Myocardial Mass Corrected CMR Feature Tracking-Based Strain Ratios are Different in Pathologies With Increased...

Acute myocarditis (AM) and hypertensive heart disease (HHD) have different pathophysiological backgrounds, thus potentially showing distinct patterns of altered myocardial deformation. Therefore, CMR left ventricular (LV) feature tracking (FT)- based strain parameters were indexed to myocardial mass index (LVMi) in order to evaluate potential additional value in the differentiation among AM, HHD, and healthy volunteers (HV) compared to non-indexed conventional strain.

Patients with AM (n = 43) and HHD (n = 28) underwent CMR at 3T. 61 HV served as controls. Cine imaging-based FT-strain analysis was performed and natural strain (nStrain) values were evaluated for gender and age specific differences in HV. Strain parameters were indexed to LVMi yielding ratio Strain (rStrain). These were evaluated for their discriminatory accuracy compared to nStrain values.

There were significant differences in nStrain between genders (p < 0.05), but not between age groups in HV. Circumferential strains differentiated best between HV and AM, reaching an area under the curve (AUC) of 0.86 (female) and 0.81 (male), yielding 93 (72) % sensitivity and 55 (75) % specificity. In discriminating between HV and HHD as well as AM and HHD, longitudinal strains outperformed all other parameters with AUCs of 1.00 (female)/ 0.92 (male) and 0.90 (female)/ 0.74 (male), respectively. Sensitivity and specificity levels of 100 %/ 100 % (female) and 91 %/ 72 % (male) for HV versus AM as well as 82 %/ 71 % (female) and 91%/ 57 % (male) for AM versus HHD could be demonstrated. The usage of rStrains significantly increased the AUC for circumferential and radial strains in male patients.

rStrain provided additional value in the differentiation of diseases with increased LVM. As rStrain is derived from standard native cine imaging, such parameters can be time efficiently and reliably calculated, giving them the potential to be a powerful addition to the currently developing multiparametric native diagnostic approaches.