Benchmarking Foundation Models for Renal Lesion Stratification in CT

TL;DR

This study benchmarks medical foundation models for renal lesion classification in CT, finding they match deep learning models but are outperformed by traditional radiomics, highlighting current limitations in FMs' fine-grained feature capture.

Contribution

It provides a comparative analysis of foundation models versus traditional methods for renal lesion classification, emphasizing the current performance gap.

Findings

FM embeddings match deep learning models in AUC but require less computation.

Radiomics significantly outperforms all deep learning approaches in AUC.

Current FMs do not yet capture the texture and shape details critical for histological differentiation.

Abstract

The rapid proliferation of open-source medical foundation models (FMs) raises a practical question: how well do their pre-trained representations transfer to clinically relevant but data-scarce classification tasks? Particularly in CT-based renal lesion classification, a push toward greater generalizability would be meaningful, as the field is constrained by inherently limited training data. We addressed this through a benchmark of three medical FMs on this specific task. This six-class problem spans common entities like cysts and clear cell renal cell carcinoma, alongside rare subtypes. Using a frozen feature-probing protocol, we compared FM embeddings against a handcrafted radiomics classifier and a 3D ResNet-50 trained from scratch. Models were trained on a composite dataset of 2,854 lesions and evaluated on an external test set of 234 lesions from The Cancer Imaging Archive. Our results reveal two key findings. First, FM performance (AUC 0.70-0.77) matched the from-scratch ResNet (AUC 0.72) while drastically reducing hardware demand, requiring only seconds on a CPU after feature extraction. However, the conventional radiomics baseline significantly outperformed all deep learning approaches, achieving an AUC of 0.88 (all p $\leq$ 0.002). This suggests that current generalist FM embeddings do not yet capture the fine-grained texture and shape heterogeneity driving histological subtype discrimination. Despite their potential in data-scarce settings, medical FMs did not surpass established models for renal lesion stratification, leaving radiomics as the current state-of-the-art.