Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.15012 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913554463981568 |
|---|---|
| author | Mittmann, Gesa Laiouar-Pedari, Sara Mehrtens, Hendrik A. Haggenmüller, Sarah Bucher, Tabea-Clara Chanda, Tirtha Gaisa, Nadine T. Wagner, Mathias Klamminger, Gilbert Georg Rau, Tilman T. Neppl, Christina Compérat, Eva Maria Gocht, Andreas Hämmerle, Monika Rupp, Niels J. Westhoff, Jula Krücken, Irene Seidl, Maximillian Schürch, Christian M. Bauer, Marcus Solass, Wiebke Tam, Yu Chun Weber, Florian Grobholz, Rainer Augustyniak, Jaroslaw Kalinski, Thomas Hörner, Christian Mertz, Kirsten D. Döring, Constanze Erbersdobler, Andreas Deubler, Gabriele Bremmer, Felix Sommer, Ulrich Brodhun, Michael Griffin, Jon Lenon, Maria Sarah L. Trpkov, Kiril Cheng, Liang Chen, Fei Levi, Angelique Cai, Guoping Nguyen, Tri Q. Amin, Ali Cimadamore, Alessia Shabaik, Ahmed Manucha, Varsha Ahmad, Nazeel Messias, Nidia Sanguedolce, Francesca Taheri, Diana Baraban, Ezra Jia, Liwei Shah, Rajal B. Siadat, Farshid Swarbrick, Nicole Park, Kyung Hassan, Oudai Sakhaie, Siamak Downes, Michelle R. Miyamoto, Hiroshi Williamson, Sean R. Holland-Letz, Tim Schneider, Carolin V. Kather, Jakob Nikolas Tolkach, Yuri Brinker, Titus J. |
| author_facet | Mittmann, Gesa Laiouar-Pedari, Sara Mehrtens, Hendrik A. Haggenmüller, Sarah Bucher, Tabea-Clara Chanda, Tirtha Gaisa, Nadine T. Wagner, Mathias Klamminger, Gilbert Georg Rau, Tilman T. Neppl, Christina Compérat, Eva Maria Gocht, Andreas Hämmerle, Monika Rupp, Niels J. Westhoff, Jula Krücken, Irene Seidl, Maximillian Schürch, Christian M. Bauer, Marcus Solass, Wiebke Tam, Yu Chun Weber, Florian Grobholz, Rainer Augustyniak, Jaroslaw Kalinski, Thomas Hörner, Christian Mertz, Kirsten D. Döring, Constanze Erbersdobler, Andreas Deubler, Gabriele Bremmer, Felix Sommer, Ulrich Brodhun, Michael Griffin, Jon Lenon, Maria Sarah L. Trpkov, Kiril Cheng, Liang Chen, Fei Levi, Angelique Cai, Guoping Nguyen, Tri Q. Amin, Ali Cimadamore, Alessia Shabaik, Ahmed Manucha, Varsha Ahmad, Nazeel Messias, Nidia Sanguedolce, Francesca Taheri, Diana Baraban, Ezra Jia, Liwei Shah, Rajal B. Siadat, Farshid Swarbrick, Nicole Park, Kyung Hassan, Oudai Sakhaie, Siamak Downes, Michelle R. Miyamoto, Hiroshi Williamson, Sean R. Holland-Letz, Tim Schneider, Carolin V. Kather, Jakob Nikolas Tolkach, Yuri Brinker, Titus J. |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_15012 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Pathologist-like explainable AI for interpretable Gleason grading in prostate cancer Mittmann, Gesa Laiouar-Pedari, Sara Mehrtens, Hendrik A. Haggenmüller, Sarah Bucher, Tabea-Clara Chanda, Tirtha Gaisa, Nadine T. Wagner, Mathias Klamminger, Gilbert Georg Rau, Tilman T. Neppl, Christina Compérat, Eva Maria Gocht, Andreas Hämmerle, Monika Rupp, Niels J. Westhoff, Jula Krücken, Irene Seidl, Maximillian Schürch, Christian M. Bauer, Marcus Solass, Wiebke Tam, Yu Chun Weber, Florian Grobholz, Rainer Augustyniak, Jaroslaw Kalinski, Thomas Hörner, Christian Mertz, Kirsten D. Döring, Constanze Erbersdobler, Andreas Deubler, Gabriele Bremmer, Felix Sommer, Ulrich Brodhun, Michael Griffin, Jon Lenon, Maria Sarah L. Trpkov, Kiril Cheng, Liang Chen, Fei Levi, Angelique Cai, Guoping Nguyen, Tri Q. Amin, Ali Cimadamore, Alessia Shabaik, Ahmed Manucha, Varsha Ahmad, Nazeel Messias, Nidia Sanguedolce, Francesca Taheri, Diana Baraban, Ezra Jia, Liwei Shah, Rajal B. Siadat, Farshid Swarbrick, Nicole Park, Kyung Hassan, Oudai Sakhaie, Siamak Downes, Michelle R. Miyamoto, Hiroshi Williamson, Sean R. Holland-Letz, Tim Schneider, Carolin V. Kather, Jakob Nikolas Tolkach, Yuri Brinker, Titus J. Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition 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. |
| title | Pathologist-like explainable AI for interpretable Gleason grading in prostate cancer |
| topic | Image and Video Processing Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2410.15012 |