Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods

This paper presents the results of an experimental study and the development of tools for automatic analysis and recognition of histological images in order to obtain quantitative estimates of the presence and degree of aggressiveness of prostate cancer in the commonly used Gleason and ISUP scales....

Full description

Saved in:
Bibliographic Details
Main Authors: V. A. Kovalev, D. M. Voynov, V. D. Malyshau, E. D. Lapo
Format: Article
Language:Russian
Published: National Academy of Sciences of Belarus, the United Institute of Informatics Problems 2021-01-01
Series:Informatika
Subjects:
Online Access:https://inf.grid.by/jour/article/view/1090
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849771481534300160
author V. A. Kovalev
D. M. Voynov
V. D. Malyshau
E. D. Lapo
author_facet V. A. Kovalev
D. M. Voynov
V. D. Malyshau
E. D. Lapo
author_sort V. A. Kovalev
collection DOAJ
description This paper presents the results of an experimental study and the development of tools for automatic analysis and recognition of histological images in order to obtain quantitative estimates of the presence and degree of aggressiveness of prostate cancer in the commonly used Gleason and ISUP scales. The input data consisted of 10 616 whole-slide histological images with the size of the largest side up to 100 000 pixels and22 089 of their image tiles of 256×256 pixels in size. Two solutions were chosen as the final ones. The first solution is based on sequential analysis of image fragments and includes feature extraction using the ResNet50 network and the subsequent generalization of particular recognition results using a small convolutional network. The second solution is based on the simultaneous analysis of the selected informative sections, presented in the form of an intermediate pseudo-image, and its subsequent recognition using an ensemble of four variants of convolutional networks with the EfficientNetB0 architecture. Being independently tested on an unknown image dataset that was not available for authors, these approaches achieved the prediction accuracy of 0,9277 according to the ISUP scale.
format Article
id doaj-art-07d2b5431ecb46fa93f8cbda6681ea59
institution DOAJ
issn 1816-0301
language Russian
publishDate 2021-01-01
publisher National Academy of Sciences of Belarus, the United Institute of Informatics Problems
record_format Article
series Informatika
spelling doaj-art-07d2b5431ecb46fa93f8cbda6681ea592025-08-20T03:02:37ZrusNational Academy of Sciences of Belarus, the United Institute of Informatics ProblemsInformatika1816-03012021-01-01174486010.37661/1816-0301-2020-17-4-48-60946Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methodsV. A. Kovalev0D. M. Voynov1V. D. Malyshau2E. D. Lapo3The United Institute of Informatics Problems of the National Academy of Sciences of Belarus; Belarusian State UniversityThe United Institute of Informatics Problems of the National Academy of Sciences of Belarus; Belarusian State UniversityThe United Institute of Informatics Problems of the National Academy of Sciences of Belarus; Belarusian State UniversityThe United Institute of Informatics Problems of the National Academy of Sciences of Belarus; Belarusian State UniversityThis paper presents the results of an experimental study and the development of tools for automatic analysis and recognition of histological images in order to obtain quantitative estimates of the presence and degree of aggressiveness of prostate cancer in the commonly used Gleason and ISUP scales. The input data consisted of 10 616 whole-slide histological images with the size of the largest side up to 100 000 pixels and22 089 of their image tiles of 256×256 pixels in size. Two solutions were chosen as the final ones. The first solution is based on sequential analysis of image fragments and includes feature extraction using the ResNet50 network and the subsequent generalization of particular recognition results using a small convolutional network. The second solution is based on the simultaneous analysis of the selected informative sections, presented in the form of an intermediate pseudo-image, and its subsequent recognition using an ensemble of four variants of convolutional networks with the EfficientNetB0 architecture. Being independently tested on an unknown image dataset that was not available for authors, these approaches achieved the prediction accuracy of 0,9277 according to the ISUP scale.https://inf.grid.by/jour/article/view/1090prostate cancerhistologywhole slide histologydeep learningconvolutional neural networks
spellingShingle V. A. Kovalev
D. M. Voynov
V. D. Malyshau
E. D. Lapo
Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods
Informatika
prostate cancer
histology
whole slide histology
deep learning
convolutional neural networks
title Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods
title_full Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods
title_fullStr Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods
title_full_unstemmed Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods
title_short Computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods
title_sort computerized diagnosis of prostate cancer based on whole slide histology images and deep learning methods
topic prostate cancer
histology
whole slide histology
deep learning
convolutional neural networks
url https://inf.grid.by/jour/article/view/1090
work_keys_str_mv AT vakovalev computerizeddiagnosisofprostatecancerbasedonwholeslidehistologyimagesanddeeplearningmethods
AT dmvoynov computerizeddiagnosisofprostatecancerbasedonwholeslidehistologyimagesanddeeplearningmethods
AT vdmalyshau computerizeddiagnosisofprostatecancerbasedonwholeslidehistologyimagesanddeeplearningmethods
AT edlapo computerizeddiagnosisofprostatecancerbasedonwholeslidehistologyimagesanddeeplearningmethods