Going Beyond H&E and Oncology: How Do Histopathology Foundation Models Perform for Multi-stain IHC and Immunology?
It addresses the problem of AI model transferability in histopathology for autoimmune disease diagnosis, highlighting challenges in applying cancer-trained models to immunology, which is incremental as it tests existing methods on new data.
This study evaluated the generalization of histopathology foundation models on multi-stain autoimmune IHC datasets, finding that models pretrained on cancer H&E images did not significantly outperform ImageNet-pretrained models and showed issues like feature misinterpretation and bias.
This study evaluates the generalisation capabilities of state-of-the-art histopathology foundation models on out-of-distribution multi-stain autoimmune Immunohistochemistry datasets. We compare 13 feature extractor models, including ImageNet-pretrained networks, and histopathology foundation models trained on both public and proprietary data, on Rheumatoid Arthritis subtyping and Sjogren's Disease detection tasks. Using a simple Attention-Based Multiple Instance Learning classifier, we assess the transferability of learned representations from cancer H&E images to autoimmune IHC images. Contrary to expectations, histopathology-pretrained models did not significantly outperform ImageNet-pretrained models. Furthermore, there was evidence of both autoimmune feature misinterpretation and biased feature importance. Our findings highlight the challenges in transferring knowledge from cancer to autoimmune histopathology and emphasise the need for careful evaluation of AI models across diverse histopathological tasks. The code to run this benchmark is available at https://github.com/AmayaGS/ImmunoHistoBench.