DeepCIN: Attention-Based Cervical histology Image Classification with Sequential Feature Modeling for Pathologist-Level Accuracy

Sudhir Sornapudi; R Joe Stanley; William V Stoecker; Rodney Long; Zhiyun Xue; Rosemary Zuna; Shellaine R Frazier; Sameer Antani

doi:10.4103/jpi.jpi_50_20

DeepCIN: Attention-Based Cervical histology Image Classification with Sequential Feature Modeling for Pathologist-Level Accuracy

J Pathol Inform. 2020 Dec 24:11:40. doi: 10.4103/jpi.jpi_50_20. eCollection 2020.

Authors

Sudhir Sornapudi¹, R Joe Stanley¹, William V Stoecker², Rodney Long³, Zhiyun Xue³, Rosemary Zuna⁴, Shellaine R Frazier⁵, Sameer Antani³

Affiliations

¹ Department of Electrical and Computer Engineering, Missouri University of Science and Technology, Rolla, MO, USA.
² Stoecker and Associates, Rolla, MO, USA.
³ Lister Hill National Center for Biomedical Communications, National Library of Medicine, National Institutes of Health, Bethesda, MD, USA.
⁴ Department of Pathology, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA.
⁵ Department of Surgical Pathology, University of Missouri Hospitals and Clinics, Columbia, MO, USA.

Abstract

Background: Cervical cancer is one of the deadliest cancers affecting women globally. Cervical intraepithelial neoplasia (CIN) assessment using histopathological examination of cervical biopsy slides is subject to interobserver variability. Automated processing of digitized histopathology slides has the potential for more accurate classification for CIN grades from normal to increasing grades of pre-malignancy: CIN1, CIN2, and CIN3.

Methodology: Cervix disease is generally understood to progress from the bottom (basement membrane) to the top of the epithelium. To model this relationship of disease severity to spatial distribution of abnormalities, we propose a network pipeline, DeepCIN, to analyze high-resolution epithelium images (manually extracted from whole-slide images) hierarchically by focusing on localized vertical regions and fusing this local information for determining Normal/CIN classification. The pipeline contains two classifier networks: (1) a cross-sectional, vertical segment-level sequence generator is trained using weak supervision to generate feature sequences from the vertical segments to preserve the bottom-to-top feature relationships in the epithelium image data and (2) an attention-based fusion network image-level classifier predicting the final CIN grade by merging vertical segment sequences.

Results: The model produces the CIN classification results and also determines the vertical segment contributions to CIN grade prediction.

Conclusion: Experiments show that DeepCIN achieves pathologist-level CIN classification accuracy.

Keywords: Attention networks; cervical cancer; cervical intraepithelial neoplasia; classification; convolutional neural networks; digital pathology; fusion-based classification; histology; recurrent neural networks.