Paper ID: 2410.15012

Pathologist-like explainable AI for interpretable Gleason grading in prostate cancer

Gesa Mittmann, Sara Laiouar-Pedari, Hendrik A. Mehrtens, Sarah Haggenmüller, Tabea-Clara Bucher, Tirtha Chanda, Nadine T. Gaisa, Mathias Wagner, Gilbert Georg Klamminger, Tilman T. Rau, Christina Neppl, Eva Maria Compérat, Andreas Gocht, Monika Hämmerle, Niels J. Rupp, Jula Westhoff, Irene Krücken, Maximillian Seidl, Christian M. Schürch, Marcus Bauer, Wiebke Solass, Yu Chun Tam, Florian Weber, Rainer Grobholz, Jaroslaw Augustyniak, Thomas Kalinski, Christian Hörner, Kirsten D. Mertz, Constanze Döring, Andreas Erbersdobler, Gabriele Deubler, Felix Bremmer, Ulrich Sommer, Michael Brodhun, Jon Griffin, Maria Sarah L. Lenon, Kiril Trpkov, Liang Cheng, Fei Chen, Angelique Levi, Guoping Cai, Tri Q. Nguyen, Ali Amin, Alessia Cimadamore, Ahmed Shabaik, Varsha Manucha, Nazeel Ahmad, Nidia Messias, Francesca Sanguedolce, Diana Taheri, Ezra Baraban, Liwei Jia, Rajal B. Shah, Farshid Siadat, Nicole Swarbrick, Kyung Park, Oudai Hassan, Siamak Sakhaie, Michelle R. Downes, Hiroshi Miyamoto, Sean R. Williamson, Tim Holland-Letz, Carolin V. Schneider, Jakob Nikolas Kather, Yuri Tolkach, Titus J. Brinker et al. (20 additional authors not shown) You must enabled JavaScript to view entire author list.

The aggressiveness of prostate cancer, the most common cancer in men worldwide, is primarily assessed based on histopathological data using the Gleason scoring system. While artificial intelligence (AI) has shown promise in accurately predicting Gleason scores, these predictions often lack inherent explainability, potentially leading to distrust in human-machine interactions. To address this issue, we introduce a novel dataset of 1,015 tissue microarray core images, annotated by an international group of 54 pathologists. The annotations provide detailed localized pattern descriptions for Gleason grading in line with international guidelines. Utilizing this dataset, we develop an inherently explainable AI system based on a U-Net architecture that provides predictions leveraging pathologists' terminology. This approach circumvents post-hoc explainability methods while maintaining or exceeding the performance of methods trained directly for Gleason pattern segmentation (Dice score: 0.713 $\pm$ 0.003 trained on explanations vs. 0.691 $\pm$ 0.010 trained on Gleason patterns). By employing soft labels during training, we capture the intrinsic uncertainty in the data, yielding strong results in Gleason pattern segmentation even in the context of high interobserver variability. With the release of this dataset, we aim to encourage further research into segmentation in medical tasks with high levels of subjectivity and to advance the understanding of pathologists' reasoning processes.

Submitted: Oct 19, 2024