Reproducibility of Structural and Diffusion Tensor Imaging in the TACERN Multi-Center Study

Area of Science:

• Neuroimaging
• Quantitative MRI
• Multi-site studies

Background:

• Multi-site MRI studies are crucial for rare disease research but face challenges in data pooling due to scanner variability.
• Evaluating reproducibility across scanners is critical for reliable multi-site quantitative MRI data.

Purpose of the Study:

• To assess the reproducibility of structural and diffusion-weighted (DW) MRI measurements across seven scanners at five medical centers.
• To establish variability benchmarks for quantitative MRI metrics in the Tuberous Sclerosis Complex Autism Center of Excellence Research Network (TACERN) study.

Main Methods:

• Utilized the American College of Radiology (ACR) phantom for monthly signal intensity and uniformity assessments across seven 3T scanners (GE, Philips, Siemens).
• Imaged a healthy volunteer repeatedly over 5 years using the TACERN structural and DW MRI protocol.
• Quantified inter- and intra-scanner reproducibility of brain volumes and diffusion tensor imaging (DTI) metrics (FA, MD) using coefficient of variation.

Main Results:

• ACR phantom signal intensity and uniformity showed low variability (mean intra-scanner CV: 3.6% and 1.8%).
• Brain structure volumes exhibited high reproducibility (mean inter-scanner CV: 3.3%, intra-scanner CV: 1.1%).
• Diffusion tensor imaging metrics demonstrated good reproducibility (FA: inter-scanner CV 4.5%, intra-scanner CV 2.5%; MD: inter-scanner CV 5.4%, intra-scanner CV 1.5%).

Conclusions:

• Volumetric and DTI measurements are highly reproducible within and between the evaluated scanners.
• The study provides typical variation amplitudes serving as valuable references for future TACERN network findings.
• Findings support the reliability of multi-site quantitative MRI data for rare disease research.