Shazam: Unifying Multiple Foundation Models for Advanced Computational Pathology
This work addresses the need for consistent performance across diverse clinical tasks in computational pathology, offering a lightweight and flexible solution for real-world applications, though it is incremental in nature.
The paper tackles the variability of foundation models in computational pathology by proposing Shazam, a framework that dynamically combines multiple models, achieving superior performance on two pathology patch classification datasets compared to existing methods.
Foundation Models (FMs) in computational pathology (CPath) have significantly advanced the extraction of meaningful features from histopathology image datasets, achieving strong performance across various clinical tasks. Despite their impressive performance, these models often exhibit variability when applied to different tasks, prompting the need for a unified framework capable of consistently excelling across various applications. In this work, we propose Shazam, a novel framework designed to efficiently combine multiple CPath models. Unlike previous approaches that train a fixed-parameter FM, Shazam dynamically extracts and refines information from diverse FMs for each specific task. To ensure that each FM contributes effectively without dominance, a novel distillation strategy is applied, guiding the student model with features from all teacher models, which enhances its generalization ability. Experimental results on two pathology patch classification datasets demonstrate that Shazam outperforms existing CPath models and other fusion methods. Its lightweight, flexible design makes it a promising solution for improving CPath analysis in real-world settings. Code will be available at https://github.com/Tuner12/Shazam.