A histopathological image analysis for the classification of endocervical adenocarcinoma Silva patterns depend on weakly supervised deep learning.

Researchers

Journal

Modalities

Models

Abstract

25% of cervical cancers are classified as endocervical adenocarcinomas (EACs), which comprise a highly heterogeneous group of tumors. A histopathological risk stratification system known as the Silva pattern system was developed based on morphology. However, accurately classifying such patterns can be challenging. The objective was to develop a deep learning pipeline (Silva3-AI) that automatically analyzes whole-slide-image-based histopathological images and identifies Silva patterns with high accuracy. Initially, a total of 202 patients with EACs and histopathological slides were obtained from QLHSDU for developing and internally testing the Silva3-AI model. Subsequently, an additional 161 patients and slides were collected from seven other medical centers for independent testing. The Silva3-AI model was developed using a ViT and RNN architecture, utilizing multi-magnification patches, and its performance was evaluated based on class-specific AUROC. Silva3-AI achieved a class-specific AUROC of 0.947 for Silva A, 0.908 for Silva B, and 0.947 for Silva C on the independent test set. Notably, the performance of Silva3-AI was consistent with that of professional pathologists possessing 10 years of diagnostic experience. The findings in this study implicate Silva3-AI, which could predict Silva pattern in an interpretable manner, achieved comparable performance to that of experienced pathologists. Furthermore, the visualization of prediction heatmaps facilitated the identification of TME heterogeneity, which is known to contribute to variations in Silva patterns.Copyright © 2024. Published by Elsevier Inc.