Computational Pathology
Computational pathology applies computational methods, primarily deep learning, to analyze digitized histopathology slides, aiming to improve diagnostic accuracy, efficiency, and accessibility. Current research emphasizes developing robust and generalizable models, focusing on architectures like Vision Transformers, Multiple Instance Learning (MIL), and foundation models, often incorporating multimodal data (e.g., pathology reports, gene expression) and addressing challenges like domain generalization and efficient WSI representation. This field holds significant promise for improving cancer diagnosis and prognosis, accelerating research, and potentially reducing the workload on pathologists.
Papers
MambaMIL: Enhancing Long Sequence Modeling with Sequence Reordering in Computational Pathology
Shu Yang, Yihui Wang, Hao Chen
ReStainGAN: Leveraging IHC to IF Stain Domain Translation for in-silico Data Generation
Dominik Winter, Nicolas Triltsch, Philipp Plewa, Marco Rosati, Thomas Padel, Ross Hill, Markus Schick, Nicolas Brieu
Beyond Multiple Instance Learning: Full Resolution All-In-Memory End-To-End Pathology Slide Modeling
Gabriele Campanella, Eugene Fluder, Jennifer Zeng, Chad Vanderbilt, Thomas J. Fuchs
Reducing self-supervised learning complexity improves weakly-supervised classification performance in computational pathology
Tim Lenz, Omar S. M. El Nahhas, Marta Ligero, Jakob Nikolas Kather