DASGAN - Joint Domain Adaptation and Segmentation for the Analysis of Epithelial Regions in Histopathology PD-L1 Images

Ansh Kapil, Tobias Wiestler, Simon Lanzmich, Abraham Silva, Keith Steele, Marlon Rebelatto, Guenter Schmidt, Nicolas Brieu

Keywords: Domain Adaptation, Semantic Segmentation, Digital Pathology, Generative Adversarial Networks

TL;DR: Joint domain adaptation and semantic segmentation enables reuse of data among two stain domains to improve segmentation performance along with reducing annotation efforts

Abstract: The analysis of the tumor environment on digital histopathology slides is becoming key for the understanding of the immune response against cancer, supporting the development of novel immuno-therapies. We introduce here a novel deep learning solution to the related problem of tumor epithelium segmentation. While most existing deep learning segmentation approaches are trained on time-consuming and costly manual annotation on single stain domain (PD-L1), we leverage here semi-automatically labeled images from a second stain domain (Cytokeratin-CK). We introduce an end-to-end trainable network that jointly segment tumor epithelium on PD-L1 while leveraging unpaired image-to-image translation between CK and PD-L1, therefore completely bypassing the need for serial sections or re-staining of slides. Extending the method to differentiate between PD-L1 positive and negative tumor epithelium regions enables the automated estimation of the clinically relevant PD-L1 Tumor Cell score. Quantitative experimental results demonstrate the accuracy of our approach against state-of-the-art segmentation methods.