Heart Rate related to Accuracy of left ventricular strain parameters derived from the single-shot compressed sensing cine CMR images of patients with cardiovascular disease

Abstract

Objectives Left ventricular (LV) global strain values from single-shot compressed sensing (ss-CS) cine imaging of cardiovascular magnetic resonance (CMR) are significantly underestimated compared with those from conventional segmented cine imaging (regarded as gold standard). This study aimed at exploring the patient-related factors of this underestimation and finding a relatively easy way of adjustment in clinical cardiovascular patients. Methods Patients with ss-CS cine and standard segmented cine CMR were retrospectively enrolled. Different groups in terms of LV wall thickness (thinned < 6mm, normal 9-12mm and thickened ≥ 15mm) and ejection fraction (EF) were established. Absolute and relative differences of LV global peak strains (GPSs) were calculated. Stepwise multivariable linear regression models were used to find the independent patient-related factors influencing the difference. Results Among 62 patients (mean age, 50 ± 16 years; heart rate, 71 ± 13 beats/min), patients with normal, thinned and thickened wall thickness were 19, 22 and 21 respectively and with LV systolic dysfunction (EF < 50%) were 31. LV GPSs by ss-CS cine were significantly underestimated. Absolute underestimations of GPSs were significantly different among different wall thickness groups and EF groups respectively (all P < 0.05), while relative underestimations of GPSs were not related to LV wall thickness or EF. In stepwise multivariable analysis, EF and heart rate were both associated with absolute decline (β values: from 0.244 to 0.485, P < 0.05). However, only heart rate was independently assassinated with relative underestimation in global circumferential and longitudinal strains (circumferential: β value = 0.320, P = 0.011; longitudinal: β value = 0.388, P = 0.002). Conclusions Compared to the traditional segment cine, ss-CS cine indeed underestimates LV strain parameters in cardiovascular patients and relative underestimations in GPSs could be adjusted by heart rate. Improving and standardized imaging techniques for high-speed CMR imaging are still needed.