Feature adaptation improves analysis of large medical slide images
LEAF-Mamba: Feature Adaptation and Fusion for Whole Slide Image Analysis
Computer Vision and Pattern Recognition
Summary
Analyzing very large medical images, like whole slide images of tissue samples, is hard because these images are huge and have few labels to learn from. The authors present LEAF-Mamba, a new method that changes and combines features extracted from small patches of these images to better capture important details. Their approach balances remembering the original features with learning task-specific information and uses two ways to understand both local and long-range connections in the image. Tests show that their method works better than previous approaches for classifying diseases and predicting patient outcomes from these slides.
What this means in practice
- •For medical imaging teams: Improve disease diagnosis and prognosis by enhancing computational analysis of large pathology slides using adapted and fused patch features.
- •For healthcare data engineers: Build slide-level models that better handle large image sequences by jointly adapting features and aggregating contextual information for clinical prediction tasks.
Authors
Lubin Gan, Jing Zhang, Mingjun Ma, Linhao Qu, Dachun Kai, Xiaoyan Sun
Abstract
Whole slide image (WSI) analysis enables computational diagnosis and prognosis, but gigapixel images and limited annotations make slide-level learning challenging. Multiple instance learning addresses this task by aggregating patch representations from frozen pretrained encoders. Recently, Mamba has enabled efficient modeling of long patch sequences. However, task supervision alone does not explicitly constrain information recovery during feature adaptation, while Mamba couples local processing with long-range propagation rather than retaining separate representations for fusion. Motivated by these challenges, we introduce LEAF-Mamba, a framework that jointly addresses latent evidence adaptation and fusion for frozen-feature WSI analysis. A reconstruction-regularized adapter learns compact, task-specific representations while reconstructing the input features. Parallel Mamba and convolutional streams then capture long-range dependencies and sequence-local interactions, with an input-dependent gate modulating their joint representation before slide-level pooling. Experiments across feature extractors demonstrate improved average performance over strong MIL baselines in both diagnostic classification and survival prediction. Component analyses further support the contributions of adaptation and fusion. These findings highlight the value of jointly designing feature transformation and contextual aggregation to translate pretrained representations into effective slide-level predictions.