Contrast-free postoperative imaging of the pulmonary arteries: Intraindividual comparison of relaxation-enhanced angiography without contrast and triggering and time-resolved contrast-enhanced magnetic resonance angiography.

BACKGROUND: Flow-independent, three-dimensional isotropic non-contrast magnetic resonance angiography (MRA) such as relaxation enhanced angiography without contrast and triggering (REACT) are increasingly established as alternatives to contrast-enhanced MRA (CE-MRA) for assessment of the pulmonary arteries (PAs). However, their performance for dedicated postoperative assessment of the PAs in patients with congenital heart disease (CHD), particularly in the presence of different implant types, remains unclear.

METHODS: In this retrospective single-center study, 53 patients with CHD underwent clinically indicated cardiovascular magnetic resonance, including both four-dimensional (4D) CE-MRA and REACT at 1.5T. Three radiologists independently scored image quality (IQ) as well as motion and susceptibility artifacts on 5-point Likert scales and measured the diameters of the PAs [main (MPA), left (LPA), and right pulmonary artery (RPA)]. Subgroup analysis was performed for stents, conduit/patch/valve (CPV), and no implant.

RESULTS: Pooled across readers and PA segments, REACT achieved higher overall IQ than 4D CE-MRA (median 3.67 [3.00-4.17] vs 3.00 [2.33-3.33]; p < 0.001) and provided significantly better motion scores (p < 0.001), whereas susceptibility scores were comparable between techniques. The proportion of fully diagnostic studies (3/3 segments) was similar (REACT 77.4%, 41/53; 4D CE-MRA 83.0%, 44/53; McNemar, p = 0.38). Diameter measurements showed excellent inter-reader agreement (intraclass correlation coefficient ≈ 0.89-0.95) and minimal bias between techniques; only the RPA yielded slightly smaller diameters in REACT (mean difference -0.85 ± 1.51 mm, p < 0.001). In subgroup analysis, stented segments showed no IQ advantage of REACT (p > 0.99) with IQ being limited due to susceptibility artifacts in both 4D CE-MRA and REACT. In the CPV and the no-implant group, REACT yielded a one point higher median IQ score (both p = 0.002) and one point less impaired by motion artifacts (CPV: p < 0.001; no implant: p = 0.002), while both techniques provided very high shares of diagnostic IQ (defined as IQ ≥ 2; both > 90%; p > 0.99).

CONCLUSION: REACT enables robust, contrast-free postoperative imaging of the PAs in patients with CHD with superior IQ and reduced motion artifacts compared to 4D CE-MRA, while maintaining highly reproducible diameter measurements. Stented segments remain a shared limitation.