Assessment of Image Quality of a Coronary MR Angiography Pulse Sequence with 100% Respiratory Scan Efficiency using 3D-PROST Reconstruction (RF_SA_444)

Coronary MRA with 100% respiratory scan efficiency using 2D image navigator (iNAV) and motion corrected reconstruction including patch based low-rank (PROST) regularisation is a promising pulse sequence for evaluating coronary artery disease. However, its effectiveness in assessing coronary segments remains underexplored. This study aims to evaluate the visibility, diagnostic image quality, and inter-rater agreement in assessing various coronary segments. 

Methods:

This prospective study included 37 patients consisting of healthy participants (n=5), and patients with acute (n=18) and long (n=14) COVID-19. All participants underwent b-SSFP coronary MRA with 1.0 mm isotropic resolution at 1.5T (Siemens Sola) immediately after administration of gadolinium between February 2022 and March 2024. No preprocedural medications, including beta blockers or nitroglycerin, were administered. Using the AHA 13-segment classification, four proximal (RCA, LAD, LM, LCX) and two mid segments (RCA, LAD), three distal segments (RCA, LAD, LCX), and four branch segments (DM1, DM2, OM, PDA) were evaluated. Coronary segments were independently assessed by a radiologist and a medical fellow evaluating the number of visible segments and the diagnostic image quality on a Likert scale of 1(not visible) to 5(best), with 3 defined as diagnostically acceptable. Weighted Kappa analysis was employed to assess interrater agreement on image quality. 

Results:

A total of 37 patients (22 men and 15 women; mean age, 46 ± 17 years; mean heart rate, 66.1 ± 13.4 bpm) successfully underwent coronary MRA (mean scan time = 7.39 ± 2.07min) were included in the study, with 13 coronary segments assessed per patient, resulting in a total of 481 segments evaluated (Figure 1 for representative renderings). The visibility rate was highest in proximal segments (93.9%), with a slight decrease in mid-segments (91.8%). Visibility declined significantly in distal (53.1%) and branch (14.8%) segments (Figure 2a). Similarly, image quality scores dropped from a mean of 3.63 ± 1.09 in proximal segments to 3.22 ± 1.01 in mid segments, 2.04 ± 1.01 in distal segments, and 1.36 ± 0.76 in branch segments (Figure 2b). Inter-rater agreement on image quality was very good for RCA proximal [κ = 0.81 (CI: 0.63-0.97)], RCA mid [κ = 0.81 (0.66-0.95)], and LM [κ = 0.87 (0.74-1.00)]; good for RCA distal [κ = 0.68 (0.50-0.86)] and LAD proximal [κ = 0.61 (0.34-0.88)]; moderate for LCX proximal [κ = 0.53 (0.33-0.74)]; and fair for LAD mid [κ = 0.37 (0.14-0.60)] (Figure 2c). 

Conclusion:

This study highlights the strengths of this coronary MRA pulse sequence without medications in evaluating proximal segments, where high visibility, reliable image quality scores, and strong inter-rater agreement were observed