Metadata-Aligned 3D MRI Representations for Contrast Understanding and Quality Control

TL;DR

This paper introduces MR-CLIP, a metadata-guided framework that creates unified 3D MRI contrast representations, enabling improved sequence classification and quality control without manual labels.

Contribution

MR-CLIP is a novel method that aligns MRI images with acquisition metadata to learn contrast representations, enhancing analysis across heterogeneous datasets.

Findings

Outperforms supervised baselines in few-shot sequence classification

Enables unsupervised detection of corrupted or inconsistent data

Creates distinct clusters of MRI sequences based on contrast

Abstract

Magnetic Resonance Imaging suffers from substantial data heterogeneity and the absence of standardized contrast labels across scanners, protocols, and institutions, which severely limits large-scale automated analysis. A unified representation of MRI contrast would enable a wide range of downstream utilities, from automatic sequence recognition to harmonization and quality control, without relying on manual annotations. To this end, we introduce MR-CLIP, a metadata-guided framework that learns MRI contrast representations by aligning volumetric images with their DICOM acquisition parameters. The resulting embeddings shows distinct clusters of MRI sequences and outperform supervised 3D baselines under data scarcity in few-shot sequence classification. Moreover, MR-CLIP enables unsupervised data quality control by identifying corrupted or inconsistent metadata through image-metadata embedding distances. By transforming routinely available acquisition metadata into a supervisory signal, MR-CLIP provides a scalable foundation for label-efficient MRI analysis across diverse clinical datasets.