Explainable Anatomy-Guided AI for Prostate MRI: Foundation Models and In Silico Clinical Trials for Virtual Biopsy-based Risk Assessment

2025-05-23

AI Summary

Current deep learning models for MRI-based risk stratification of prostate cancer (PCa) suffer from limited interpretability and clinical applicability. Method: We propose an anatomy-guided, fully automated deep learning framework featuring a novel Swin Transformer backbone (UMedPT) infused with anatomical priors, integrated with nnU-Net for high-accuracy prostate zonal segmentation (Dice scores: 0.95 for whole gland, 0.94 for peripheral zone, 0.92 for transition zone). A VAE-GAN–based counterfactual heatmap generation module enhances decision interpretability, while multi-scale feature fusion and clinical metadata integration optimize classification performance. Results: The system achieves an AUC of 0.79 for risk prediction, improves radiologists’ diagnostic accuracy to 0.77, reduces reading time by 40%, and enables virtual biopsy–informed clinical decision-making—thereby significantly advancing the trustworthiness and practical utility of AI in urological imaging.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Segmentation
• Natural Language Processing: Safety and Robustness

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

We present a fully automated, anatomically guided deep learning pipeline for prostate cancer (PCa) risk stratification using routine MRI. The pipeline integrates three key components: an nnU-Net module for segmenting the prostate gland and its zones on axial T2-weighted MRI; a classification module based on the UMedPT Swin Transformer foundation model, fine-tuned on 3D patches with optional anatomical priors and clinical data; and a VAE-GAN framework for generating counterfactual heatmaps that localize decision-driving image regions. The system was developed using 1,500 PI-CAI cases for segmentation and 617 biparametric MRIs with metadata from the CHAIMELEON challenge for classification (split into 70% training, 10% validation, and 20% testing). Segmentation achieved mean Dice scores of 0.95 (gland), 0.94 (peripheral zone), and 0.92 (transition zone). Incorporating gland priors improved AUC from 0.69 to 0.72, with a three-scale ensemble achieving top performance (AUC = 0.79, composite score = 0.76), outperforming the 2024 CHAIMELEON challenge winners. Counterfactual heatmaps reliably highlighted lesions within segmented regions, enhancing model interpretability. In a prospective multi-center in-silico trial with 20 clinicians, AI assistance increased diagnostic accuracy from 0.72 to 0.77 and Cohen's kappa from 0.43 to 0.53, while reducing review time per case by 40%. These results demonstrate that anatomy-aware foundation models with counterfactual explainability can enable accurate, interpretable, and efficient PCa risk assessment, supporting their potential use as virtual biopsies in clinical practice.

Problem

• Automated prostate cancer risk stratification using MRI
• Anatomy-guided AI for interpretable lesion localization
• Improving diagnostic accuracy and efficiency in clinical practice

Innovation

• Anatomy-guided nnU-Net for prostate segmentation
• Swin Transformer with anatomical priors for classification
• VAE-GAN framework for counterfactual heatmaps