T1/T2 Mapping across 3T Scanner Fleet Using Ismrm / NIST Model 130 Premium System Phantom
Abstract
Purpose
To establish a rapid quality-control (QC) methodology and baseline performance metrics for T1 and T2 relaxation time estimation across a multi-vendor fleet of 3T MRI systems.
Methods
Scans of the ISMRM/NIST Model 130 Premium System Phantom (Caliber MRI, Boulder, CO) were acquired on multiple 3T MRI scanners from GE and Siemens. Variable TI (TI range 35-4000ms) acquisitions were used for T1 mapping and multi– echo spin-echo (TE range 10-320ms) acquisitions for T2 mapping. Other parameters were matched to the phantom manual (Rev. L). Semi-automated 10 pixel-diameter ROI placement, measurement, and analysis were performed using Python. T1 analysis used a nonlinear least-squares fit to a magnitude inversion-recovery model with an inversion efficiency term. T2 analysis used a mono-exponential decay model with a constant noise offset. For T2, spheres 1 and 5 were excluded per phantom vendor recommendations. Spheres 13 and 14 were omitted for both data sets. For Siemens systems, the first echo (10 ms) was excluded. Estimated relaxation times were compared to temperature-corrected phantom reference values.
Results
Percent deviation [((calculated_T1 – reference_T1)/reference_T1) x 100] was calculated for each sphere. T1 measurements were generally within ±7% of reference values across all scanners. T2 measurements exhibited greater variability, with most values falling within approximately ±30% of nominal values. Data fits from Siemens tended to overestimate T2, while those from GE underestimated.
Conclusion
This abbreviated scan technique QC method yielded consistent results for T1 calculation across all scanners. T2 behavior differed between GE and Siemens and generally had greater deviation from reference values. These results establish baseline performance metrics that can support routine MRI QC and longitudinal consistency monitoring across the scanner fleet. Future work will involve determining sources of error and improving T2 accuracy. This study shows promise for validation of clinical quantitative measurements.