Accuracy of accelerated cine MR imaging at 3 Tesla in longitudinal follow-up of cardiac function

The ability of fast, parallel-imaging-based cine magnetic resonance (MR) to monitor global cardiac function in longitudinal exams at 3 Tesla was evaluated. Seventeen patients with chronic cardiac disease underwent serial cine MR imaging exams (n=3) at 3 Tesla. Data were acquired in short-axis orientation using cine steady-state free precession (SSFP) with a spatial resolution of 2.5 x 1.9 mm(2) at 45 ms temporal resolution. Multislice imaging (three slices/breath-hold) was performed using TSENSE acceleration (R=3) and standard single-slice cine (non-TSENSE) was performed at identical locations in consecutive breath-holds. End-diastolic volume (EDV), end-systolic volume (ESV), ejection fraction (EF) and myocardial mass (MM) of both cine approaches were compared for individual time-points as well as for longitudinal comparison. TSENSE-cine did not show significant differences for EDV (2.6 ml; P=0.79), ESV (2.2 ml; P=0.81), EF (-0.3%; P=0.95) and MM (2.4 g; P=0.72) in comparison with non-TSENSE. Longitudinal ANOVA analysis did not reveal significant differences for any parameter, neither for non-TSENSE data (all P>0.7) nor for TSENSE data (all P>0.9). Multifactorial ANOVA showed non-significant differences (all P>0.7) at comparable data variances. Data acquisition was significantly shortened using TSENSE. Threefold accelerated multislice cine at 3 Tesla allows accurate assessment of volumetric LV data and accurate longitudinal monitoring of global LV function at a substantially shorter overall examination time.