A machine learning model for separating epithelial and stromal regions in oral cavity squamous cell carcinomas using H&E-stained histology images: A multi-center, retrospective study

• Published in: Oral oncology 2022-08, Vol.131, p.105942-105942, Article 105942
• Main Authors: Wu, Yuxin, Koyuncu, Can F., Toro, Paula, Corredor, German, Feng, Qianyu, Buzzy, Christina, Old, Matthew, Teknos, Theodoros, Connelly, Stephen Thaddeus, Jordan, Richard C., Lang Kuhs, Krystle A., Lu, Cheng, Lewis, James S., Madabhushi, Anant
• Format: Article
• Language: English
• Subjects: Carcinoma, Squamous Cell - diagnostic imaging; Carcinoma, Squamous Cell - pathology; Deep Learning; Head and Neck Neoplasms; Humans; Image Processing, Computer-Assisted - methods; Machine Learning; Mouth Neoplasms - diagnostic imaging; Mouth Neoplasms - pathology; Retrospective Studies; Squamous Cell Carcinoma of Head and Neck
• ISSN: 1368-8375; 1879-0593
• DOI: 10.1016/j.oraloncology.2022.105942

[Display omitted] •Deep learning model achieved a consistently accurate performance for epithelium segmentation•The model trained on 10x magnification images achieved the best performance.•Morphologic features extracted from human- and AI-annotated epithelial regions are equivalent. Tissue slides from Oral cavity squamous cell carcinoma (OC-SCC), particularly the epithelial regions, hold morphologic features that are both diagnostic and prognostic. Yet, previously developed approaches for automated epithelium segmentation in OC-SCC have not been independently tested in a multi-center setting. In this study, we aimed to investigate the effectiveness and applicability of a convolutional neural network (CNN) model to perform epithelial segmentation using digitized H&E-stained diagnostic slides from OC-SCC patients in a multi-center setting. A CNN model was developed to segment the epithelial regions of digitized slides (n = 810), retrospectively collected from five different centers. Deep learning models were trained and validated using well-annotated tissue microarray (TMA) images (n = 212) at various magnifications. The best performing model was locked down and used for independent testing with a total of 478 whole-slide images (WSIs). Manually annotated epithelial regions were used as the reference standard for evaluation. We also compared the model generated results with IHC-stained epithelium (n = 120) as the reference. The locked-down CNN model trained on the TMA image training cohorts with 10x magnification achieved the best segmentation performance. The locked-down model performed consistently and yielded Pixel Accuracy, Recall Rate, Precision Rate, and Dice Coefficient that ranged from 95.8% to 96.6%, 79.1% to 93.8%, 85.7% to 89.3%, and 82.3% to 89.0%, respectively for the three independent testing WSI cohorts. The automated model achieved a consistently accurate performance for automated epithelial region segmentation compared to manual annotations. This model could be integrated into a computer-aided diagnosis or prognosis system.