Room: AAPM ePoster Library
Purpose: To estimate the variability and accuracy of T1 and T2 mapping using MRI phantoms with known T1 and T2 values on Siemens MRI platforms.
Methods: A prostate Phantom with 6 T1/T2 inserts (T1/T2 ranging 126-1989ms/32-278ms respectively) and a NIST Phantom with 14 T1/T2 inserts (T1/T2 ranging 22-1989ms/6-581ms respectively)(both QalibreMD Boulder, CO, USA) were scanned on Siemens 3T Skyra and Vida MRI scanners (Siemens Healthineers, Erlangen, Germany). The phantoms were scanned using 18 and 30 channel body coils. 2D T1 and T2 maps of the phantoms were obtained using standard clinical T1/T2 mapping sequences. T2 mapping was performed using spin-echo multi-contrast (SE-MC) sequences, T1 mapping using turbo spin-echo inversion recovery (TSE-IR) sequences. Image analysis was performed in Siemens Syngo.via.
Results: For T2 measurements, we observe large deviation from the values listed in the manufacturer’s manuals of both phantoms. T2 quantification performed poorly on the high T2 tubes in the prostate phantom; in the NIST phantom, results were poor for the highest and lowest T2 tubes. Changing the number of contrasts greatly improves accuracy: increasing contrasts from the default 6 to 30 improved maximum error in the prostate phantom from 471% to 57% on Vida. 15 contrasts yielded best results on the NIST phantom.
On both phantoms, T1 measurements on long T1 inserts were discrepant between Skyra and Vida by up to 25%. In the NIST phantom, Skyra T1 measurements were higher; for the prostate phantom, Vida measurements were higher. T1 values overall were slightly higher than the manuals’ stated values. There was little change in T1/T2 measurements when the coil was changed.
Conclusion: The variability of measurements of T1 and particularly T2 indicates that accuracy and reproducibility of quantitative MRI is highly dependent on pulse sequence parameters, such as the number of echoes acquired, even when scanning standardized phantoms.
MRI, Quantitative Imaging, Tissue Characterization