DISARM++: Beyond scanner-free harmonization

2025-05-06

AI Summary

To address unreliable feature extraction in multi-center T1-weighted MRI caused by scanner-specific biases, this paper proposes an end-to-end cross-scanner harmonization method. Methodologically, it introduces a novel dual-path generative architecture that jointly integrates scanner-decoupled representation learning, domain adaptation, and geometric invariance constraints—enabling direct mapping to an unbiased latent space or adaptation to a target scanner domain without error-prone preprocessing (e.g., skull-stripping) and supporting zero-shot generalization. Key contributions include: (i) the first harmonization framework that is preprocessing-free, plug-and-play, and requires no retraining for new scanners; and (ii) substantial improvements in downstream tasks: brain age prediction achieves R² = 0.60 ± 0.05, Alzheimer’s disease classification attains accuracy = 0.86 ± 0.03 and diagnostic AUC = 0.95—consistently surpassing current state-of-the-art methods.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Harmonization of T1-weighted MR images across different scanners is crucial for ensuring consistency in neuroimaging studies. This study introduces a novel approach to direct image harmonization, moving beyond feature standardization to ensure that extracted features remain inherently reliable for downstream analysis. Our method enables image transfer in two ways: (1) mapping images to a scanner-free space for uniform appearance across all scanners, and (2) transforming images into the domain of a specific scanner used in model training, embedding its unique characteristics. Our approach presents strong generalization capability, even for unseen scanners not included in the training phase. We validated our method using MR images from diverse cohorts, including healthy controls, traveling subjects, and individuals with Alzheimer's disease (AD). The model's effectiveness is tested in multiple applications, such as brain age prediction (R2 = 0.60 pm 0.05), biomarker extraction, AD classification (Test Accuracy = 0.86 pm 0.03), and diagnosis prediction (AUC = 0.95). In all cases, our harmonization technique outperforms state-of-the-art methods, showing improvements in both reliability and predictive accuracy. Moreover, our approach eliminates the need for extensive preprocessing steps, such as skull-stripping, which can introduce errors by misclassifying brain and non-brain structures. This makes our method particularly suitable for applications that require full-head analysis, including research on head trauma and cranial deformities. Additionally, our harmonization model does not require retraining for new datasets, allowing smooth integration into various neuroimaging workflows. By ensuring scanner-invariant image quality, our approach provides a robust and efficient solution for improving neuroimaging studies across diverse settings. The code is available at this link.

Problem

• Harmonizing T1-weighted MR images across different scanners for consistency
• Enabling image transfer to scanner-free or specific scanner domains
• Improving reliability and accuracy in neuroimaging without extensive preprocessing

Innovation

• Maps images to scanner-free space
• Transforms images to specific scanner domain
• Eliminates need for skull-stripping preprocessing