Multicenter and multi-vendor reproducibility of T1, T2 and ADC phantom data on 1.5T and 3T MRI scanners

This multicentre study aimed at assessing multi-vendor accuracy and reproducibility of typical MR biomarkers on 3T and 1.5T scanners from 5 clinical centers. MRI acquisitions were performed using NIST protocols, on Essential System and Diffusion NIST phantoms. T1 was measured by 3-parameter fitting, while T2 and ADC by mono-exponential fitting, using PhantomViewer software. When compared with reference values, non-negligible discrepancies were found across vendors, especially for T1 and T2 measurements. Reproducibility was vendor-dependent, and higher on 3T than on 1.5T scanners. ADC and T1 showed highest reproducibility. Corrections are likely needed to account for possible temperature and scanner differences.