VISTA: Variance-Gated Inter-Sequence Test-Time Adaptation for Multi-Sequence MRI Segmentation

2026-05-17

AI Summary

This work addresses the performance degradation of multi-sequence MRI segmentation models when deployed in new clinical settings, primarily caused by differences in scanners and acquisition protocols that induce inter-modal interaction shifts. To tackle this challenge, the authors propose VISTA, a source-free test-time adaptation framework that introduces an interventional consistency probe and a variance-gated pseudo-labeling strategy. Specifically, intervention samples are generated by cross-sequence swapping of low-frequency spectral components and entropy-localized patches, while voxel-wise reliability is dynamically estimated through cross-view disagreement to guide self-training. Evaluated on domain shifts from adult MRI to both African low-field and pediatric datasets, VISTA achieves Dice score improvements of 1.89% and 2.82%, respectively, significantly outperforming existing methods.

Abstract

Deploying multi-sequence magnetic resonance imaging (MRI) segmentation models to new clinical environments is challenging due to variations in scanners and acquisition protocols. Although existing TTA methods handle basic per-modality shifts, they often fail under a fundamental dual-shift problem, as their adaptation signals fail to capture modality-interaction shifts that disrupt inter-sequence consistency. To address this, we propose Variance-gated Inter-Sequence Test-time Adaptation (VISTA), a source-free framework that tackles modality-interaction shifts. First, we design an Inter-Sequence Intervention Generator (ISIG) that generates a set of consistency probes by swapping low-frequency spectra and entropy-localized patches across sequences, preserving anatomical semantics while challenging inter-sequence dependencies. Second, we introduce Cross-View Disagreement-Aware Pseudo Labeling (CDPL), which establishes a voxel-wise reliability metric using cross-view disagreement variance to dynamically gate self-training and enforce interventional consistency, encouraging the network to rely on robust anatomical semantics. Extensive experiments adapting from standard adult MRI (BraTS-GLI-Pre) to African low-field (BraTS-SSA) and pediatric (BraTS-PED) cohorts show improved performance over competing methods under clinical shifts, achieving absolute Dice improvements of +1.89% (SSA) and +2.82% (PED) over the source model. The code is available at https://github.com/dzp2095/VISTA.

Problem

• multi-sequence MRI
• test-time adaptation
• domain shift
• inter-sequence consistency
• modality-interaction shift

Innovation

• Test-Time Adaptation
• Multi-Sequence MRI
• Inter-Sequence Consistency
• Source-Free Domain Adaptation
• Pseudo Labeling