Evaluation of Real-Time Cardiovascular Flow MRI Using Compressed Sensing in a Phantom and in Patients With Valvular Disease or Arrhythmia.

Background: Real-time (RT) phase contrast (PC) flow MRI can potentially be used to measure blood flow in arrhythmic patients. Undersampled RT PC has been combined with online compressed sensing (CS) reconstruction (CS RT) enabling clinical use. However, CS RT flow has not been validated in a clinical setting.

Purpose: Evaluate CS RT in phantom and patients.

Study type: Prospective.

Population: Flow phantom (60 cycles/min: N = 10, 120 cycles/min: N = 12), sinus rhythm patients, no regurgitation (N = 20) or suspected aortic regurgitation (N = 10), arrhythmia patients (N = 10).

Field strength/sequence: 1.5 T, 2D gated PC, CS RT PC, RT cine with arrhythmia rejection.

Assessment: Phantom experiments tested the accuracy of CS RT cardiac output and peak flow rate at 60 and 120 cycles/min against gated PC. For sinus rhythm patients, cardiac output, peak flow rate, and regurgitation fraction in the ascending aorta and/or pulmonary artery were evaluated against gated PC. Cardiac output in patients with arrythmia was evaluated against RT cine with arrhythmia rejection.

Statistical tests: Bland Altman, correlation, Mann-Whitney test, Wilcoxon signed-rank test.

Results: Cardiac output bias ± SD for CS RT in the phantom was -0.0 ± 0.2 L/min (0.5 ± 3%, P = 0.76) at 60 cycles/min and 0.2 ± 0.3 L/min (4 ± 4%, P = 0.0016) at 120 cycles/min. Correspondingly, peak flow rate bias was -23 ± 6 mL/s (-7 ± 2%, P < 0.0001) and -73 ± 25 mL/s (-23 ± 4%, P < 0.0001). In patients, regurgitant fraction was -4 ± 0.5% (-23 ± 4%, P = 0.0025). Cardiac output bias in patients in sinus rhythm was -0.1 ± 0.5 L/min (-2 ± 10%, P = 0.99) (with regurgitation) and -0.3 ± 0.6 L/min (-5 ± 11%, P = 0.035) (without regurgitation). Peak flow rate bias was -60 ± 31 mL/s (-13 ± 6%, P < 0.0001) (with regurgitation) and -64 ± 32 mL/s (-16 ± 8%, P < 0.0001) (without regurgitation). Cardiac output bias was -0.4 ± 0.6 L/min (-9 ± 11%, P < 0.003) in arrhythmia patients.

Data conclusions: CS RT flow could potentially serve as a clinical tool for patients with or without valvular disease or arrhythmia, with accurate cardiac output and regurgitation fraction quantification.

Plain language summary: Accurate flow assessment is important in clinical evaluation of cardiac patients, but in the presence of irregular heart rhythm flow assessment is challenging. We have evaluated a new method using cardiac magnetic resonance imaging and real-time flow for blood flow assessment in cardiac patients. The method was tested against a reference method in a phantom flow model in low and high heart rates, and in cardiac patients with and without irregular heart rhythm and in different vessels. We found the cardiac magnetic resonance imaging real time flow method accurate and therefore promising for clinical implementation.

Evidence level: 1 TECHNICAL EFFICACY: Stage 1.