Journal of imaging informatics in medicine · 2026
Field-Strength Dependence of Automated Subcortical Volumetry: A Paired 3 T-7 T Benchmarking Study of BrainSuite, FastSurfer and NextBrain.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Emekli Emre, Demirel Burak Can, Emekli Esra
Аннотация
Automated segmentation pipelines are widely used for quantitative assessment of subcortical brain structures, yet it remains unclear whether measurements obtained at different magnetic field strengths are numerically interchangeable. This study evaluated the cross-field agreement and inter-software consistency of automated subcortical volumetry using paired 3T and 7T structural MRI data from 30 healthy adults. T1-weighted images were independently processed using BrainSuite, FastSurfer and NextBrain. Volumes of the bilateral putamen, globus pallidus, nucleus accumbens, thalamus and caudate nucleus were analysed. Within-software agreement between 3 and 7T measurements was assessed using two-way mixed-effects, single-measure intraclass correlation coefficients for absolute agreement. Inter-software agreement was examined separately at 3T and 7T using both absolute agreement and consistency intraclass correlation coefficients. All within-software regional comparisons and selected inter-software comparisons were further evaluated using Bland-Altman analyses. Cross-field agreement varied markedly across software packages and anatomical regions. BrainSuite showed poor absolute agreement across all regions, whereas FastSurfer and NextBrain demonstrated moderate-to-good agreement in selected structures. Inter-software consistency was generally higher than absolute agreement, indicating preservation of relative participant ranking despite systematic differences in measured volumes. Segmentation failures were more frequent at 7T, particularly for BrainSuite. Bland-Altman analyses confirmed software- and region-specific systematic differences and wide limits of agreement. These findings indicate that the evaluated automated segmentation pipelines do not provide field-strength-invariant subcortical volume estimates. Measurements obtained at 3T and 7T, or using different software packages, should not be considered directly interchangeable without pipeline-specific validation or harmonisation.