Reproducibility of T1 and T2 relaxation times calculated from routine MR imaging sequences: phantom study

Measurement of T1 and T2 relaxation times has been sought as one fundamental way to characterize tissue. Relaxation times can be calculated from routine spin-echo (SE) imaging sequences using two distinct repetition times (TRs), each with two SE samplings of signal intensity. Previous reports have quantified relaxation times without discussing the variation in their measurements. By imaging a phantom containing different samples with known T1 and T2 relaxation times on three separate occasions, the variation in relaxation time measurements inherent in different routine imaging sequences was studied. For the present study a more complete and accurate equation was used to calculate T1 values. The variation in T1 and T2 relaxation times for samples with relaxation times similar to solid tissue was 2%-4%. The amount of variability in calculated relaxation times was found to be dependent on the magnitude of the relaxation times themselves. However, the mean values were independent of the imaging sequences used to calculate the relaxation times. No significant differences were seen between left-to-right or section-to-section position within the same study or between studies performed on different occasions. The variability in the calculated T1 was dependent on the pair of TR sequences used to calculate T1. Samples with long T1 and T2 relaxation times, similar to many body fluids, had much larger variability. A computer simulation of measurement error was created to explain these results. This study indicates that properly performed routine imaging studies do yield reproducible T1 and T2 measurements.