Beyond the LUMIR challenge: The pathway to foundational registration models

TL;DR

This paper introduces the LUMIR challenge, a large-scale benchmark for unsupervised brain MRI registration, demonstrating deep learning's state-of-the-art performance and robustness across diverse neuroimaging tasks.

Contribution

It presents a new large-scale, unlabeled MRI dataset and a comprehensive evaluation framework for deep learning-based registration, advancing towards foundational models in medical imaging.

Findings

Deep learning methods achieved state-of-the-art registration accuracy.

Models produced anatomically plausible, diffeomorphic deformation fields.

Deep learning methods outperformed traditional optimization-based approaches and were robust to domain shifts.

Abstract

Medical image challenges have played a transformative role in advancing the field, catalyzing innovation and establishing new performance benchmarks. Image registration, a foundational task in neuroimaging, has similarly advanced through the Learn2Reg initiative. Building on this, we introduce the Large-scale Unsupervised Brain MRI Image Registration (LUMIR) challenge, a next-generation benchmark for unsupervised brain MRI registration. Previous challenges relied upon anatomical label maps, however LUMIR provides 4,014 unlabeled T1-weighted MRIs for training, encouraging biologically plausible deformation modeling through self-supervision. Evaluation includes 590 in-domain test subjects and extensive zero-shot tasks across disease populations, imaging protocols, and species. Deep learning methods consistently achieved state-of-the-art performance and produced anatomically plausible, diffeomorphic deformation fields. They outperformed several leading optimization-based methods and remained robust to most domain shifts. These findings highlight the growing maturity of deep learning in neuroimaging registration and its potential to serve as a foundation model for general-purpose medical image registration.