Highly Accelerated Aortic 4D Flow MRI: Implications for Pulse Wave Velocity Measurements.

Background: Aortic pulse wave velocity (PWV) is a biomarker of arterial stiffness. Although compressed sensing (CS)-accelerated 4D flow MRI reduces scan time, the influence of CS acceleration factors on the consistency of PWV measurements remains unclear.

Purpose: To evaluate the consistency of different PWV calculation methods under varying CS acceleration factors (R), identify the most robust method, and investigate the impact of CS acceleration on PWV measurements.

Study type: Prospective.

Population: Sixteen healthy adults (23-39 years, 8 males).

Field strength/sequence: A 3D Cartesian retrospective ECG-triggered 4D flow sequence, using GRAPPA (R = 2) and CS (R = 4.4, 7.7, 11.5, 15.3, 20.5) at 3 T.

Assessment: PWV was calculated from GRAPPA and CS datasets using five methods: time-to-foot (TTf), time-to-median (TTm), cross-correlation (Xcor), fitting plane (plane), and maximum likelihood estimation (MLE) across the thoracic aorta. Scan-rescan repeatability was also assessed.

Statistical tests: Paired t-test, Bland-Altman analysis, Levene's test, root-mean-square-error (RMSE), intraclass correlation coefficients (ICC), and Pearson's correlation coefficients (r). A p value < 0.05 was considered statistically significant.

Results: All CS reconstructions systematically overestimated PWV compared with GRAPPA R = 2 (bias +7.7% to +50.5%, all p < 0.05). Among the five evaluated PWV calculation methods, the TTf method exhibited the largest overestimation (+29.3% to +50.5% vs. GRAPPA R = 2; +3.2% to +16.3% vs. CS R = 4.4). In contrast, the Xcor method yielded much smaller biases (+8.6% to +13.4% vs. GRAPPA R = 2; -1.3% to +3.1% vs. CS R = 4.4) and demonstrated the highest scan-rescan repeatability (ICC up to 0.94 across CS accelerations).

Data conclusions: A systematic overestimation of PWV was observed in CS reconstructions compared with GRAPPA. Nevertheless, in our experiments, the Xcor method demonstrated the smallest bias and highest repeatability, which may enable more consistent PWV measurements when using high-acceleration CS 4D flow.

Evidence level: 2.

Technical efficacy: Stage 1.