Top-Down Attention-Based Multiple Instance Learning for Whole Slide Image Analysis
Reisenbüchler D, Deng R, Matek C, Feuerhake F, Merhof D (2026)
Publication Type: Conference contribution
Publication year: 2026
Journal
Publisher: Springer Science and Business Media Deutschland GmbH
Book Volume: 15960 LNCS
Pages Range: 651-660
Conference Proceedings Title: Lecture Notes in Computer Science
Event location: Daejeon, KOR
ISBN: 9783032049261
DOI: 10.1007/978-3-032-04927-8_62
Abstract
Multiple instance learning (MIL) has become the de facto standard approach for whole-slide image analysis in computational pathology (CPath). While instance-wise attention tends to miss correlations between instances, self-attention can capture these interactions, but remains agnostic to the particular task. To address this issue, we introduce Top-Down Attention-based Multiple Instance Learning (TDA-MIL), an architecture that first learns a general representation from the data via self-attention in an initial inference step, then identifies task-relevant instances through a feature selection module, and finally refines these representations by injecting the selected instances back into the attention mechanism for a second inference step. By focusing on task-specific signals, TDA-MIL effectively discerns subtle, yet significant, regions within each slide, leading to more precise classification. Extensive experiments on detecting lymph node metastasis in breast cancer, biomarker screening for microsatellite instability in different organs, and challenging molecular status prediction for HER2 in breast cancer show that TDA-MIL consistently surpasses other MIL baselines, underscoring the effectiveness of our proposed task-relevant refocusing and its broad applicability across CPath tasks. Our implementation is released at https://github.com/agentdr1/TDA_MIL.
Authors with CRIS profile
Involved external institutions
Universität Regensburg
Germany (DE)
Cardiovascular Research Foundation
United States (USA) (US)
Medizinische Hochschule Hannover (MHH) / Hannover Medical School
Germany (DE)
Fraunhofer-Institut für Bildgestützte Medizin (MEVIS)
Germany (DE)
Germany (DE)
Cardiovascular Research Foundation
United States (USA) (US)
Medizinische Hochschule Hannover (MHH) / Hannover Medical School
Germany (DE)
Fraunhofer-Institut für Bildgestützte Medizin (MEVIS)
Germany (DE)
How to cite
APA:
Reisenbüchler, D., Deng, R., Matek, C., Feuerhake, F., & Merhof, D. (2026). Top-Down Attention-Based Multiple Instance Learning for Whole Slide Image Analysis. In James C. Gee, Jaesung Hong, Carole H. Sudre, Polina Golland, Daniel C. Alexander, Juan Eugenio Iglesias, Archana Venkataraman, Jong Hyo Kim (Eds.), Lecture Notes in Computer Science (pp. 651-660). Daejeon, KOR: Springer Science and Business Media Deutschland GmbH.
MLA:
Reisenbüchler, Daniel, et al. "Top-Down Attention-Based Multiple Instance Learning for Whole Slide Image Analysis." Proceedings of the 28th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2025, Daejeon, KOR Ed. James C. Gee, Jaesung Hong, Carole H. Sudre, Polina Golland, Daniel C. Alexander, Juan Eugenio Iglesias, Archana Venkataraman, Jong Hyo Kim, Springer Science and Business Media Deutschland GmbH, 2026. 651-660.
BibTeX: Download