Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
To address the pervasive failure modes of whole-brain registration metrics in neurovascular visualization, this work establishes an ROI-centered MRI-MRA registration benchmark using 149 trigeminal neuralgia patient datasets with expert-annotated nerve targets. Across six representative registration pipelines, the authors show that standard image-similarity metrics fail to co-rank with downstream vessel localization, and one-sided vessel distances are easily confounded by partial annotation volumes. Crucially, deformable refinement yielded only marginal, field-of-view-dependent improvements over affine alignment while frequently producing globally implausible warps that went undetected by local distance metrics.
Deformable registration barely outperforms simple affine baselines for neurovascular MRI-MRA fusion, frequently generating anatomically nonsensical deformations that easily trick standard image-similarity metrics.
Preoperative evaluation of trigeminal neuralgia (TN) often requires joint interpretation of structural MRI, which depicts the trigeminal nerve and surrounding cisternal anatomy, and time-of-flight MRA, which highlights vascular structures. Although MRI-MRA fusion is clinically attractive for visualizing neurovascular compression, this task is poorly captured by conventional whole-brain registration evaluation because the clinically relevant target is a small trigeminal ROI, vessel annotations are partial and clinically focused, local TOF-MRA contrast is variable, and field-of-view mismatch can limit deformable alignment. We formulate TN MRI-MRA fusion as an ROI-centered neurovascular registration-evaluation problem and construct a benchmark from 149 patients with clinician-annotated bilateral trigeminal ROIs. Six representative registration pipelines were evaluated using local image-based metrics, segmentation-derived vessel-localization metrics, prediction-volume analysis, and contrast- and FOV-stratified comparisons. Conventional evaluation summaries were often misleading: local image similarity, vessel-background separability, and downstream vessel localization did not co-rank methods; one-sided vessel distances were strongly affected by predicted vessel extent under partial annotations; and local MRA contrast determined when vessel-separability metrics were informative. Deformable refinement provided only a small, FOV-dependent benefit over affine alignment, while reader review showed that locally favorable vessel distances could coexist with globally implausible registrations. These findings indicate that TN MRI-MRA registration should be evaluated as a local, vessel-aware, contrast-sensitive, and FOV-aware visualization task rather than as generic multimodal brain registration. Our code is publicly available at https://github.com/jhuldr/TN-Reg-Benchmark.
Johns Hopkins