A Target-Free Harmonization Method for MRI

2026-05-02

AI Summary

This work addresses the challenge of domain shift in MRI images caused by variations in scanning parameters, sequences, or hardware, which severely degrades model generalization. Existing harmonization methods typically require access to target-domain data, raising privacy and data-sharing concerns. To overcome this limitation, we propose TgtFreeHarmony, a novel framework that achieves MRI style alignment without any target-domain data—a first in the field. Our approach leverages a disentangled generator to model the MRI style manifold and employs Bayesian optimization guided by downstream task performance feedback to adaptively estimate the target style. Evaluated on multi-site brain tissue segmentation tasks, TgtFreeHarmony significantly improves the analysis performance of source-domain images when applied to target domains, offering high-quality harmonization while preserving patient privacy and enabling practical, privacy-preserving deployment in clinical settings.

Abstract

In MRI, variations in scan parameters, sequence, or hardware can lead to discrepancies in image appearance, even for the same subject. These inconsistencies, known as domain shifts, can hinder image analysis and degrade the performance of deep learning models trained on data from specific target domains. MRI image harmonization aims to address these issues by aligning source domain images to the target domain images while preserving biological information such as anatomical structures. However, most existing harmonization approaches require access to both source and target domain data in training or test time. This dependence induces data sharing between institutions, raising concerns about patient privacy and substantially limiting the harmonization approaches that can be practically deployed in clinical settings. To overcome these limitations, we introduce TgtFreeHarmony, the harmonization framework tailored for target-free scenarios, eliminating the need for target domain data and any data sharing, enabling privacy-preserving harmonization directly within the source institution. Our approach estimates the target domain style by searching the manifold of MRI domain style constructed via a disentanglement-based generator using Bayesian optimization guided by the performance of a downstream task model, which is trained on target domain data. We evaluated our method on the brain tissue segmentation task across multiple institutes and demonstrated that it effectively harmonizes source images into target images, leading to improved downstream task performance. By enabling harmonization without any access to target-domain data, TgtFreeHarmony establishes a new direction of harmonization preserving data privacy that can be realistically deployed within clinical environments.

Problem

• MRI harmonization
• domain shift
• target-free
• data privacy
• clinical deployment

Innovation

• target-free harmonization
• MRI domain shift
• privacy-preserving
• disentanglement-based generator
• Bayesian optimization