1348991 - 3D Blood Flow Measurements in the Coronary Arteries

Phase contrast (PC) MRI can be used to measure blood flow velocities in 3D volumes [1]. Flow measurements in the coronary arteries are particularly challenging due their tortuousness and small vessel diameters as well as cardiac and respiratory motion. Previous studies performed single-slice measurements with 1D velocity encoding, considering only small sections of the coronary arteries and yielding limited flow information [2,3,4]. The aim of this work was to acquire 3D PC measurements of the coronary arteries.

Methods:

3D PC measurements (VENC=50 cm/s, undersampling factor = 14, (Δx)³ = (1.2mm)³) were performed in diastole on 15 healthy volunteers (8 females, 59.5±12.9 years) on a 1.5 T MRI system (MAGNETOM Sola, Siemens, Erlangen) using a research application. T2 preparation and fat saturation were used to improve the contrast between the coronary arteries and surrounding tissue. Motion effects were mitigated through ECG triggering and navigator-based respiratory gating. Subject-specific optimum trigger delays were obtained from preceding single-slice cine measurements, ensuring minimum cardiac motion during data acquisition. A 3D post-processing and visualization research tool was developed for qualitative flow analysis and velocity quantification in slices perpendicular to the centerlines of the left (LCA) and right (RCA) coronary arteries. Inter-scan variability was analyzed by repeating the scan in 12 volunteers within the same scan slot.

Results:

Figure 1 shows coronary velocity maximum intensity projections (MIP) (a-d) as well as maximum flow velocities along the centerlines (e-f). Along the LCA, flow velocity peaks shortly downstream the ostium and then decreases distally. No general trend can be observed for the RCA.
Figure 2 shows MIP images obtained from two successive scans of volunteer #7 (a,b) and the corresponding flow velocities (c). Qualitative similarities can be observed while quantitative differences remain evident. Inter-scan variability is further evaluated in Table 1, giving peak flow velocities in the proximal arteries (up to 25mm along the centerline). For all volunteers except #8 and #10, the peak velocity in the LCA exceeds that of the RCA. The respective mean peak velocities for the LCA and RCA are 15.0±3.2cm/s and 9.6±1.6cm/s. The mean deviations in peak velocities along the LCA and RCA between the first and second scan are 7±22% and 2±20% respectively. The interplay between physiology (heart rate variation, pulsatile coronary flow, stenoses, …) and scan parameters (trigger delay, acquisition window, sampling pattern, reordering) is expected to have substantial impact on the quantification and inter-scan variability results and is subject of future investigation.

Conclusion:

This study showed the feasibility of 3D flow measurements in the coronary arteries of healthy volunteers. Although technical hurdles remain, a non-invasive measurement of hemodynamic relevance of coronary artery stenoses in the proximal branches might be feasible in the future.