Alert button
Picture for Willem Vreuls

Willem Vreuls

Alert button

Automated Scoring of Nuclear Pleomorphism Spectrum with Pathologist-level Performance in Breast Cancer

Dec 24, 2020
Caner Mercan, Maschenka Balkenhol, Roberto Salgado, Mark Sherman, Philippe Vielh, Willem Vreuls, Antonio Polonia, Hugo M. Horlings, Wilko Weichert, Jodi M. Carter, Peter Bult, Matthias Christgen, Carsten Denkert, Koen van de Vijver, Jeroen van der Laak, Francesco Ciompi

Figure 1 for Automated Scoring of Nuclear Pleomorphism Spectrum with Pathologist-level Performance in Breast Cancer
Figure 2 for Automated Scoring of Nuclear Pleomorphism Spectrum with Pathologist-level Performance in Breast Cancer
Figure 3 for Automated Scoring of Nuclear Pleomorphism Spectrum with Pathologist-level Performance in Breast Cancer
Figure 4 for Automated Scoring of Nuclear Pleomorphism Spectrum with Pathologist-level Performance in Breast Cancer

Nuclear pleomorphism, defined herein as the extent of abnormalities in the overall appearance of tumor nuclei, is one of the components of the three-tiered breast cancer grading. Given that nuclear pleomorphism reflects a continuous spectrum of variation, we trained a deep neural network on a large variety of tumor regions from the collective knowledge of several pathologists, without constraining the network to the traditional three-category classification. We also motivate an additional approach in which we discuss the additional benefit of normal epithelium as baseline, following the routine clinical practice where pathologists are trained to score nuclear pleomorphism in tumor, having the normal breast epithelium for comparison. In multiple experiments, our fully-automated approach could achieve top pathologist-level performance in select regions of interest as well as at whole slide images, compared to ten and four pathologists, respectively.

* 16 pages, 11 figures 
Viaarxiv icon

Whole-Slide Mitosis Detection in H&E Breast Histology Using PHH3 as a Reference to Train Distilled Stain-Invariant Convolutional Networks

Aug 17, 2018
David Tellez, Maschenka Balkenhol, Irene Otte-Holler, Rob van de Loo, Rob Vogels, Peter Bult, Carla Wauters, Willem Vreuls, Suzanne Mol, Nico Karssemeijer, Geert Litjens, Jeroen van der Laak, Francesco Ciompi

Figure 1 for Whole-Slide Mitosis Detection in H&E Breast Histology Using PHH3 as a Reference to Train Distilled Stain-Invariant Convolutional Networks
Figure 2 for Whole-Slide Mitosis Detection in H&E Breast Histology Using PHH3 as a Reference to Train Distilled Stain-Invariant Convolutional Networks
Figure 3 for Whole-Slide Mitosis Detection in H&E Breast Histology Using PHH3 as a Reference to Train Distilled Stain-Invariant Convolutional Networks
Figure 4 for Whole-Slide Mitosis Detection in H&E Breast Histology Using PHH3 as a Reference to Train Distilled Stain-Invariant Convolutional Networks

Manual counting of mitotic tumor cells in tissue sections constitutes one of the strongest prognostic markers for breast cancer. This procedure, however, is time-consuming and error-prone. We developed a method to automatically detect mitotic figures in breast cancer tissue sections based on convolutional neural networks (CNNs). Application of CNNs to hematoxylin and eosin (H&E) stained histological tissue sections is hampered by: (1) noisy and expensive reference standards established by pathologists, (2) lack of generalization due to staining variation across laboratories, and (3) high computational requirements needed to process gigapixel whole-slide images (WSIs). In this paper, we present a method to train and evaluate CNNs to specifically solve these issues in the context of mitosis detection in breast cancer WSIs. First, by combining image analysis of mitotic activity in phosphohistone-H3 (PHH3) restained slides and registration, we built a reference standard for mitosis detection in entire H&E WSIs requiring minimal manual annotation effort. Second, we designed a data augmentation strategy that creates diverse and realistic H&E stain variations by modifying the hematoxylin and eosin color channels directly. Using it during training combined with network ensembling resulted in a stain invariant mitosis detector. Third, we applied knowledge distillation to reduce the computational requirements of the mitosis detection ensemble with a negligible loss of performance. The system was trained in a single-center cohort and evaluated in an independent multicenter cohort from The Cancer Genome Atlas on the three tasks of the Tumor Proliferation Assessment Challenge (TUPAC). We obtained a performance within the top-3 best methods for most of the tasks of the challenge.

* Accepted to appear in IEEE Transactions on Medical Imaging 
Viaarxiv icon