Scanner-Agnostic MRI Harmonization via SSIM-Guided Disentanglement

2025-10-24

AI Summary

Multi-center brain MRI suffers from substantial domain shifts due to variations in scanner models, acquisition protocols, and imaging sites, severely impairing model generalizability and reproducibility. To address this, we propose a SSIM-guided 3D disentanglement and harmonization framework that employs differentiable SSIM loss to separately optimize luminance, contrast, and structural components—explicitly decoupling anatomical content from device-specific stylistic artifacts—and enabling multi-target style harmonization. Evaluated on T1-weighted images, our method achieves structural SSIM of 0.97 and luminance SSIM of 0.98–0.99, while significantly reducing voxel-wise distribution divergence (Wasserstein distance). Downstream tasks demonstrate marked improvements: brain age prediction error decreases from 5.36 to 3.30 years, and Alzheimer’s disease classification AUC rises from 0.78 to 0.85. Crucially, the approach preserves biological fidelity while delivering high-fidelity, interpretable cross-center image standardization.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Humans and AI: Brain-Sensing and Analysis

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

The variability introduced by differences in MRI scanner models, acquisition protocols, and imaging sites hinders consistent analysis and generalizability across multicenter studies. We present a novel image-based harmonization framework for 3D T1-weighted brain MRI, which disentangles anatomical content from scanner- and site-specific variations. The model incorporates a differentiable loss based on the Structural Similarity Index (SSIM) to preserve biologically meaningful features while reducing inter-site variability. This loss enables separate evaluation of image luminance, contrast, and structural components. Training and validation were performed on multiple publicly available datasets spanning diverse scanners and sites, with testing on both healthy and clinical populations. Harmonization using multiple style targets, including style-agnostic references, produced consistent and high-quality outputs. Visual comparisons, voxel intensity distributions, and SSIM-based metrics demonstrated that harmonized images achieved strong alignment across acquisition settings while maintaining anatomical fidelity. Following harmonization, structural SSIM reached 0.97, luminance SSIM ranged from 0.98 to 0.99, and Wasserstein distances between mean voxel intensity distributions decreased substantially. Downstream tasks showed substantial improvements: mean absolute error for brain age prediction decreased from 5.36 to 3.30 years, and Alzheimer's disease classification AUC increased from 0.78 to 0.85. Overall, our framework enhances cross-site image consistency, preserves anatomical fidelity, and improves downstream model performance, providing a robust and generalizable solution for large-scale multicenter neuroimaging studies.

Problem

• Reducing MRI scanner variability across multicenter studies
• Preserving anatomical fidelity while harmonizing image styles
• Improving downstream model performance in neuroimaging analysis

Innovation

• Disentangles anatomical content from scanner variations
• Uses SSIM loss to preserve biological features
• Enables harmonization across diverse scanners and sites