Quantitative Inter-Vendor Evaluation of PDFF and R2* Mapping Accuracy for 3T MRI Scanners
Presentations
TH-IePD-TRACK 2-6 (Thursday, 7/29/2021) 12:30 PM - 1:00 PM [Eastern Time (GMT-4)]
Purpose: To evaluate the relative accuracy of proton density fat fraction (PDFF) and R2* mapping sequences at 3T from different MRI scanner manufacturers and pulse sequence types using standardized quantitative MRI phantoms.
Methods: PDFF and R2* phantoms manufactured by Calimetrix were scanned twice on GE PETMR, GE 750W, Siemens Vida, and Siemens Skyra platforms using the body coil and the two measurements averaged. For GE, IDEAL-IQ sequences were used to generate the fat fraction and R2* maps. For Siemens, product q-Dixon sequences were used, as well as a work-in-progress (WIP) q-Dixon sequence on the Skyra. MR parameters were kept at manufacturer’s defaults. Data was processed by drawing a 1.6cm diameter circular ROI over the phantom vials DICOM images and recording the mean ROI value.
Results: IDEAL-IQ sequences had lower overall average percent error for PDFF calculation, with 750W and PETMR measuring 2.6% and 7.1% error respectively. Skyra WIP q-Dixon had 4.8% error for PDFF; Vida product q-Dixon had 16.1% error. Siemens scanners were more accurate than GE for R2* measurements: Skyra WIP q-Dixon measured 7.1% error, and Vida product q-Dixon measured 7.8% error. In contrast, 750W and PETMR measured 14.6% and 17.7% error, respectively. Data from the Skyra product q-Dixon sequence was not accurate for both PDFF and R2* calculations, with an average error of 72% and 59%, respectively. Greater inaccuracy was observed at the high and low ends of the phantom’s PDFF and R2* scale for both vendors, suggesting that extreme values are more challenging to reconstruct accurately.
Conclusion: We observe modest accuracy for PDFF and R2* mapping sequences on both GE and Siemens MRI scanners, suggesting further refinement of reconstruction algorithms is possible. These results also highlight the importance of validation of quantitative results before integration into clinical workflows.
ePosters
Keywords
Taxonomy
IM- MRI : Quantitative Imaging
Contact Email