Abstract
T1 mapping and extracellular volume (ECV) calculations in cardiac magnetic resonance (CMR) have the potential to identify early fibrosis that is not yet visible using late gadolinium enhancement; however, the need for same-day blood draws due to the temporal variations in hematocrit (Hct) limits the use of ECV. We aimed to determine the reproducibility of synthetic Hct and ECV using different published models among groups of subjects. Healthy subjects and those with diagnosed cardiac amyloidosis, sarcoidosis, and hypertrophic cardiomyopathy (HCM) scanned using a 1.5T scanner with native and post-contrast T1 maps and same-day Hct were included. Among 148 subjects, there was excellent reproducibility (all ICCs ~0.98) between synthetic and measured ECV across the six formulas, despite only modest reproducibility of synthetic/measured Hct (ICCs 0.52-0.66). The levels of accuracy predicting abnormal measured ECV were consistently excellent among the different synthetic ECV models. The difference in the CMR vendor used to generate models did not seem to affect the results of the comparisons. We conclude that synthetic ECV yielded excellent reproducibility compared with ECV calculated using measured hematocrit, possibly obviating the need for blood extraction in cardiac MRI settings without point-of-care Hct.