Cross-Vendor Reproducibility of Radiomics-based Machine Learning Models for Computer-aided Diagnosis

TL;DR

This study tackles the cross-vendor reproducibility of radiomics-based ML for prostate cancer detection in MRI by comparing SVM and Random Forest models trained on radiomic features from two toolkits (Pyradiomics and MRCradiomics) with MRMR feature selection. Using axial T2-weighted images from IMPROD/MULTI-IMPROD and related trials, it evaluates three feature sets (Pyradiomics, MRCradiomics, and their combination) under cross-vendor testing ( Siemens vs Philips). The results show notable cross-vendor variability, with higher AUCs on the Siemens-based dataset and substantial drops on Philips data, though Pyradiomics features demonstrate stronger robustness in several configurations. The work highlights the potential of multimodal feature fusion for improved CDS while underscoring the critical need for rigorous cross-platform validation to ensure reliable AI-driven diagnostics across imaging platforms.

Abstract

Background: The reproducibility of machine-learning models in prostate cancer detection across different MRI vendors remains a significant challenge. Methods: This study investigates Support Vector Machines (SVM) and Random Forest (RF) models trained on radiomic features extracted from T2-weighted MRI images using Pyradiomics and MRCradiomics libraries. Feature selection was performed using the maximum relevance minimum redundancy (MRMR) technique. We aimed to enhance clinical decision support through multimodal learning and feature fusion. Results: Our SVM model, utilizing combined features from Pyradiomics and MRCradiomics, achieved an AUC of 0.74 on the Multi-Improd dataset (Siemens scanner) but decreased to 0.60 on the Philips test set. The RF model showed similar trends, with notable robustness for models using Pyradiomics features alone (AUC of 0.78 on Philips). Conclusions: These findings demonstrate the potential of multimodal feature integration to improve the robustness and generalizability of machine-learning models for clinical decision support in prostate cancer detection. This study marks a significant step towards developing reliable AI-driven diagnostic tools that maintain efficacy across various imaging platforms.

Figures (7)

• Figure 1: Result of SVM trained over Pyradiomics and MRCradiomics
• Figure 2: Results of Random Forest trained over Pyradiomics and MRCradiomics
• Figure 3: Result of Random Forest trained over Pyradiomics
• Figure 4: Results of SVM trained over Pyradiomics
• Figure 5: Result of Random Forest trained over MRCradiomics.