Domain Generalization in Computational Pathology: Survey and Guidelines

Abstract

Deep learning models have exhibited exceptional effectiveness in Computational Pathology (CPath) for various tasks on multi-gigapixel histology images. Nevertheless, the presence of out-of-distribution data (stemming from different sources such as disparate imaging devices) can cause domain shift (DS). DS decreases the generalization of trained models to unseen datasets with slightly different data distributions, prompting the need for innovative domain generalization (DG) solutions. Recognizing the potential of DG to significantly influence diagnostic and prognostic models in cancer studies and clinical practice, we present this survey along with guidelines on achieving DG in CPath. We rigorously define various DS types, systematically review and categorize existing DG approaches and resources in CPath, and provide insights into their advantages, limitations, and applicability. We also conduct thorough benchmarking experiments with 28 cutting-edge DG algorithms to address a complex DG example problem. Our findings suggest that careful experiment design and Stain Augmentation technique can be very effective. However, there is no one-size-fits-all solution for DG in CPath. Therefore, we establish guidelines for detecting and managing DS in different scenarios. While most of the concepts and recommendations are given for applications in CPath, they apply to most medical image analysis tasks as well.