

Although the initial equipment and software cost for creating VM is high, this new technology has the potential to revolutionize the way individuals teach and learn from microscopic images. With VM, the most representative slides with the best quality material can reassuringly be included in teaching sets. Not only can such materials be easily added to the virtual sets, but compared with glass slides these digital slides will not fade, break, or disappear. Scanned slides for dedicated teaching should be de‑identified prior to making them available for general users. One of the main advantages of VM is the portability (time and location), and ease of maintenance. Finally, this tool may allow reducing or even eliminating the expensive laboratories of microscopy.[2‑5]

Figure 3: Time of accesses to the virtual slides during the day. The X axis shows the time of the day and the Y axis shows the absolute number of accessions

Figure 4: Mean students’ ratings for the main items of the survey regarding the use of virtual microscopy



The main strength of our study is that it allows adequate comparison of two very similar groups from the same course working with the same material and that it provides objective data on how students learn pathology. A possible limitation of our study is the use of an examination system restricted to standard static images. However, this allowed comparison of the results with the CM group, as all the image captures were made from the same cases. Examinations that apply virtual technology require more sophisticated management software.[4] Finally, the very good results obtained in the examinations should be considered as the consequence of the extremely high marks required in Spain to accede to the medical schools.

In conclusion, evidence showing that the microscopic skills achieved by students with VM are comparable to these acquired with CM indicates that this technology can effectively replace the traditional methods of learning pathology. One of its main advantages is that it provides mobility and convenience to medical students.

REFERENCES

1. Williams G, Lau A. Reform of undergraduate medical teaching in the United Kingdom: A triumph of evangelism over common sense. BMJ 2004;329:92‑4.

2. Al‑Janabi S, Huisman A, Van Diest PJ. Digital pathology: Current status and future perspectives. Histopathology 2012;61:1‑9.

3. Pantanowitz L, Valenstein PN, Evans AJ, Kaplan KJ, Pfeifer JD, Wilbur DC, et al. Review of the current state of whole slide imaging in pathology. J Pathol Inform 2011;2:36.

4. Pantanowitz L, Szymas J, Yagi Y, Wilbur D. Whole slide imaging for educational purposes. J Pathol Inform 2012;3:46.

5. Blake CA, Lavoie HA, Millette CF. Teaching medical histology at the University of South Carolina School of Medicine: Transition to virtual slides

and virtual microscopes. Anat Rec B New Anat 2003;275:196‑206.6. Boutonnat J, Paulin C, Faure C, Colle PE, Ronot X, Seigneurin D. A pilot

study in two French medical schools for teaching histology using virtual microscopy. Morphologie 2006;90:21‑5.

7. Foster K. Medical education in the digital age: Digital whole slide imaging as an e‑learning tool. J Pathol Inform 2010;1.

8. Goldberg HR, Dintzis R. The positive impact of team‑based virtual microscopy on student learning in physiology and histology. Adv Physiol Educ 2007;31:261‑5.

9. Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F. Comparison of a virtual microscope laboratory to a regular microscope laboratory for teaching histology. Anat Rec 2001;265:10‑4.

10. Chen YK, Hsue SS, Lin DC, Wang WC, Chen JY, Lin CC, et al. An application of virtual microscopy in the teaching of an oral and maxillofacial pathology laboratory course. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008;105:342‑7.

11. Weaker FJ, Herbert DC. Transition of a dental histology course from light to virtual microscopy. J Dent Educ 2009;73:1213‑21.

12. Dee FR, Meyerholz DK. Teaching medical pathology in the twenty‑first century: Virtual microscopy applications. J Vet Med Educ 2007;34:431‑6.

13. Mills PC, Bradley AP, Woodall PF, Wildermoth M. Teaching histology to first‑year veterinary science students using virtual microscopy and traditional microscopy: A comparison of student responses. J Vet Med Educ 2007;34:177‑82.

14. Neel JA, Grindem CB, Bristol DG. Introduction and evaluation of virtual microscopy in teaching veterinary cytopathology. J Vet Med Educ 2007;34:437‑44.

15. Doménech Martínez E, Armas Ramos H, Castro Conde JR, González Díaz JP, Méndez Pérez A, Ormazábal Ramos C, et al. Study of the introduction of the European Credit Transfer System (ECTS) in pediatrics and modification of the teaching methodology. An Pediatr (Barc) 2006;65:415‑27.

16. Inuwa IM, Taranikanti V, Al‑Rawahy M, Habbal O. Anatomy practical examinations: How does student performance on computerized evaluation compare with the traditional format? Anat Sci Educ 2012;5:27‑32.

17. Bloodgood RA, Ogilvie RW. Trends in histology laboratory teaching in United States medical schools. Anat Rec B New Anat 2006;289:169‑75.

18. Sims MH, Mendis‑Handagama C, Moore RN. Virtual microscopy in a veterinary curriculum. J Vet Med Educ 2007;34:416‑22.

19. Szymas J, Lundin M. Five years of experience teaching pathology to dental students using the WebMicroscope. Diagn Pathol 2011;6 Suppl 1:S13.



[103]

V. Discusión


[104]


[105]

El uso de la MV en el diagnóstico primario se está extendiendo

considerablemente en los últimos años, y cada vez más centros disponen de esta

herramienta en los Servicios de Anatomía Patológica. Esto es debido a las numerosas

ventajas que ofrece este sistema, como las herramientas informáticas, la portabilidad

o la facilidad para visualizar la misma preparación por un grupo amplio de personas. A

estas cualidades hay que añadir otras de más reciente aparición, como la

cuantificación automática de células marcadas con tinciones inmunohistoquímicas o el

reconocimiento de patrones histológicos, las cuales sirven de ayuda a la hora de

realizar el diagnóstico y disminuyen la variabilidad entre observadores. A pesar de

estas ventajas existen reticencias a su uso en el diagnóstico primario, siendo la

principal el desconocimiento de si existe suficiente evidencia científica que asegure su

no inferioridad respecto a la MC. La realización de los estudios de validación de la MV

en el diagnóstico primario se ve facilitada por la existencia de guías publicadas por la

American Telemedicine Association, el College of American Pathologists y la Canadian

Association of Pathologist, así como del Libro Blanco de la Sociedad Española de

Patología; donde se recogen una serie de recomendaciones útiles a la hora de realizar

la validación del diagnóstico primario con MV.

Nuestro estudio 1 tiene como principal objetivo evaluar si existe suficiente

bibliografía que ponga de manifiesto una buena concordancia entre los diagnósticos

con MV y MC; para ello hemos llevado a cabo una revisión de los estudios de

validación publicados hasta el momento, valorando si tanto la metodología como las

muestras evaluadas son adecuadas. Los resultados de este estudio pusieron de

manifiesto que, independientemente de la subespecialidad, todos los estudios sobre

validación tienen una muy buena correlación entre los diagnósticos alcanzados con MV

y MC; por lo cual la MV parece ser una herramienta adecuada para el diagnóstico

histológico de rutina, presentando además múltiples ventajas sobre la MC. Sin

embargo, a pesar de la buena evidencia demostrada con la MV en la práctica rutinaria

existen numerosas áreas de la Anatomía Patológica que presentan deficiencias en los

estudios de validación o una total ausencia de los mismos. Entre estas áreas se

encuentran la hematopatología, la patología hepática, ginecológica, ósea, endocrina y

de partes blandas. Algunas de estas áreas son similares entre ellas y presentan


[106]

características superponibles a otras que sí tienen estudios de validación, por lo cual

no resultan necesarios estudios adicionales. Sin embargo, otras como la

hematopatología o la patología hepática presentan características propias, por lo que

resultan imprescindibles nuevos estudios de validación antes del uso de la MV para

realizar el diagnóstico primario de forma generalizada en un Servicio de Anatomía

Patológica.

La citología parece ser una excepción, pues la aplicación de la MV en esta área

presenta una mayor controversia por la necesidad usar ejes adicionales a la hora de

escanear para permitir el enfoque a distintos planos, incrementando de forma muy

significativa el tamaño de las imágenes generadas.

Varios estudios ponían de manifiesto que, como ocurre con otras herramientas

nuevas, existe una curva de aprendizaje por lo que el tiempo empleado en el

diagnóstico, y en menor medida la concordancia intra e inter-observador, pueden ser

subóptimos en las fases iniciales del uso de esta tecnología [43–46,48].

Algunos estudios destacan la existencia de otras aplicaciones para la MV además

del diagnóstico primario, como por ejemplo la creación de reconstrucciones 3D de

imágenes en 2D de biopsias, lo cual puede ayudar a mejorar la comprensión de los

patrones de crecimiento y de la disposición de las células en el espacio [21,68]. Otra

área que se encuentra en reciente expansión es la referente al reconocimiento de

patrones histológicos usando análisis de imagen. Esta herramienta puede incrementar

significativamente la reproducibilidad de los diagnósticos entre patólogos en muchas

subespecialidades, consiguiendo un diagnóstico más preciso, con la repercusión que

esto tiene sobre el pronóstico y tratamiento de los pacientes [69–74].

Ante la luz de estos resultados realizamos nuevos estudios de validación del uso

de la MV en el diagnóstico de biopsias ginecológicas (estudio 2) y biopsias pequeñas

hepáticas (estudio 3), pues no existía en la bibliografía actual suficiente evidencia

sobre la no inferioridad del diagnóstico de rutina con MV comparado con MC.

Los resultados del estudio 2 muestran una elevada concordancia entre los

diagnósticos realizados con MC y MV (superior al 94%) en una serie amplia de

muestras ginecológicas de rutina. El valor de Kappa, considerado como la medida del


[107]

nivel de concordancia entre observadores corregido por el azar, resultó casi perfecto

(0,914). La cantidad de discrepancias observadas en nuestro estudio se encuentra

dentro del rango de la variabilidad intra-observador que se suele observar en patología

[75,76]. El diagnóstico final de consenso coincidió con el diagnóstico realizado con MV

en un 22,2% de las discrepancias mayores y en un 35,3% de las menores. Ninguna de

las discrepancias se consideró como relacionada con una pobre calidad de imagen de

MV o con la imposibilidad de conseguir una magnificación suficiente, sino que

estuvieron asociadas fundamentalmente a interpretaciones diferentes de casos

difíciles o a la presencia de lesiones de pequeño tamaño que no fueron identificadas

en la evaluación. Por tanto, nuestros resultados confirman que la MV puede ser

utilizada de forma segura en el diagnóstico histológico primario de la patología

ginecológica.

Ocho de las nueve discrepancias observadas en el estudio (88,9%) fueron en el

diagnóstico de H-SIL como L-SIL o como negativas o cambios reactivos en el cérvix

uterino (cuatro casos de cada uno). Así mismo, 13 de las 17 discrepancias menores

(76,5%) estuvieron relacionadas con el diagnóstico de L-SIL versus epitelio cervical

normal o reactivo. Como consecuencia, el valor Kappa de las biopsias o escisiones

cervicales de pacientes referidas por citología cérvico-vaginal anormal fue 0,832,

claramente inferior al valor observado en el grupo general de todas las biopsias. Esta

observación es concordante con los resultados de varios estudios que muestran la

existencia de una importante variación intra e inter-observador en la interpretación de

biopsias de cérvix uterino teñidas con H&E, con coeficientes de Kappa situados en

general entre 0,45 y 0,50; indicando un nivel de concordancia solo moderado [75,77–

82]. La estimación de la reproducibilidad en la interpretación de las muestras cérvico-

vaginales en el curso de un estudio de triaje de lesiones ASCUS/LSIL comparando los

diagnósticos del centro de origen con los resultados de la revisión de patólogos

expertos mostró una reproducibilidad moderada (Kappa<0,5) en la interpretación de

biopsias [83]. En este estudio la ausencia de reproducibilidad fue sustancialmente

mayor en biopsias por punción tipo punch que en muestras por escisión, y la

variabilidad fue mayor en la evaluación de lesiones de bajo grado; al igual que fue

observado en nuestro estudio.


[108]

En nuestro estudio el patólogo responsable de la evaluación con MV tenía

solamente una mínima experiencia previa en el uso de la herramienta, pero ello no

afectó de forma significativa a la exactitud del diagnóstico, ni siquiera en el periodo

inicial del estudio. No obstante, sí se detectó un incremento claro del nivel de

reproducibilidad durante el periodo de la realización del estudio, así como una

disminución en el número de casos en los que el patólogo requirió el uso de la MC para

confirmar el diagnóstico. El diagnóstico de consenso en los casos discrepantes

coincidió con el de la evaluación con MC en el 82,4% de las biopsias en el periodo

inicial, presentando una tendencia más equilibrada en el segundo periodo,

coincidiendo con el de la evaluación con MV en un 55,5% de los casos. Todas estas

evidencias indican que el incremento en la experiencia con el uso de la MV mejora los

resultados diagnósticos y que los patólogos, tras un periodo de adquisición de

experiencia, son capaces de diagnosticar prácticamente todas las muestras de forma

fiable exclusivamente con el uso de la MV.

El patólogo que trabajó con la MV no refirió dificultades para alcanzar el

diagnóstico con el aumento de 200x con el que se realizó el escaneo, indicando que no

resulta necesaria una mayor resolución en la mayor parte de las biopsias ginecológicas.

Esta estrategia de escaneo es la usada en la mayor parte de los estudios de validación

porque comporta un ahorro en el tiempo de escaneo y en los requerimientos de disco

para el almacenaje de imágenes [43,84–89]. Sin embargo, es probable que se requiera

un incremento en la magnificación del escaneo (400x) para un porcentaje pequeño de

casos para poder alcanzar un diagnóstico de forma segura.

La principal fortaleza de nuestro estudio es que el mayor análisis de validación

realizado hasta la fecha en biopsias ginecológicas, el cual incluye un número

significativo de casos que permiten un estudio estadístico robusto. Hasta el momento

solo una publicación había analizado la correlación entre MV y MC en la evaluación de

secciones congeladas de 52 lesiones ováricas, mostrando que, al igual que se observa

en nuestro estudio, la correlación entre ambos diagnósticos es muy buena [7]. La

segunda fortaleza de nuestro estudio es que los patólogos implicados en el diagnóstico

tuvieron a su disposición la información clínica relevante para el diagnóstico.


[109]

La principal limitación de nuestro estudio es que no se evaluó la variabilidad

intra-observador de los diagnósticos con MV y MC, lo que se considera como el mejor

diseño a la hora de evaluar la concordancia entre los dos sistemas [3]. Sin embargo, la

buena reproducibilidad inter-observador alcanzada con el estudio sugiere que los

resultados serían muy similares si las evaluaciones fuesen realizadas por el mismo

observador.

En conclusión, los diagnósticos realizados con MV son altamente concordantes

con los emitidos usando MC en las biopsias ginecológicas de rutina, por lo que la MV es

una herramienta que permite realizar diagnósticos primarios, presentando además

múltiples ventajas respecto a la MC.

La patología hepática resultó ser otra de las subespecialidades sin estudios de

validación que permitieran el uso de la MV en el diagnóstico primario. Esta área de la

patología tiene características que la diferencian de las demás subespecialidades;

entre ellas el pequeño tamaño de algunas biopsias y la presencia de muestras

derivadas del trasplante hepático, donde pequeñas variaciones en la histología pueden

cambiar de forma significativa el diagnóstico y el manejo clínico del paciente. Por este

motivo llevamos a cabo el estudio 3, donde evaluamos la concordancia tanto intra

como inter-observador en el diagnóstico de biopsias con aguja hepáticas usando MC y

MV.

Los resultados muestran una elevada concordancia intra-observador entre la

evaluación con MC y con MV de dos patólogos, siendo mayor al 90% con ambos

observadores. El valor de Kappa, considerado como la medida del nivel de

concordancia entre observadores corregido por el azar, fue casi perfecto; siendo de 0.9

para las comparaciones intra-observador de ambos observadores, y de 0.9 y 1 para las

comparaciones inter-observador con MC y MV respectivamente.

El nivel de discrepancias entre los diagnósticos con MV y MC fue menor al 10%, y

solo se identificaron discrepancias menores, pues no representaron ningún impacto en

el manejo del paciente. Además, cabe destacar que ninguna de ellas fue relativa a una

deficiente calidad de imagen o a un aumento insuficiente. Todas las discrepancias


[110]

encontradas fueron secundarias al pequeño tamaño del material o a la dificultad

intrínseca del caso.

Existen numerosos estudios que muestran variaciones sustanciales inter e intra-

observador en el diagnóstico de biopsias hepáticas, aunque se use MC en ambas

visualizaciones. Estos estudios se encuentran limitados a patologías concretas como la

esteatohepatitis no alcohólica o la hepatitis crónica de origen viral [90–94]. En uno de

los estudios donde se evaluaba la concordancia intra-observador de 50 biopsias

hepáticas orientadas como esteatohepatitis no alcohólica, Kleiner et al reportaron un

valor de Kappa de 0.61 [91]. Nuestro estudio mostró una tasa de concordancia mayor,

con un valor de Kappa que variaba entre 0.7 y 0.9 en las diferentes comparaciones, a

pesar de que se evaluó un numero mayor de casos y se incluyó esteatohepatis

alcoholica y no alcoholica. Tres de los estudios evaluaban la concordancia intra-

observador en el diagnóstico de la hepatitis crónica de origen viral. La evaluación de

los grados de fibrosis y el estadio en esos estudios mostraron valores de Kappa que

oscilaban entre 0.72 y 1 [92–94], los cuales son comparables con los hallados en

nuestro estudio (entre 0.7 y 0.9). Estas discrepancias pueden ser atribuidas a la

inherente variabilidad intra-observador en el diagnóstico de las biopsias con aguja

hepáticas. Algunos de estos estudios evaluaron distintas características histológicas

por separado, mostrando altas tasas de concordancia en la evaluación de la esteatosis

(K=0,79), necrosis periportal (K=0,74) y fibrosis (K=0,86), y tasas más bajas en la

necrosis lobulillar (K=0,42) [90,91]. En nuestro estudio los niveles de concordancia

entre la evaluación con MV y MC de todos esos hallazgos histológicos fueron aún

mejores. Estos resultados confirmaron que la MV puede ser usada para el diagnóstico

primario de rutina de las biopsias hepáticas.

Una posible limitación de nuestro estudio es el bajo número de casos de cada

una de las patologías hepáticas, comparado con el de los estudios previos [90–94]. Sin

embargo, nuestro estudio está diseñado para evaluar el diagnóstico en la práctica

rutinaria, donde se recibe una gran variedad de patologías distintas, y no para evaluar

una patología concreta.


[111]

El análisis por separado de la patología de trasplante (n=64) mostró una alta

concordancia intra-observador entre diagnósticos con MV y MC, siendo casi perfecta

(93.7%; κ= 0.9 para un observador y 87.5%; κ= 0.8 para el otro). No hubo diferencias en

el diagnóstico de rechazo.

Cabe destacar, que los patólogos no refirieron ninguna dificultad para alcanzar el

diagnóstico con la magnitud de 400x usada en las preparaciones. Al contrario de lo que

sucede en otras áreas de la patología, donde objetivos menores son suficientes, en la

patología hepática 400x parece la magnitud más aducada debido a las características

propias de esta subespecialidad; como son el pequeño tamaño de las biopsias con

aguja o la necesidad de evaluar cambios histológicos muy sutiles [43,84,85,87,88].

En conclusión, los diagnósticos de biopsias pequeñas hepáticas realizados con

MV tienen una gran concordancia intra e inter-observador respecto a los llevados a

cabo con MC. Nuestros resultados confirman que la MV puede ser usada de forma

segura para el diagnóstico histológico primario de biopsias hepáticas, tanto de hígados

de trasplante como de nativos.

Cabe destacar que las altas tasas de concordancia halladas tanto en patología

hepática (estudio 3) como ginecológica (estudio 2) son similares a las publicadas en

otros estudios de validación del uso de la MV en biopsias de piel [43,95], mama

[45,96], próstata [76,85], vejiga urinaria [97], patología gastrointestinal [86,98] y

pediátrica [89], recogidos en la revisión realizada en el estudio 1. Los patólogos

encargados del diagnóstico con MV también destacaron algunas de sus ventajas, como

la posibilidad de usar herramientas informáticas y de análisis de imagen. Éstas

permiten realizar de forma más objetiva la medición de lesiones y de la profundidad de

infiltración, lo cual es una información muy relevante en algunas áreas anatómicas

como la vulva, cérvix uterino o endometrio, entre otras. La posibilidad de visualización

de varias preparaciones a la vez en una misma pantalla también resulta muy util,

especialmente en las biopsias hepáticas donde se realizan múltiples tinciones

frecuentemente. Además, la MV facilita la realización de consultas diagnósticas, así

como los comités multidiscipliarios.


[112]

A pesar de los buenos resultados de la MV en el diangóstico rutinario, hay varias

consideraciones a tener en cuenta que dificultan el uso de este sistema, siendo las

principales el rechazo de los patólogos a abandonar la MC y el elevado coste

económico de la adquisición de equipamiento y del archivo de las imágenes generadas.

También cabe destacar los aspectos legales que envuelven al diagnóstico primario con

MV referentes a la confidencialidad de datos, la calidad de imagen, la calidad de los

monitores, el espacio de almacenamiento y la confianza a la hora de realizar los

diagnósticos [17–22]. Estas cuestiones comienzan a ser reguladas en otros países como

EEUU, donde la Food and Drug Administration (FDA) empieza a establecer la normativa

a seguir a la hora de diagnosticar usando MV. En el ámbito nacional, aún no existe

legislación al respecto, pero cabe esperar que en un futuro cercano ésta exista.

La docencia es otro de los ámbitos donde la MV puede tener un gran impacto

positivo debido a sus numerosas ventajas, destacando la portabilidad o la posibilidad

de visualizar una misma preparación por un gran número de personas. Al igual que

ocurre con el diagnóstico primario, resulta fundamental asegurarse de que la MV es

capaz de sustituir a la MC sin que el aprendizaje de los alumnos de pre y post-grado se

vea afectado de forma desfavorable. Con este objetivo realizamos el estudio 4, el cual

mostro la existencia de numerosas publicaciones científicas con unos buenos

resultados en el uso de la MV aplicada a la docencia de pre y post-grado, evidenciando

la no inferioridad de esta nueva tecnología respecto a la MC. En el caso de la docencia

de pre-grado tanto en facultades de Medicina como de Odontología, Veterinaria o

Parasitología, los resultados muestran que la MV puede sustituir a la MC sin que los

conocimientos de los estudiantes se vean afectados, presentando además numerosas

ventajas sobre esta última [32,49,57,59,99–104]. La opinión de los estudiantes sobre el

uso de la MV es muy positiva; siendo la fácil accesibilidad a las preparaciones la

característica mejor valorada [35,37]. Otras ventajas con buena aceptación por parte

del alumnado fueron la facilidad de uso, el fomento de la cooperación entre

estudiantes y el autoaprendizaje. La realización de los exámenes mediante MV

también destaca por sus ventajas, pues se valora realmente el conocimiento de los

alumnos, el material se encuentra homogeneizado y con una buena calidad de imagen

en todo momento.


[113]

El resultado final de las evaluaciones en los diferentes estudios o bien no mostró

diferencias entre los estudiantes de pre-grado que se prepararon usando MV y MC, o

bien resultó favorable al uso de MV, pues estos estudiantes parecían reconocer

patrones histológicos con mayor facilidad debido a que no tenían que centrarse en el

aprendizaje del uso del microscopio convencional [105–108]

En el caso de la docencia de post-grado, los estudios también ponen de

manifiesto unos resultados favorables al uso de la MV. La MV permite crear series de

casos interesantes que ayudan a unificar y homogenizar el aprendizaje,

independientemente del centro donde se encuentre el patólogo en formación,

preservando a su vez el material histológico.

Algunos estudios indican que el uso de preparaciones digitales con anotaciones

previamente creadas por los docentes hace que el aprendizaje sea más dirigido a las

características histológicas que llevan al diagnóstico, haciendo el proceso de

aprendizaje más eficiente y mejorando las puntuaciones finales de los exámenes [109].

La MV aplicada al aprendizaje de citopatología también presenta muy buenos

resultados en distintos estudios, solucionando los problemas técnicos que impedían la

homogeneización de las preparaciones. En este ámbito, el uso de anotaciones parece

mejorar aún más los resultados finales de los examenes [110].

Existen otros usos de la MV aplicados a la docencia que incluyen “tumor boards”

y tele-aprendizaje; este último permite tanto a patologos en formación como a

especialistas disponer de series de casos interesantes destinados a la formación

continuada [111,112].

En resumen, el estudio 4 ha puesto de manifiesto que a pesar de que en el

momento actual el uso de la MV en la docencia suele ser complementario a la MC,

especialmente en el caso de la educación de post-grado, existen sufientes evidencias

de que puede sustituir totalmente a la MC sin que el proceso de aprendizaje se vea

afectado negativamente. Además, la MV presenta numerosas ventajas y mejora la

homogeneización del material destinado al estudio.


[114]

Los buenos resultados puestos de manifiesto en este estudio y las ventajas de la

MV propiciaron la implementación de esta tecnología en la Universidad de Barcelona,

con el objetivo de realizar las clases de Anatomía Patológica a alumnos de la Facultad

de Medicina. Durante este proceso realizamos el estudio 5 con el fin de documentar la

transición de MC a MV y para validar la aplicación de esta última la docencia de pre-

grado, así como recoger su valoración por parte del alumnado.

Los resultados confirman que la MV puede reemplazar de forma efectiva a la MC

en la docencia de estudiantes de pre-grado, y muestran que las habilidades alcanzadas

con MV son comparables a aquellas alcanzadas con MC. La aceptación de esta nueva

tecnología por parte de los estudiantes fue muy positiva, resaltando el fácil uso del

programa informático. La característica mejor valorada de la MV fue la posibilidad de

acceder a todas las imágenes en cualquier momento y en cualquier ubicación. Esta

ventaja fue remarcada por los datos derivados de la auditoría del acceso al navegador,

mostrando que sobre la mitad de los accesos se llevaron a cabo en vacaciones y un

tercio fuera de horas lectivas.

La introducción de la MV se tradujo en una disminución del trabajo realizado por

los estudiantes y docentes en términos de tiempo empleado en las clases, aunque el

tiempo de preparación del material se vio incrementado de forma significativa.

En nuestro estudio la transición entre MC y MV no se realizó progresivamente,

sino que se llevó a cabo de forma abrupta, demostrando que los resultados positivos

se pueden ver de forma inmediata.

Las recientes reformas en los programas docentes se centran en la reducción de

horas de clase y prácticas presenciales, para dar mayor énfasis al autoaprendizaje, el

desarrollo de habilidades personales y la capacidad para la resolución de problemas

[104]. Con el fin de alcanzar estos objetivos el tiempo empleado en actividades más

tradicionales ha sido reubicado en nuevas actividades más importantes; lo que ha

dado paso a que en algunas facultades de medicina hayan disminuido

considerablemente las oportunidades de los estudiantes de aprender ciencias médicas

básicas [57,104,113,114]. Por esta razón resulta necesario establecer nuevas

estrategias, con el objetivo de mejorar el aprendizaje de Anatomía Patológica. La


[115]

introducción de la MV es una alternativa adecuada a los sistemas de aprendizaje

tradicionales, pudiendo ayudar a los estudiantes a alcanzar un conocimiento

satisfactorio en estas disciplinas básicas.

Nuestro estudio no mostró diferencias significativas en los resultados de los

exámenes entre el grupo que realizó las practicas con MC y el que lo hizo con MV, con

la ventaja adicional de que los estudiantes se sintieron más cómodos con el uso de la

MV para el aprendizaje. Para los docentes, la MV no solo disminuyó el tiempo

empleado en las clases, sino que también ayudó a conocer cómo es el proceso de

aprendizaje de los alumnos de Anatomía Patológica, cómo de instructivas son las

clases con MC y qué preparaciones resultan realmente didácticas.

La VM imita el funcionamiento de la MC, con la ventaja de que las preparaciones

se encuentran siempre enfocadas y con un óptimo contraste e iluminación. Además,

hasta un 70% de los estudiantes encontraron más sencilla la navegación de las

preparaciones digitales que las convencionales con microscopio óptico.

Las encuestas anónimas pusieron de manifiesto que la mayoría de estudiantes

encontraba útil la MV; este resultado concuerda con el de otros estudios donde la MV

también fue valorada de forma positiva por parte del alumnado [104]. Esto pone de

manifiesto los beneficios de la aplicación de la MV a la docencia, pues permite que los

estudiantes exploren las preparaciones de forma independiente, controlando el

contenido y el ritmo de aprendizaje. Además, el uso generalizado de ordenadores

personales y la experiencia previa de los estudiantes con dispositivos informáticos,

hacen que la adaptación y el manejo de la MV resulte muy sencilla [104].

Los resultados son concordantes con los de otros estudios, recogidos en nuestro

estudio 4, los cuales ponen de manifiesto que esta tecnología hace que las prácticas de

laboratorio de microscopía sean más eficientes y hace posible la portabilidad de los

recursos docentes sin depender de calendarios académicos [30,37,57,104]. Además,

en nuestro estudio la característica mejor valorada de la MV fue la posibilidad de

acceder al sistema desde cualquier lugar y en cualquier momento, lo que quedó

confirmado con los datos de la auditoría del acceso al navegador.


[116]

A pesar de que el coste de la implementación de la MV es alto, este nuevo

sistema puede representar un gran avance en la forma de enseñar y aprender

Anatomía Patológica. La MV posibilita que el material más representativo y de mejor

calidad de cada caso pueda ser incluido, eliminando la posibilidad de deterioro o

pérdida de las preparaciones; aunque es necesario que las muestras pasen por un

proceso de anonimato previo, para respetar la confidencialidad de datos de los

pacientes. Otra de las principales ventajas de la MV es la facilidad de mantenimiento,

pues esta herramienta puede reducir e incluso eliminar los gastos destinados al

mantenimiento de los laboratorios de microscopía [2,38,42,57].

La principal fortaleza de nuestro estudio es la comparación de dos grupos muy

similares del mismo curso, que trabajaban con el mismo material; lo cual provee datos

objetivos sobre el efecto de la MV en el proceso de aprendizaje. Una posible limitación

de nuestro estudio es que los exámenes se realizaron usando imágenes estáticas en

vez de preparaciones virtuales que permitan la navegación, pues se requieren

programas informáticos más sofisticados para poder realizarlos de esta manera.

Aunque hay que tener en consideración que en ambos grupos los exámenes se

realizaron con imágenes provenientes de los mismos casos. Por último, cabe destacar

que las altas calificaciones alcanzadas por los estudiantes son consecuencia de las altas

notas necesarias para ingresar en la Facultad de Medicina.

En resumen, la evidencia muestra que las habilidades en microscopía alcanzadas

con MV son comparables con aquellas alcanzadas con MC, lo que indica que esta

tecnología puede reemplazar de forma satisfactoria los métodos de aprendizaje

tradicionales, aportando a su vez mayor portabilidad entre otras ventajas.

La conclusión final de nuestros estudios es que la MV es una herramienta capaz

de sustituir a la MC tanto en la docencia de pre y post-grado, como en el diagnóstico

rutinario de biopsias. Existen suficientes estudios de validación que aseguran la no

inferioridad de la MV respecto a la MC en docencia y en el diagnóstico primario de

gran parte de las áreas de la Anatomía Patológica; siendo necesaria la realización de

nuevos estudios de validación que incluyan algunas subespecialidades que no

disponen de los mismos. La expansión de esta tecnología por los Servicios de Anatomía


[117]

Patológica y las Facultades hará que cada vez más centros puedan disponer de sus

múltiples ventajas.


[118]


[119]

VI. Conclusiones


[120]


[121]

1. Existe una muy buena correlación entre los diagnósticos realizados con MC y

MV en gran parte de las subespecialidades de la Anatomía Patológica, por lo

que se puede considerar que la implementación de esta tecnología para el

diagnóstico primario es segura (estudio 1)

2. Existe una ausencia total de estudios de validación en algunas áreas de la

Anatomía Patológica como en la patología ginecológica, la patología hepática,

la hematopatología, la patología ósea, endocrina y de partes blandas, razón por

la cual es necesario realizar estudios de validación que incluyan muestras de

estas subespecialidades (estudio 1)

3. La MV puede ser utilizada de forma segura en el diagnóstico histológico

primario de la patología ginecológica, puesto que la concordancia inter-

observador entre los diagnósticos obtenidos con la MV y la MC es muy alta

(estudio 2)

4. Existe una curva de aprendizaje en el uso de la MV, pues la experiencia se

traduce en una mejoría de los resultados y permite a los patólogos diagnosticar

con seguridad prácticamente todas las muestras usando exclusivamente la MV

(estudio 2)

5. El escaneo a 200x aumentos es adecuado para realizar un diagnóstico de la

mayoría de las biopsias ginecológicas de forma fiable (estudio 2)

6. La MV puede ser usada de forma segura para el diagnóstico histológico

primario de biopsias hepáticas con aguja, tanto en hígado nativo como en

hígado de trasplante, mostrando unos altos índices de concordancia tanto intra

como inter-observador (estudio 3)

7. El escaneo a 400x aumentos ha demostrado ser adecuado para realizar el

diagnóstico de las biopsias hepáticas con aguja de forma satisfactoria (estudio

3)


[122]

8. Existe suficiente evidencia de que la MV es una herramienta adecuada para la

docencia de pre y post-grado tanto en los estudios de Medicina, como de

Veterinaria u Odontología (estudio 4)

9. Las habilidades en microscopía alcanzadas por los estudiantes de pregrado de

Medicina usando la MV son comparables con las alcanzadas con la MC.

Además, este cambio puede realizarse de forma rápida (estudio 5)

10. Los alumnos valoran de forma positiva la implementación de la MV en la

docencia. Las principales ventajas reconocidas por ellos son el fácil uso del

programa informático y la posibilidad de realizar las prácticas a cualquier hora y

desde cualquier lugar (estudio 5)


[123]

VII. Bibliografía


[124]


[125]

[1] Weinstein RS. Prospects for telepathology. Hum Pathol 1986;17:433–4.

[2] Al-Janabi S, Huisman AAA, Van Diest PJ. Digital pathology: Current status and

future perspectives. Histopathology 2012;61:1–9. doi:10.1111/j.1365-

2559.2011.03814.x.

[3] Pantanowitz L, Sinard JH, Henricks WH, Fatheree LA, Carter AB, Contis L, et al.

Validating whole slide imaging for diagnostic purposes in pathology: guideline

from the College of American Pathologists Pathology and Laboratory Quality

Center. Arch Pathol Lab Med 2013;137:1710–22. doi:10.5858/arpa.2013-0093-

CP.

[4] Cross SS, Dennis T, Start RD. Telepathology: current status and future prospects

in diagnostic histopathology. Histopathology 2002;41:91–109.

[5] Evans AJ, Kiehl T-R, Croul S. Frequently asked questions concerning the use of

whole-slide imaging telepathology for neuropathology frozen sections. Semin

Diagn Pathol 2010;27:160–6.

[6] Evans AJ, Chetty R, Clarke BA, Croul S, Ghazarian DM, Kiehl T-R, et al. Primary

frozen section diagnosis by robotic microscopy and virtual slide telepathology:

the University Health Network experience. Semin Diagn Pathol 2009;26:165–76.

[7] Fallon MA, Wilbur DC, Prasad M. Ovarian frozen section diagnosis: use of whole-

slide imaging shows excellent correlation between virtual slide and original

interpretations in a large series of cases. Arch Pathol Lab Med 2010;134:1020–3.

doi:10.1043/2009-0320-OA.1.

[8] Gould P V, Saikali S. A comparison of digitized frozen section and smear

preparations for intraoperative neurotelepathology. Anal Cell Pathol (Amst)


[126]

2012;35:85–91. doi:10.3233/ACP-2011-0026.

[9] Kaplan KJ, Burgess JR, Sandberg GD, Myers CP, Bigott TR, Greenspan RB. Use of

robotic telepathology for frozen-section diagnosis: a retrospective trial of a

telepathology system for intraoperative consultation. Mod Pathol an Off J

United States Can Acad Pathol Inc 2002;15:1197–204.

doi:10.1097/01.MP.0000033928.11585.42.

[10] Pantanowitz L, Dickinson K, Evans AJ, Hassell LA, Henricks WH, Lennerz JK, et al.

American Telemedicine Association clinical guidelines for telepathology. J Pathol

Inform 2014;5:39. doi:10.4103/2153-3539.143329.

[11] Slodkowska J, Pankowski J, Siemiatkowska K, Chyczewski L. Use of the virtual

slide and the dynamic real-time telepathology systems for a consultation and

the frozen section intra-operative diagnosis in thoracic/pulmonary pathology.

Folia Histochem Cytobiol 2009;47:679–84. doi:10.2478/v10042-010-0009-z.

[12] Wilbur DC. Digital pathology: get on board-the train is leaving the station.

Cancer Cytopathol 2014;122:791–5. doi:10.1002/cncy.21479.

[13] Ayad E. Virtual telepathology in Egypt, applications of WSI in Cairo University.

Diagn Pathol 2011;6 Suppl 1:S1. doi:10.1186/1746-1596-6-S1-S1.

[14] Romero Lauro G, Cable W, Lesniak A, Tseytlin E, McHugh J, Parwani A, et al.

Digital pathology consultations-a new era in digital imaging, challenges and

practical applications. J Digit Imaging 2013;26:668–77. doi:10.1007/s10278-013-

9572-0.

[15] Wienert S, Beil M, Saeger K, Hufnagl P, Schrader T. Integration and acceleration

of virtual microscopy as the key to successful implementation into the routine


[127]

diagnostic process. Diagn Pathol 2009;4:3. doi:10.1186/1746-1596-4-3.

[16] Wilbur DC, Madi K, Colvin RB, Duncan LM, Faquin WC, Ferry JA, et al. Whole-

slide imaging digital pathology as a platform for teleconsultation: a pilot study

using paired subspecialist correlations. Arch Pathol Lab Med 2009;133:1949–53.

doi:10.1043/1543-2165-133.12.1949.

[17] Bernard C, Chandrakanth SA, Cornell IS, Dalton J, Evans A, Garcia BM, et al.

Guidelines from the Canadian Association of Pathologists for establishing a

telepathology service for anatomic pathology using whole-slide imaging. J Pathol

Inform 2014;5:15. doi:10.4103/2153-3539.129455.

[18] Hedvat C V. Digital microscopy: past, present, and future. Arch Pathol Lab Med

2010;134:1666–70. doi:10.1043/2009-0579-RAR1.1.

[19] Ho J, Ahlers SM, Stratman C, Aridor O, Pantanowitz L, Fine JL, et al. Can digital

pathology result in cost savings? A financial projection for digital pathology

implementation at a large integrated health care organization. J Pathol Inform

2014;5:33. doi:10.4103/2153-3539.139714.

[20] Isaacs M, Lennerz JK, Yates S, Clermont W, Rossi J, Pfeifer JD. Implementation of

whole slide imaging in surgical pathology: A value added approach. J Pathol

Inform 2011;2:39. doi:10.4103/2153-3539.84232.

[21] Pantanowitz L. Digital images and the future of digital pathology. J Pathol Inform

2010;1. doi:10.4103/2153-3539.68332.

[22] Thorstenson S, Molin J, Lundstrom C. Implementation of large-scale routine

diagnostics using whole slide imaging in Sweden: Digital pathology experiences

2006-2013. J Pathol Inform 2014;5:14. doi:10.4103/2153-3539.129452.


[128]

[23] Hartman DJ, Parwani A V, Cable B, Cucoranu IC, McHugh JS, Kolowitz BJ, et al.

Pocket pathologist: A mobile application for rapid diagnostic surgical pathology

consultation. J Pathol Inform 2014;5:10. doi:10.4103/2153-3539.129443.

[24] Roy S, Pantanowitz L, Amin M, Seethala RR, Ishtiaque A, Yousem SA, et al.

Smartphone adapters for digital photomicrography. J Pathol Inform 2014;5:24.

doi:10.4103/2153-3539.137728.

[25] Speiser JJ, Hughes I, Mehta V, Wojcik EM, Hutchens KA. Mobile

teledermatopathology: using a tablet PC as a novel and cost-efficient method to

remotely diagnose dermatopathology cases. Am J Dermatopathol 2014;36:54–7.

doi:10.1097/DAD.0b013e3182863186.

[26] Gavrielides MA, Conway C, O’Flaherty N, Gallas BD, Hewitt SM. Observer

performance in the use of digital and optical microscopy for the interpretation

of tissue-based biomarkers. Anal Cell Pathol (Amst) 2014;2014:157308.

doi:10.1155/2014/157308.

[27] Nassar A, Cohen C, Agersborg SS, Zhou W, Lynch KA, Barker EA, et al. A multisite

performance study comparing the reading of immunohistochemical slides on a

computer monitor with conventional manual microscopy for estrogen and

progesterone receptor analysis. Am J Clin Pathol 2011;135:461–7.

doi:10.1309/AJCP4VFKA5FCMZNA.

[28] Micsik T, Kiszler G, Szabo D, Krecsak L, Hegedus C, Tibor K, et al. Computer Aided

Semi-Automated Evaluation of HER2 Immunodetection--A Robust Solution for

Supporting the Accuracy of Anti HER2 Therapy. Pathol Oncol Res 2015;21:1005–

11. doi:10.1007/s12253-015-9927-6.


[129]

[29] Krenacs T, Zsakovics I, Diczhazi C, Ficsor L, Varga VS, Molnar B. The potential of

digital microscopy in breast pathology. Pathol Oncol Res 2009;15:55–8.

doi:10.1007/s12253-008-9087-z.

[30] Boutonnat J, Paulin C, Faure C, Colle PE, Ronot X, Seigneurin D. A pilot study in

two French medical schools for teaching histology using virtual microscopy.

Morphologie 2006;90:21–5.

[31] Braun MW, Kearns KD. Improved learning efficiency and increased student

collaboration through use of virtual microscopy in the teaching of human

pathology. Anat Sci Educ 2008;1:240–6. doi:10.1002/ase.53.

[32] Linder E, Lundin M, Thors C, Lebbad M, Winiecka-Krusnell J, Helin H, et al. Web-

based virtual microscopy for parasitology: A novel tool for education and quality

assurance. PLoS Negl Trop Dis 2008;2:e315. doi:10.1371/journal.pntd.0000315.

[33] Paulsen FP, Eichhorn M, Brauer L, Bräuer L. Virtual microscopy-The future of

teaching histology in the medical curriculum? Ann Anat 2010;192:378–82.

doi:10.1016/j.aanat.2010.09.008.

[34] Foster K. Medical education in the digital age: Digital whole slide imaging as an

e-learning tool. J Pathol Inform 2010;1:38–40. doi:10.4103/2153-3539.68331.

[35] Merk M, Knuechel R, Perez-Bouza A. Web-based virtual microscopy at the RWTH

Aachen University: Didactic concept, methods and analysis of acceptance by the

students. Ann Anat 2010;192:383–7. doi:10.1016/j.aanat.2010.01.008.

[36] Collier L, Dunham S, Braun MW, O’Loughlin VD. Optical versus virtual: Teaching

assistant perceptions of the use of virtual microscopy in an undergraduate

human anatomy course. Anat Sci Educ 2012;5:10–9. doi:10.1002/ase.262.


[130]

[37] Szymas J, Lundin M. Five years of experience teaching pathology to dental

students using the WebMicroscope. Diagn Pathol 2011;6:S13.

doi:10.1186/1746-1596-6-S1-S13.

[38] Pantanowitz L, Szymas J, Yagi Y, Wilbur D. Whole slide imaging for educational

purposes. J Pathol Inform 2012;3:46. doi:10.4103/2153-3539.104908.

[39] Husmann PR, O’Loughlin VD, Braun MW. Quantitative and qualitative changes in

teaching histology by means of virtual microscopy in an introductory course in

human anatomy. Anat Sci Educ 2009;2:218–26. doi:10.1002/ase.105.

[40] Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F. Comparison of a virtual

microscope laboratory to a regular microscope laboratory for teaching

histology. Anat Rec 2001;265:10–4.

[41] Cornish TC, Swapp RE, Kaplan KJ. Whole-slide imaging: routine pathologic

diagnosis. Adv Anat Pathol 2012;19:152–9.

doi:10.1097/PAP.0b013e318253459e.

[42] Pantanowitz L, Valenstein PN, Evans AJ, Kaplan KJ, Pfeifer JD, Wilbur DC, et al.

Review of the current state of whole slide imaging in pathology. J Pathol Inform

2011;2:36. doi:10.4103/2153-3539.83746.

[43] Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJA, Ten Kate FJW, et al.

Whole slide images for primary diagnostics in dermatopathology: a feasibility

study. J Clin Pathol 2012;65:152–8. doi:10.1136/jclinpath-2011-200277.

[44] Randell R, Ruddle RA, Mello-Thoms C, Thomas RG, Quirke P, Treanor D. Virtual

reality microscope versus conventional microscope regarding time to diagnosis:

an experimental study. Histopathology 2013;62:351–8. doi:10.1111/j.1365-


[131]

2559.2012.04323.x.

[45] Krishnamurthy S, Mathews K, McClure S, Murray M, Gilcrease M, Albarracin C,

et al. Multi-institutional comparison of whole slide digital imaging and optical

microscopy for interpretation of hematoxylin-eosin-stained breast tissue

sections. Arch Pathol Lab Med 2013;137:1733–9. doi:10.5858/arpa.2012-0437-

OA.

[46] Houghton JP, Ervine AJ, Kenny SL, Kelly PJ, Napier SS, McCluggage WG, et al.

Concordance between digital pathology and light microscopy in general surgical

pathology: a pilot study of 100 cases. J Clin Pathol 2014;67:1052–5.

doi:10.1136/jclinpath-2014-202491.

[47] Sanders DSA, Grabsch H, Harrison R, Bateman A, Going J, Goldin R, et al.

Comparing virtual with conventional microscopy for the consensus diagnosis of

Barrett’s neoplasia in the AspECT Barrett’s chemoprevention trial pathology

audit. Histopathology 2012;61:795–800. doi:10.1111/j.1365-2559.2012.04288.x.

[48] Randell R, Ruddle RA, Thomas RG, Mello-Thoms C, Treanor D. Diagnosis of major

cancer resection specimens with virtual slides: impact of a novel digital

pathology workstation. Hum Pathol 2014;45:2101–6.

doi:10.1016/j.humpath.2014.06.017.

[49] Fonseca F-P, Santos-Silva A-R, Lopes M-A, Almeida O-P de, Vargas P-A.

Transition from glass to digital slide microscopy in the teaching of oral pathology

in a Brazilian dental school. Med Oral Patol Oral Cir Bucal 2015;20:e17-22.

[50] Kumar RK, Velan GM, Korell SO, Kandara M, Dee FR, Wakefield D. Virtual

microscopy for learning and assessment in pathology. J Pathol 2004;204:613–8.


[132]

doi:10.1002/path.1658.

[51] Ho J, Parwani A V, Jukic DM, Yagi Y, Anthony L, Gilbertson JR. Use of whole slide

imaging in surgical pathology quality assurance: design and pilot validation

studies. Hum Pathol 2006;37:322–31.

[52] Garcia Rojo M, Felix Conde A, Ordi J, Ruiz Martin J, Corominas JM, Alvarez

Alegret R, et al. Libro blanco de la Anatomía Patológica en España. 2015.

[53] Winokur TS, McClellan S, Siegal GP, Redden D, Gore P, Lazenby A, et al. A

prospective trial of telepathology for intraoperative consultation (frozen

sections). Hum Pathol 2000;31:781–5. doi:10.1053/hupa.2000.8452.

[54] Neel JA, Grindem CB, Bristol DG. Introduction and evaluation of virtual

microscopy in teaching veterinary cytopathology. J Vet Med Educ 2007;34:437–

44. doi:10.3138/jvme.34.4.437.

[55] Mills PC, Bradley AP, Woodall PF, Wildermoth M. Teaching histology to first-year

veterinary science students using virtual microscopy and traditional microscopy:

a comparison of student responses. J Vet Med Educ 2007;34:177–82.

[56] Dee FR, Meyerholz DK. Teaching medical pathology in the twenty-first century:

virtual microscopy applications. J Vet Med Educ 2007;34:431–6.

doi:10.3138/jvme.34.4.431.

[57] Blake C a, Lavoie HA, Millette CF. Teaching medical histology at the University of

South Carolina School of Medicine: Transition to virtual slides and virtual

microscopes. Anat Rec B New Anat 2003;275:196–206. doi:10.1002/ar.b.10037.

[58] Goldberg HR, Dintzis R. The positive impact of team-based virtual microscopy on

student learning in physiology and histology. AJP Adv Physiol Educ 2007;31:261–


[133]

5. doi:10.1152/advan.00125.2006.

[59] Chen Y-K, Hsue S-S, Lin D-C, Wang W-C, Chen J-Y, Lin C-C, et al. An application of

virtual microscopy in the teaching of an oral and maxillofacial pathology

laboratory course. Oral Surg Oral Med Oral Pathol Oral Radiol Endod

2008;105:342–7. doi:10.1016/j.tripleo.2007.03.020.

[60] Wellnitz U, Fritz P, Voudouri V, Linder A, Toomes H, Schmid J, et al. The validity

of telepathological frozen section diagnosis with ISDN-mediated remote

microscopy. Virchows Arch 2000;437:52–7.

[61] Oberholzer M, Fischer HR, Christen H, Gerber S, Bruhlmann M, Mihatsch MJ, et

al. Telepathology: frozen section diagnosis at a distance. Virchows Arch

1995;426:3–9.

[62] Della Mea V, Cataldi P, Pertoldi B, Beltrami CA. Combining dynamic and static

robotic telepathology: a report on 184 consecutive cases of frozen sections,

histology and cytology. Anal Cell Pathol 2000;20:33–9.

[63] Dawson PJ, Johnson JG, Edgemon LJ, Brand CR, Hall E, Van Buskirk GF.

Outpatient frozen sections by telepathology in a Veterans Administration

medical center. Hum Pathol 2000;31:786–8. doi:10.1053/hupa.2000.8451.

[64] Becker RLJ, Specht CS, Jones R, Rueda-Pedraza ME, O’Leary TJ. Use of remote

video microscopy (telepathology) as an adjunct to neurosurgical frozen section

consultation. Hum Pathol 1993;24:909–11.

[65] Adachi H, Inoue J, Nozu T, Aoki H, Ito H. Frozen-section services by

telepathology: experience of 100 cases in the San-in District, Japan. Pathol Int

1996;46:436–41.


[134]

[66] Nakayama I, Matsumura T, Kamataki A, Uzuki M, Saito K, Hobbs J, et al.

Development of a teledermatopathology consultation system using virtual

slides. Diagn Pathol 2012;7:177. doi:10.1186/1746-1596-7-177.

[67] Jones NC, Nazarian RM, Duncan LM, Kamionek M, Lauwers GY, Tambouret RH,

et al. Interinstitutional whole slide imaging teleconsultation service

development: assessment using internal training and clinical consultation cases.

Arch Pathol Lab Med 2015;139:627–35. doi:10.5858/arpa.2014-0133-OA.

[68] Song Y, Treanor D, Bulpitt AJ, Magee DR. 3D reconstruction of multiple stained

histology images. J Pathol Inform 2013;4:S7. doi:10.4103/2153-3539.109864.

[69] Helin HH, Lundin M, Lundin J, Martikainen P, Tammela T, Helin HH, et al. Web-

based virtual microscopy in teaching and standardizing Gleason grading. Hum

Pathol 2005;36:381–6. doi:10.1016/j.humpath.2005.01.020.

[70] Caie PD, Turnbull AK, Farrington SM, Oniscu A, Harrison DJ. Quantification of

tumour budding, lymphatic vessel density and invasion through image analysis

in colorectal cancer. J Transl Med 2014;12:156. doi:10.1186/1479-5876-12-156.

[71] Neil DAH, Roberts ISD, Bellamy COC, Wigmore SJ, Neuberger JM. Improved

access to histopathology using a digital system could increase the organ donor

pool and improve allocation. Transpl Int 2014;27:759–64. doi:10.1111/tri.12320.

[72] Neltner JH, Abner EL, Schmitt FA, Denison SK, Anderson S, Patel E, et al. Digital

pathology and image analysis for robust high-throughput quantitative

assessment of Alzheimer disease neuropathologic changes. J Neuropathol Exp

Neurol 2012;71:1075–85. doi:10.1097/NEN.0b013e3182768de4.

[73] Riber-Hansen R, Vainer B, Steiniche T. Digital image analysis: a review of


[135]

reproducibility, stability and basic requirements for optimal results. APMIS

2012;120:276–89. doi:10.1111/j.1600-0463.2011.02854.x.

[74] Webster JD, Dunstan RW. Whole-slide imaging and automated image analysis:

considerations and opportunities in the practice of pathology. Vet Pathol

2014;51:211–23. doi:10.1177/0300985813503570.

[75] Nelson D, Ziv A, Bandali KS. Going glass to digital: virtual microscopy as a

simulation-based revolution in pathology and laboratory science. J Clin Pathol

2012;65:877–81. doi:10.1136/jclinpath-2012-200665.

[76] Rodriguez-Urrego PA, Cronin AM, Al-Ahmadie HA, Gopalan A, Tickoo SK, Reuter

VE, et al. Interobserver and intraobserver reproducibility in digital and routine

microscopic assessment of prostate needle biopsies. Hum Pathol 2011;42:68–

74. doi:10.1016/j.humpath.2010.07.001.

[77] Robertson AJ, Anderson JM, Beck JS, Burnett RA, Howatson SR, Lee FD, et al.

Observer variability in histopathological reporting of cervical biopsy specimens. J

Clin Pathol 1989;42:231–8.

[78] McCluggage WG, Walsh MY, Thornton CM, Hamilton PW, Date A, Caughley LM,

et al. Inter- and intra-observer variation in the histopathological reporting of

cervical squamous intraepithelial lesions using a modified Bethesda grading

system. Br J Obstet Gynaecol 1998;105:206–10.

[79] McCluggage WG, Bharucha H, Caughley LM, Date A, Hamilton PW, Thornton

CM, et al. Interobserver variation in the reporting of cervical colposcopic biopsy

specimens: comparison of grading systems. J Clin Pathol 1996;49:833–5.

[80] de Vet HC, Koudstaal J, Kwee WS, Willebrand D, Arends JW. Efforts to improve


[136]

interobserver agreement in histopathological grading. J Clin Epidemiol

1995;48:869–73.

[81] Creagh T, Bridger JE, Kupek E, Fish DE, Martin-Bates E, Wilkins MJ. Pathologist

variation in reporting cervical borderline epithelial abnormalities and cervical

intraepithelial neoplasia. J Clin Pathol 1995;48:59–60.

[82] Bergeron C, Ordi J, Schmidt D, Trunk MJ, Keller T, Ridder R. Conjunctive

p16INK4a testing significantly increases accuracy in diagnosing high-grade

cervical intraepithelial neoplasia. Am J Clin Pathol 2010;133:395–406.

doi:10.1309/AJCPXSVCDZ3D5MZM.

[83] Stoler MH, Schiffman M. Interobserver reproducibility of cervical cytologic and

histologic interpretations: realistic estimates from the ASCUS-LSIL Triage Study.

JAMA 2001;285:1500–5.

[84] Bauer TW, Schoenfield L, Slaw RJ, Yerian L, Sun Z, Henricks WH. Validation of

whole slide imaging for primary diagnosis in surgical pathology. Arch Pathol Lab

Med 2013;137:518–24. doi:10.5858/arpa.2011-0678-OA.

[85] Camparo P, Egevad L, Algaba F, Berney DM, Boccon-Gibod L, Comperat E, et al.

Utility of whole slide imaging and virtual microscopy in prostate pathology.

APMIS 2012;120:298–304. doi:10.1111/j.1600-0463.2011.02872.x.

[86] Al-Janabi S, Huisman A, Vink A, Leguit RJ, Offerhaus GJA, ten Kate FJW, et al.

Whole slide images for primary diagnostics of gastrointestinal tract pathology: a

feasibility study. Hum Pathol 2012;43:702–7.

doi:10.1016/j.humpath.2011.06.017.

[87] Gilbertson JR, Ho J, Anthony L, Jukic DM, Yagi Y, Parwani A V. Primary histologic


[137]

diagnosis using automated whole slide imaging: a validation study. BMC Clin

Pathol 2006;6:4. doi:10.1186/1472-6890-6-4.

[88] Bauer TW, Slaw RJ. Validating whole-slide imaging for consultation diagnoses in

surgical pathology. Arch Pathol Lab Med 2014;138:1459–65.

doi:10.5858/arpa.2013-0541-OA.

[89] Al-Janabi S, Huisman A, Nikkels PGJ, ten Kate FJW, van Diest PJ. Whole slide

images for primary diagnostics of paediatric pathology specimens: a feasibility

study. J Clin Pathol 2013;66:218–23. doi:10.1136/jclinpath-2012-201104.

[90] Rousselet M-C, Michalak S, Dupre F, Croue A, Bedossa P, Saint-Andre J-P, et al.

Sources of variability in histological scoring of chronic viral hepatitis. Hepatology

2005;41:257–64. doi:10.1002/hep.20535.

[91] Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al.

Design and validation of a histological scoring system for nonalcoholic fatty liver

disease. Hepatology 2005;41:1313–21. doi:10.1002/hep.20701.

[92] Regev A, Berho M, Jeffers LJ, Milikowski C, Molina EG, Pyrsopoulos NT, et al.

Sampling error and intraobserver variation in liver biopsy in patients with

chronic HCV infection. Am J Gastroenterol 2002;97:2614–8. doi:10.1111/j.1572-

0241.2002.06038.x.

[93] Robert M, Sofair AN, Thomas A, Bell B, Bialek S, Corless C, et al. A comparison of

hepatopathologists’ and community pathologists’ review of liver biopsy

specimens from patients with hepatitis C. Clin Gastroenterol Hepatol

2009;7:335–8. doi:10.1016/j.cgh.2008.11.029.

[94] Skripenova S, Trainer TD, Krawitt EL, Blaszyk H. Variability of grade and stage in


[138]

simultaneous paired liver biopsies in patients with hepatitis C. J Clin Pathol

2007;60:321–4. doi:10.1136/jcp.2005.036020.

[95] Brick KE, Sluzevich JC, Cappel MA, DiCaudo DJ, Comfere NI, Wieland CN.

Comparison of virtual microscopy and glass slide microscopy among

dermatology residents during a simulated in-training examination. J Cutan

Pathol 2013;40:807–11. doi:10.1111/cup.12189.

[96] Reyes C, Ikpatt OF, Nadji M, Cote RJ. Intra-observer reproducibility of whole

slide imaging for the primary diagnosis of breast needle biopsies. J Pathol Inform

2014;5:5. doi:10.4103/2153-3539.127814.

[97] Comperat E, Egevad L, Lopez-Beltran A, Camparo P, Algaba F, Amin M, et al. An

interobserver reproducibility study on invasiveness of bladder cancer using

virtual microscopy and heatmaps. Histopathology 2013;63:756–66.

doi:10.1111/his.12214.

[98] Singson RP, Natarajan S, Greenson JK, Marchevsky AM. Virtual microscopy and

the Internet as telepathology consultation tools. A study of gastrointestinal

biopsy specimens. Am J Clin Pathol 1999;111:792–5.

[99] Farah CS, Maybury TS. The e-evolution of microscopy in dental education. J Dent

Educ 2009;73:942–9.

[100] Diaz-Perez JA, Raju S, Echeverri JH. Evaluation of a teaching strategy based on

integration of clinical subjects, virtual autopsy, pathology museum, and digital

microscopy for medical students. J Pathol Inform 2014;5:25. doi:10.4103/2153-

3539.137729.

[101] Gatumu MK, MacMillan FM, Langton PD, Headley PM, Harris JR. Evaluation of


[139]

usage of virtual microscopy for the study of histology in the medical, dental, and

veterinary undergraduate programs of a UK University. Anat Sci Educ

2014;7:389–98.

[102] Helle L, Nivala M, Kronqvist P, Gegenfurtner A, Bjork P, Saljo R. Traditional

microscopy instruction versus process-oriented virtual microscopy instruction: a

naturalistic experiment with control group. Diagn Pathol 2011;6 Suppl 1:S8.

doi:10.1186/1746-1596-6-S1-S8.

[103] McCready ZR, Jham BC. Dental students’ perceptions of the use of digital

microscopy as part of an oral pathology curriculum. J Dent Educ 2013;77:1624–

8.

[104] Weaker FJ, Herbert DC. Transition of a Dental Histology Course from Light to

Virtual Microscopy. J Dent Educ 2009;73:1213–21.

[105] Anyanwu GE, Agu AU, Anyaehie UB. Enhancing learning objectives by use of

simple virtual microscopic slides in cellular physiology and histology: impact and

attitudes. Adv Physiol Educ 2012;36:158–63. doi:10.1152/advan.00008.2012.

[106] Krippendorf BB, Lough J. Complete and rapid switch from light microscopy to

virtual microscopy for teaching medical histology. Anat Rec - Part B New Anat

2005;285:19–25. doi:10.1002/ar.b.20066.

[107] Scoville SA, Buskirk TD. Traditional and virtual microscopy compared

experimentally in a classroom setting. Clin Anat 2007;20:565–70.

doi:10.1002/ca.20440.

[108] Sivamalai S, Murthy SV, Gupta T Sen, Woolley T. Teaching pathology via online

digital microscopy: Positive learning outcomes for rurally based medical


[140]

students. Aust J Rural Health 2011;19:45–51. doi:10.1111/j.1440-

1584.2010.01176.x.

[109] Marsch AF, Espiritu B, Groth J, Hutchens KA. The effectiveness of annotated (vs.

non-annotated) digital pathology slides as a teaching tool during dermatology

and pathology residencies. J Cutan Pathol 2014;41:513–8.

doi:10.1111/cup.12328.

[110] Stewart J 3rd, Bevans-Wilkins K, Bhattacharya A, Ye C, Miyazaki K, Kurtycz DFI.

Virtual microscopy: an educator’s tool for the enhancement of cytotechnology

students’ locator skills. Diagn Cytopathol 2008;36:363–8. doi:10.1002/dc.20821.

[111] Hamilton PW, Wang Y, McCullough SJ. Virtual microscopy and digital pathology

in training and education. Apmis 2012;120:305–15. doi:10.1111/j.1600-

0463.2011.02869.x.

[112] Rosai J. Digital images of case reports and other articles. Int J Surg Pathol

2007;15:5. doi:10.1177/1084713806296004.

[113] Williams G, Lau A. Reform of undergraduate medical teaching in the United

Kingdom: a triumph of evangelism over common sense. BMJ 2004;329:92–4.

doi:10.1136/bmj.329.7457.92.

[114] Bloodgood RA, Ogilvie RW. Trends in histology laboratory teaching in United

States medical schools. Anat Rec B New Anat 2006;289:169–75.

doi:10.1002/ar.b.20111.


[141]


[142]

Microscopía virtual en el diagnóstico de rutina y la ...

Documents

Transcript of Microscopía virtual en el diagnóstico de rutina y la ...