A compressed sensing multi-slice cine CMR approach for the accurate assessment of left ventricular volumes and function

Background

CMR is generally accepted as the gold standard for left ventricular (LV) volumes and function assessment. The conventional CMR approach involves several breath-holds to cover the entire heart with short-axis acquisitions. Recently, compressed sensing (CS) techniques emerged as a means to considerably accelerate data acquisition. CS principally relies on: 1) transform sparsity, 2) incoherence of undersampling artifacts, and 3) nonlinear reconstruction. PURPOSE: To compare a novel CS-based single breath-hold multi-slice cine technique with the standard multi-breath-hold technique for the assessment of LV volumes and function.

Methods

The novel single breathhold multi-slice CS cine sequence was implemented on a 1.5T MAGNETOM Aera (Siemens Healthcare, Germany). It realizes incoherent sampling by distributing the readouts pseudo-randomly on the Cartesian k space grid and by applying a pseudo-random offset from frame-to-frame during prospective ECG-triggering. Thus, 3 long-axis and 4 short-axis slices (Figure 1A) were acquired in a single breath-hold of 14 heart beats (temporal/spatial resolution: 30 ms/1.5 × 1.5 mm2, respectively, acceleration factor: 11.0). The CS cine data were analyzed by the Siemens Argus 4DVF software (Figure 1B) which is based on a 3D LV-model that takes the cyclic motion of the mitral valve plane (yellow plane in 2B) into account. For comparison, a conventional stack of cine SSFP images was acquired (temporal/spatial resolution 40 ms/1.2 × 1.6 mm2, slice thickness: 8 mm, gap: 2 mm) and analyzed by the Siemens Argus VF software. As a reference for the LV stroke volume (LVSV), the aortic flow (AoFlow) was measured by a phase-contrast acquisition (temporal/spatial resolution 40 ms/1.8 × 1.8 mm2).

Figure 1

Results

Twelve volunteers (75% male, age 33 ± 7 y) and 14 patients (85% male, age 67 ± 10 y) with different LV pathologies were studied. Good image quality of the single breathhold multi-slice CS acquisitions was obtained in 10/12 volunteers and 12/14 patients and all CS data were of adequate quality for 4D-analysis. LVEFCS and LVEFStandard were similar (49.3 ± 16.4% vs 50.6 ± 16.4%, respectively, p = 0.17) with an excellent correlation (Figure 2). Correlation of LVSVCS with AoFlow was also excellent (r = 0.93, p < 0.0001). For LVSVStandard the correlation was r = 0.70 (p < 0.01). Bland-Altman analyses showed underestimation of CS vs standard by 10.5 ml and 7.2 ml for LV end-diastolic (p < 0.005) and LV stroke volume (p < 0.002), respectively, no difference for LV end-systolic volume (p = ns).

Figure 2

Conclusions

This is the first clinical experience with a novel CS-based cine sequence. These preliminary results demonstrate the potential of this fast single breath-hold technique to replace the multi-breath-hold standard technique for the assessment of LV volumes and function.

Funding

D. Piccini, M.Zenge, M.Schmidt and M. Nadar are Siemens employees. Other funding than that was not received.