Improvements in Between‐Vendor MRI Harmonization of Renal T2 Mapping using Stimulated Echo Compensation

Abstract

BackgroundT2 mapping is valuable to evaluate pathophysiology in kidney disease. However, variations in T2 relaxation time measurements across MR scanners and vendors may occur requiring additional correction.PurposeTo harmonize renal T2 measurements between MR vendor platforms, and use an extended‐phase‐graph‐based fitting method (“StimFit”) to correct stimulated echoes and reduce between‐vendor variations.Study TypeProspective.Subjects8 healthy “travelling” volunteers (37.5% female, 32 ± 6 years) imaged on four MRI systems across three vendors at four sites, 10 healthy volunteers (50% female, 32 ± 8 years) scanned multiple times on a given MR scanner for repeatability evaluation. ISMRM/NIST system phantom scanned for evaluation of T2 accuracy.Field Strength/Sequence3T, multiecho spin‐echo sequence.AssessmentT2 images fit using conventional monoexponential fitting and “StimFit.” Mean absolute percentage error (MAPE) of phantom measurements with reference T2 values. Average cortex and medulla T2 values compared between MR vendors, with masks obtained from T2‐weighted images and T1 maps. Full‐width‐at‐half‐maximum (FWHM) T2 distributions to evaluate local homogeneity of measurements.Statistical TestsCoefficient of variation (CV), linear mixed‐effects model, analysis of variance, student's t‐tests, Bland–Altman plots, P‐value <0.05 considered statistically significant.ResultsIn the ISMRM/NIST phantom, “StimFit” reduced the MAPE from 4.9%, 9.1%, 24.4%, and 18.1% for the four sites (three vendors) to 3.3%, 3.0%, 6.6%, and 4.1%, respectively. In vivo, there was a significant difference in kidney T2 measurements between vendors using a monoexponential fit, but not with “StimFit” (P = 0.86 and 0.92, cortex and medulla, respectively). The intervendor CVs of T2 measures were reduced from 8.0% to 2.6% (cortex) and 7.1% to 2.8% (medulla) with StimFit, resulting in no significant differences for the CVs of intravendor repeat acquisitions (P = 0.13 and 0.05). “StimFit” significantly reduced the FWHM of T2 distributions in the cortex and whole kidney.Data ConclusionStimulated‐echo correction reduces renal T2 variation across MR vendor platforms.Level of Evidence2Technical EfficacyStage 1