Radiomics enhances kidney cancer type classification from CT images
Radiomics--Foundation Fusion for Interpretable RCC Classification: Internal Benchmarking and Exploratory External Transfer
Computer Vision and Pattern Recognition
Summary
Classifying types of kidney cancer from CT scans before surgery is difficult because the cancer subtypes look very similar. The authors tested different machine learning features and combined them with traditional radiomics features, which are hand-crafted measurements from images that help explain what the model sees. They found that combining deep learning features with radiomics improved accuracy and kept the results understandable to doctors. This shows that radiomics is still useful alongside newer AI methods for interpreting tumors in kidney cancer.
What this means in practice
- •For radiology teams: Improve preoperative classification of kidney tumors on CT scans using models that combine radiomics and deep learning features to aid treatment planning.
- •For medical image software developers: Integrate interpretable radiomics and foundation model fusion methods into diagnostic software to support clinicians in evaluating renal cancer subtypes.$Commercial implications: Enables development of enhanced, interpretable kidney cancer classification tools for diagnostic imaging companies to sell to hospitals.
Authors
Yuan Liang, Fangyijie Wang, Kathleen M. Curran, Guénolé Silvestre, Sourav Bhattacharjee, Abraham Campbell
Abstract
Accurate preoperative subtype classification of renal cell carcinoma (RCC) from contrast-enhanced CT remains clinically challenging because clear cell RCC (ccRCC) and non-clear cell RCC often show overlapping imaging appearances. This study evaluates whether foundation representations reduce reliance on handcrafted radiomics, or whether radiomics remains complementary for interpretable tumour characterisation. We compared radiomics, conventional CNN features, MedicalNet-pretrained features, MedVAE representations, and fusion variants for binary ccRCC classification on KiTS23, reporting area under the receiver operating characteristic curve (AUC) with bootstrap confidence intervals and average precision (AP) as a complementary class-imbalance-sensitive metric. We further assessed branch-removal ablation, TCGA/AIMI external transfer, and interpretability using radiomics permutation importance and gate-level analysis. Internally, 3D MedVAE gated fusion achieved the best performance, with an AUC of 82.7% and AP of 92.2%. On the external TCGA cohort, the same model achieved an AUC of 79.5% and AP of 98.9%, although specificity remains uncertain because only two external non-ccRCC cases were available. Gate analysis showed a radiomics-dominant fusion regime, suggesting that foundation representations acted as case-dependent refinement signals rather than replacements for structured tumour descriptors. These findings support radiomics as a complementary and clinically interpretable component of CT-based RCC characterisation in the foundation-model era.