Pooling quantitative MRI data: A multi-protocol study of healthy subcortical aging

Area of Science:

• Neuroimaging
• Biophysics
• Medical Physics

Background:

• Quantitative MRI (qMRI) offers biophysical tissue properties, ideally independent of scanner and protocol.
• Current qMRI methods exhibit variability across sites and acquisition schemes, hindering data pooling.
• Pooling data enhances statistical power for longitudinal and cross-sectional studies, especially in aging research.

Purpose of the Study:

• To investigate the impact of protocol and hardware differences on pooling large ultra-high-field (UHF) qMRI data in healthy aging.
• To evaluate how pooling affects age-dependent qMRI parameters and identify protocol-related biases, particularly in subcortical structures.
• To assess the combined effects of data pooling and protocol variations on age-related changes in qMRI metrics.

Main Methods:

• Combined an open-access aging UHF qMRI MP2RAGE dataset with two multiparameter mapping (MPM) datasets.
• Evaluated age dependence of qMRI parameters (R1, R2*, volume) in pooled and individual datasets.
• Analyzed protocol-specific biases and their influence on inferred age-related changes, focusing on subcortical structures.

Main Results:

• Pooled data revealed age-related R1 changes of 4-17% and R2* variations of 6-30% across structures.
• Subcortical structure volume changes ranged from 5-26% over the lifespan in the pooled dataset.
• R1 and volume showed larger protocol-dependent differences, while R2* remained more consistent across regions.

Conclusions:

• Pooling UHF qMRI data from diverse sources can be both beneficial (statistical power) and detrimental (protocol bias) for aging analyses.
• Protocol standardization is crucial for reliable data pooling in multi-site qMRI studies.
• Understanding and quantifying protocol effects is essential for accurate interpretation of age-related changes in qMRI parameters.