MRI-derived Right Ventricular Global Longitudinal Strain Predicts Heart Failure.

Abstract

Purpose To evaluate the association of right ventricular (RV) strain metrics derived from feature-tracking cardiac MRI with incident heart failure (HF). Materials and Methods This retrospective analysis included data of 45 700 participants (23 764 female) from the UK Biobank Imaging Study between August 2006 and July 2010. Cardiac MRI scans were performed using standardized UK Biobank protocols and equipment. Image analysis was performed with feature-tracking module of a certified postprocessing software (cvi42 prototype version 5.13.7). Biologically implausible and statistical outliers were excluded. The association of each RV global strain marker (global longitudinal strain [GLS], global circumferential strain [GCS], and global radial strain [GRS]) with incident HF was assessed using univariable and multivariable Cox regression models adjusted for established clinical features and RV imaging markers with known prognostic values. Results The mean age ± SD of the 45 700 study participants (23 764 female) was 65 years ± 7.7, and median follow-up was 3 years. Lower absolute RV GLS, RV GCS, and RV GRS were associated with increased HF risk (hazard ratio [HR]: 1.34, 1.37, and 0.71, respectively; P < .001 for all). According to multivariable analyses adjusted for conventional RV imaging markers and clinical covariates, only RV GLS independently predicted HF (HR: 1.16; 95% CI: 1.02, 1.33; P = .04). Conclusion In the general population, RV GLS, RV GCS, and RV GRS are independently predictive of HF after accounting for clinical features. RV GLS demonstrated independent predictive value against conventionally used RV imaging markers and clinical features. Keywords: Cardiac MRI, Right Ventricle, Myocardial Strain, Heart Failure, Survival Analysis Supplemental material is available for this article. © RSNA, 2025.