CNN stability training improves robustness to scanner and IHC-based image variability for epithelium segmentation in cervical histology

Frontiers in Medicine

Medicine

Amy S. Tao¹  Nicolas Wentzensen¹  Megan A. Clarke¹  Liam Bartels²  Alexandra Krauthoff²  Philipp Ströbel³  Andreas Keil⁴  Felipe Miranda Ruiz⁴  Bernd Lahrmann⁴  Niels Grabe⁵  Steffen Härtel⁶ 

[1] Division of Cancer Epidemiology and Genetics, US National Cancer Institute (NCI), Bethesda, MD, United States;Hamamatsu Tissue Imaging and Analysis Center (TIGA), BIOQUANT Center, Heidelberg University, Heidelberg, Germany;Medical Oncology Department, National Center for Tumor Diseases (NCT), Heidelberg, Germany;Institute of Pathology, University Medical Center Göttingen UMG, Göttingen, Germany;Institute of Pathology, University Medical Center Göttingen UMG, Göttingen, Germany;Hamamatsu Tissue Imaging and Analysis Center (TIGA), BIOQUANT Center, Heidelberg University, Heidelberg, Germany;Institute of Pathology, University Medical Center Göttingen UMG, Göttingen, Germany;Hamamatsu Tissue Imaging and Analysis Center (TIGA), BIOQUANT Center, Heidelberg University, Heidelberg, Germany;Medical Oncology Department, National Center for Tumor Diseases (NCT), Heidelberg, Germany;Medical Faculty, Center of Medical Informatics and Telemedicine (CIMT), University of Chile, Santiago, Chile;

关键词: digital pathology;    deep learning;    robustness;    image variability;    stability training;    cervical intraepithelial neoplasia;   

DOI  :  10.3389/fmed.2023.1173616

 received in 2023-02-24, accepted in 2023-06-06,  发布年份 2023

来源: Frontiers

【 摘 要 】

BackgroundIn digital pathology, image properties such as color, brightness, contrast and blurriness may vary based on the scanner and sample preparation. Convolutional Neural Networks (CNNs) are sensitive to these variations and may underperform on images from a different domain than the one used for training. Robustness to these image property variations is required to enable the use of deep learning in clinical practice and large scale clinical research.AimsCNN Stability Training (CST) is proposed and evaluated as a method to increase CNN robustness to scanner and Immunohistochemistry (IHC)-based image variability.MethodsCST was applied to segment epithelium in immunohistological cervical Whole Slide Images (WSIs). CST randomly distorts input tiles and factors the difference between the CNN prediction for the original and distorted inputs within the loss function. CNNs were trained using 114 p16-stained WSIs from the same scanner, and evaluated on 6 WSI test sets, each with 23 to 24 WSIs of the same tissue but different scanner/IHC combinations. Relative robustness (rAUC) was measured as the difference between the AUC on the training domain test set (i.e., baseline test set) and the remaining test sets.ResultsAcross all test sets, The AUC of CST models outperformed “No CST” models (AUC: 0.940–0.989 vs. 0.905–0.986, p < 1e − 8), and obtained an improved robustness (rAUC: [−0.038, −0.003] vs. [−0.081, −0.002]). At a WSI level, CST models showed an increase in performance in 124 of the 142 WSIs. CST models also outperformed models trained with random on-the-fly data augmentation (DA) in all test sets ([0.002, 0.021], p < 1e-6).ConclusionCST offers a path to improve CNN performance without the need for more data and allows customizing distortions to specific use cases. A python implementation of CST is publicly available at https://github.com/TIGACenter/CST_v1.

【 授权许可】

Unknown   
Copyright © 2023 Miranda Ruiz, Lahrmann, Bartels, Krauthoff, Keil, Härtel, Tao, Ströbel, Clarke, Wentzensen and Grabe.