Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Glioma subtype classification from histopathological images using in-domain and out-of-domain transfer learning: An experimental study

Vladimir Despotovic, Sang-Yoon Kim, Ann-Christin Hau, Aliaksandra Kakoichankava, Gilbert Georg Klamminger, Felix Bruno Kleine Borgmann, Katrin B. M. Frauenknecht, Michel Mittelbronn, Petr V. Nazarov

TL;DR

A semi-supervised learning approach is proposed, where the fine-tuned models are utilized to predict the labels of unannotated regions of the whole slide images (WSI), providing insights to pathologists concerning the most informative parts of the WSI.

Abstract

We provide in this paper a comprehensive comparison of various transfer learning strategies and deep learning architectures for computer-aided classification of adult-type diffuse gliomas. We evaluate the generalizability of out-of-domain ImageNet representations for a target domain of histopathological images, and study the impact of in-domain adaptation using self-supervised and multi-task learning approaches for pretraining the models using the medium-to-large scale datasets of histopathological images. A semi-supervised learning approach is furthermore proposed, where the fine-tuned models are utilized to predict the labels of unannotated regions of the whole slide images (WSI). The models are subsequently retrained using the ground-truth labels and weak labels determined in the previous step, providing superior performance in comparison to standard in-domain transfer learning with balanced accuracy of 96.91% and F1-score 97.07%, and minimizing the pathologist's efforts for annotation. Finally, we provide a visualization tool working at WSI level which generates heatmaps that highlight tumor areas; thus, providing insights to pathologists concerning the most informative parts of the WSI.

Glioma subtype classification from histopathological images using in-domain and out-of-domain transfer learning: An experimental study

TL;DR

A semi-supervised learning approach is proposed, where the fine-tuned models are utilized to predict the labels of unannotated regions of the whole slide images (WSI), providing insights to pathologists concerning the most informative parts of the WSI.

Abstract

We provide in this paper a comprehensive comparison of various transfer learning strategies and deep learning architectures for computer-aided classification of adult-type diffuse gliomas. We evaluate the generalizability of out-of-domain ImageNet representations for a target domain of histopathological images, and study the impact of in-domain adaptation using self-supervised and multi-task learning approaches for pretraining the models using the medium-to-large scale datasets of histopathological images. A semi-supervised learning approach is furthermore proposed, where the fine-tuned models are utilized to predict the labels of unannotated regions of the whole slide images (WSI). The models are subsequently retrained using the ground-truth labels and weak labels determined in the previous step, providing superior performance in comparison to standard in-domain transfer learning with balanced accuracy of 96.91% and F1-score 97.07%, and minimizing the pathologist's efforts for annotation. Finally, we provide a visualization tool working at WSI level which generates heatmaps that highlight tumor areas; thus, providing insights to pathologists concerning the most informative parts of the WSI.
Paper Structure (19 sections, 3 figures, 8 tables)

This paper contains 19 sections, 3 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. Material and methods
  4. Dataset and study design
  5. Data preprocessing
  6. Transfer learning
  7. Out-of-domain transfer learning
  8. In-domain transfer learning
  9. Semi-supervised transfer learning
  10. Performance metrics
  11. Results
  12. Out-of-domain transfer learning
  13. In-domain transfer learning
  14. Semi-supervised in-domain transfer learning
  15. Quantitative localization of diffuse gliomas in Whole Slide Images
  16. ...and 4 more sections

Figures (3)

  • Figure 1: Confusion matrix for the best performing in-domain transfer learning model (semi-supervised ResNet50+ViT).
  • Figure 2: Per class performance for the best performing in-domain transfer learning model (semi-supervised ResNet50+ViT).
  • Figure 3: WSI level prediction (semi-supervised ResNet50+ViT) (a) IDH-wildtype (glioblastoma); (b) IDH-mutant, 1p/19q codeleted (oligodendroglioma). *Correct class