An EM-based semi-supervised deep learning approach for semantic segmentation of histopathological images from radical prostatectomies. (November 2018)

• Record Type: Journal Article
• Authors: Li, Jiayun; Speier, William; Ho, King Chung; Sarma, Karthik V.; Gertych, Arkadiusz; Knudsen, Beatrice S.; Arnold, Corey W.
• Is Part Of: Computerized medical imaging and graphics. Volume 69(2018)
• Page Start: 125
• Page End: 133
• Publication Date: 2018-11
• Subjects: Histopathological image segmentation -- Prostate cancer -- Expectation maximization -- Semi-supervised deep learning; Diagnostic imaging -- Periodicals; Imaging systems in medicine -- Periodicals; Diagnosis, Radioscopic -- Data processing -- Periodicals; Diagnostic Imaging -- Periodicals; Imagerie pour le diagnostic -- Périodiques; Diagnostic imaging; Periodicals; Electronic journals; Electronic journals; 616.0754
• DOI: 10.1016/j.compmedimag.2018.08.003
• Languages: English
• ISSNs: 0895-6111

Abstract

Graphical abstract: Highlights: A semi-supervised model for histopathological image segmentation. An EM-based framework to improve semantic segmentation with a large-scale weakly-labeled dataset. Adaptive biases improve performance by incorporating prior knowledge during EM training. Abstract: Automated Gleason grading is an important preliminary step for quantitative histopathological feature extraction. Different from the traditional task of classifying small pre-selected homogeneous regions, semantic segmentation provides pixel-wise Gleason predictions across an entire slide. Deep learning-based segmentation models can automatically learn visual semantics from data, which alleviates the need for feature engineering. However, performance of deep learning models is limited by the scarcity of large-scale fully annotated datasets, which can be both expensive and time-consuming to create. One way to address this problem is to leverage external weakly labeled datasets to augment models trained on the limited data. In this paper, we developed an expectation maximization-based approach constrained by an approximated prior distribution in order to extract useful representations from a large number of weakly labeled images generated from low-magnification annotations. This method was utilized to improve the performance of a model trained on a limited fully annotated dataset. Our semi-supervised approach trained with 135 fully annotated and 1800 weakly annotated tiles achieved a mean Jaccard Index of 49.5% on an independent test set, which was 14% higher than the initial model trained only on the fully annotated dataset.