Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoring
Abstract Segmenting histological images and analyzing relevant regions are crucial for supporting pathologists in diagnosing various diseases. In prostate cancer diagnosis, Gleason grading and scoring relies on the recognition of different patterns in tissue samples. However, annotating large histol...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-05-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-00048-9 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849312069221875712 |
|---|---|
| author | Meili Ren Mengxing Huang Yu Zhang Zhijun Zhang Meiyan Ren |
| author_facet | Meili Ren Mengxing Huang Yu Zhang Zhijun Zhang Meiyan Ren |
| author_sort | Meili Ren |
| collection | DOAJ |
| description | Abstract Segmenting histological images and analyzing relevant regions are crucial for supporting pathologists in diagnosing various diseases. In prostate cancer diagnosis, Gleason grading and scoring relies on the recognition of different patterns in tissue samples. However, annotating large histological datasets is laborious, expensive, and often limited to slide-level or limited instance-level labels. To address this, we propose an enhanced hierarchical attention mechanism within a mixed multiple instance learning (MIL) model that effectively integrates slide-level and instance-level labels. Our hierarchical attention mechanism dynamically suppresses noisy instance-level labels while adaptively amplifying discriminative features, achieving a synergistic integration of global slide-level context and local superpixel patterns. This design significantly improves label utilization efficiency, leading to state-of-the-art performance in Gleason grading. Experimental results on the SICAPv2 and TMAs datasets demonstrate the superior performance of our model, achieving AUC scores of 0.9597 and 0.8889, respectively. Our work not only advances the state-of-the-art in Gleason grading but also highlights the potential of hierarchical attention mechanisms in mixed MIL models for medical image analysis. |
| format | Article |
| id | doaj-art-05d2f4c01306429280510ddf4bbc63d1 |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-05d2f4c01306429280510ddf4bbc63d12025-08-20T03:53:12ZengNature PortfolioScientific Reports2045-23222025-05-0115111410.1038/s41598-025-00048-9Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoringMeili Ren0Mengxing Huang1Yu Zhang2Zhijun Zhang3Meiyan Ren4Hainan Provincial Key Laboratory of Big Data and Smart Service, Hainan UniversityHainan Provincial Key Laboratory of Big Data and Smart Service, Hainan UniversityHainan Provincial Key Laboratory of Big Data and Smart Service, Hainan UniversityCenter of Network and Information Education Technology, Shanxi University of Finance and EconomicsSchool of Medical, Shanxi Datong UniversityAbstract Segmenting histological images and analyzing relevant regions are crucial for supporting pathologists in diagnosing various diseases. In prostate cancer diagnosis, Gleason grading and scoring relies on the recognition of different patterns in tissue samples. However, annotating large histological datasets is laborious, expensive, and often limited to slide-level or limited instance-level labels. To address this, we propose an enhanced hierarchical attention mechanism within a mixed multiple instance learning (MIL) model that effectively integrates slide-level and instance-level labels. Our hierarchical attention mechanism dynamically suppresses noisy instance-level labels while adaptively amplifying discriminative features, achieving a synergistic integration of global slide-level context and local superpixel patterns. This design significantly improves label utilization efficiency, leading to state-of-the-art performance in Gleason grading. Experimental results on the SICAPv2 and TMAs datasets demonstrate the superior performance of our model, achieving AUC scores of 0.9597 and 0.8889, respectively. Our work not only advances the state-of-the-art in Gleason grading but also highlights the potential of hierarchical attention mechanisms in mixed MIL models for medical image analysis.https://doi.org/10.1038/s41598-025-00048-9Gleason gradingHierarchical attention mechanismMultiple instance learning |
| spellingShingle | Meili Ren Mengxing Huang Yu Zhang Zhijun Zhang Meiyan Ren Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoring Scientific Reports Gleason grading Hierarchical attention mechanism Multiple instance learning |
| title | Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoring |
| title_full | Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoring |
| title_fullStr | Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoring |
| title_full_unstemmed | Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoring |
| title_short | Enhanced hierarchical attention mechanism for mixed MIL in automatic Gleason grading and scoring |
| title_sort | enhanced hierarchical attention mechanism for mixed mil in automatic gleason grading and scoring |
| topic | Gleason grading Hierarchical attention mechanism Multiple instance learning |
| url | https://doi.org/10.1038/s41598-025-00048-9 |
| work_keys_str_mv | AT meiliren enhancedhierarchicalattentionmechanismformixedmilinautomaticgleasongradingandscoring AT mengxinghuang enhancedhierarchicalattentionmechanismformixedmilinautomaticgleasongradingandscoring AT yuzhang enhancedhierarchicalattentionmechanismformixedmilinautomaticgleasongradingandscoring AT zhijunzhang enhancedhierarchicalattentionmechanismformixedmilinautomaticgleasongradingandscoring AT meiyanren enhancedhierarchicalattentionmechanismformixedmilinautomaticgleasongradingandscoring |