arXiv:2510.27237v2 Announce Type: replace 
Abstract: Whole slide image (WSI) analysis has emerged as an increasingly essential technique in computational pathology. Recent advances in the pathology foundation models (FMs) have demonstrated significant advantages in deriving meaningful patch-level or slide-level multi-scale features from WSIs. However, current pathology FMs have exhibited substantial heterogeneity caused by diverse private training datasets and different network architectures. This heterogeneity introduces performance variability when we utilize the features from different FMs in the downstream tasks. To fully explore the advantages of multiple FMs effectively, in this work, we propose a novel framework for the fusion of multi-scale heterogeneous pathology FMs, called FuseCPath, yielding a model with a superior ensemble performance. The main contributions of our framework can be summarized as follows: (i) To guarantee the representativeness of the training patches, we propose a multi-view clustering-based method to filter out the discriminative patches via multiple FMs' embeddings. (ii) To effectively fuse the patch-level FMs, we devise a cluster-level re-embedding strategy to online capture patch-level local features. (iii) To effectively fuse the slide-level FMs, we devise a collaborative distillation strategy to explore the connections between slide-level FMs. Extensive experiments demonstrate that the proposed FuseCPath achieves state-of-the-art performance across multiple tasks on diverse datasets.

تناقش المقالة ظهور تحليل صور الشرائح الكاملة (WSI) كتقنية حاسمة في علم الأمراض الحسابي. تبرز تطوير إطار عمل جديد، FuseCPath، الذي يهدف إلى دمج نماذج أساس الأمراض المتنوعة متعددة المقاييس. تتناول هذه الطريقة تباين الأداء الناتج عن الاختلافات في مجموعات بيانات التدريب وهياكل الشبكات، مما يحسن تحليل صور الشرائح الكاملة.

El artículo discute la aparición del análisis de imágenes de diapositivas completas (WSI) como una técnica crucial en la patología computacional. Destaca el desarrollo de un nuevo marco, FuseCPath, que tiene como objetivo fusionar modelos de fundación patológicos heterogéneos a múltiples escalas. Este enfoque aborda la variabilidad del rendimiento causada por las diferencias en los conjuntos de datos de entrenamiento y las arquitecturas de red, mejorando así el análisis de las WSI.

L'article traite de l'émergence de l'analyse d'images de diapositives entières (WSI) comme une technique cruciale en pathologie computationnelle. Il met en lumière le développement d'un nouveau cadre, FuseCPath, qui vise à fusionner des modèles de fondation pathologique hétérogènes à plusieurs échelles. Cette approche aborde la variabilité de performance causée par les différences dans les ensembles de données d'entraînement et les architectures de réseau, améliorant ainsi l'analyse des WSI.

The article discusses the emergence of whole slide image (WSI) analysis as a crucial technique in computational pathology. It highlights the development of a novel framework, FuseCPath, which aims to fuse multi-scale heterogeneous pathology foundation models. This approach addresses the performance variability caused by differences in training datasets and network architectures, ultimately enhancing the analysis of WSIs.

Fusion of Multi-scale Heterogeneous Pathology Foundation Models for Whole Slide Image Analysis

Was this article worth reading? Share it

PlantFCE Model Builder

LucidQuery AI

MyFramework

AIPortalX

Metaflow AI

Octofy

Ready to build your own newsroom?