Bridging Single Distortion Artifacts and Mmultifactorial Clinical Quality: Few-shot Biparametric MRI Quality Assessment via Distortion-trained Prototypical Networks

2026-06-17

AI Summary

This study addresses the challenges in clinical quality assessment of multiparametric prostate MRI, which are hindered by scarce annotations, limited availability of low-quality samples, and the subjective, time-consuming nature of PI-QUAL scoring—particularly due to geometric distortions in diffusion-weighted imaging (DWI). To overcome these limitations, the authors propose a novel few-shot, dual-parameter (T2-weighted and DWI) quality evaluation method based on prototypical networks. Their approach innovatively leverages easily obtainable single-distortion artifacts as a proxy task, enabling efficient meta-transfer to complex, multifactorial clinical quality ratings with as few as five annotated samples. The method employs a dual-branch 3D ResNet to fuse multimodal features, integrating FiLM and gradient reversal layers to align feature distributions across different b-values and mitigate acquisition bias, while incorporating anatomical context to enhance discriminative power. Evaluated on two datasets, the proposed framework significantly outperforms existing few-shot baselines, offering a practical and efficient solution for automated MRI quality control in clinical settings.

Abstract

Clinical prostate multi-parametric MRI relies heavily on high-quality diffusion-weighted imaging (DWI), yet reading DWI is frequently compromised by geometric distortion, often caused by rectal air. Assessing quality via the PI-QUAL scoring system is an emerging clinical standard, but it is subjective, time-consuming and suffers from a class imbalance where low-quality cases are diverse and relatively scarce. Using the PRIME clinical trial as an example, there are $6\%$ images with PI-QUAL scores lower than 4, $87\%$ of DWI issues are due to distortion. Many of the other clinical quality issues are under-represented. To address this common dual-scarcity of annotated clinical data, we propose a few-shot biparametric prototypical network for automated image quality assessment (IQA). Our framework utilizes a dual-branch 3D ResNet to fuse T2-weighted and DWI features, providing anatomical context to distinguish true morphology from distortion. To handle real-world heterogeneity, we introduce feature-wise linear modulation (FiLM) and a gradient reversal layer (GRL) to align feature distributions conditioned on varying b-values while suppressing acquisition-related biases. We demonstrate that a model meta-trained solely on comparatively objective, readily obtainable distortion labels can effectively adapt to predicting complex, multi-factorial clinical quality scores such as PI-QUAL using only five representative samples. Experimental results on two datasets show that our method significantly outperforms few-shot learning baselines for this challenging IQA task, offering a practically feasible and data-efficient solution for standardizing prostate MRI quality control in clinical workflows.

Problem

• MRI quality assessment
• few-shot learning
• geometric distortion
• class imbalance
• PI-QUAL

Innovation

• few-shot learning
• prototypical networks
• biparametric MRI
• distortion-aware IQA
• feature-wise linear modulation