Performance of respiratory gated 4D flow MRI with adaptive k-space reordering in healthy controls and aortic dissection: reproducibility and agreement with 2D phase contrast MRI

A four-dimensional phase-contrast magnetic resonance imaging sequence with respiratory-controlled adaptive k-space reordering (ReCAR-4DPC) offers potential benefits of improved scan efficiency and motion robustness. The purpose of this study was to evaluate the reproducibility of flow measurement using this technique and to compare hemodynamic metrics obtained to two-dimensional phase contrast MRI (2DPC)-derived metrics of the thoracic aorta. ReCAR-4DPC was performed with identical scan parameters in 15 healthy volunteers (6M,9F, mean [range] 37 [23-47] years) and 11 patients with thoracic aortic dissection (6M,5F, 56 [31-81] years) and acquisition time was recorded. Peak systolic velocity (PSV), average flow (AF) and net forward volume (NFV) were quantified by two readers for ReCAR-4DPC at ascending, descending and diaphragmatic aorta levels. Reference standard 2DPC measurements at the same levels were performed by a separate experienced cardiovascular radiologist. ReCAR-4DPC intra-reader agreement, inter-reader agreement, inter-scan repeatability and concordance with 2DPC-derived metrics (all segments combined) were evaluated with Lin's concordance correlation coefficient (LCCC) and reduced major axis regression. The overall average ± SD MRI acquisition time of all subjects was 11:59 ± 3:57 min, with shorter average times (9:37 ± 1:57 min) in healthy volunteers compared to patients (15:13 ± 3:44 min). There was near-perfect intra-reader, inter-reader and inter-scan concordance (LCCC for all metrics > 0.97, > 0.98 and > 0.92 respectively) for ReCAR-4DPC. Concordance with 2DPC was also high (LCCC all > 0.89), with overall minimally lower PSV, AF and NFV values derived from ReCAR-4DPC compared to reference 2DPC derived metrics. ReCAR-4DPC is a reproducible and relatively fast approach for comprehensive measurement of thoracic aortic flow metrics, with robust correlation to conventional 2DPC.