Abstract
BACKGROUND: Pediatric cardiac magnetic resonance imaging (MRI) examinations often present significant challenges. artificial intelligence-assisted compressed sensing (ACS) technique is widely used to reduce scan time and improve image quality. Our purpose was to explore the value of the ACS technique for detecting coronary artery aneurysms (CAA) in children with Kawasaki disease (KD), as compared to the united compressed sensing (uCS) technique. METHODS: Sixty patients with KD complicated with suspected coronary artery disease (CAD) who underwent 3T cardiovascular magnetic resonance (CMR) were recruited. Contrast-enhanced magnetic resonance angiography (MRA) images were obtained by both ACS and uCS techniques. The two sequences' scan time, subjective and objective image quality, and diagnostic performance for detecting CAA were measured and compared. RESULTS: Sixty participants (39 males and 21 females; mean age ± standard deviation, 7±3 years) completed two sequences. The ACS technique exhibited a significantly shorter scan time compared to the uCS technique group (230.6±35.7 vs. 335.3±70.8 seconds, P<0.001). The subjective image quality scores, signal-to-noise ratios (SNRs), and contrast-to-noise ratios (CNRs) of the ACS technique were significantly higher than those of the uCS technique (all P<0.05). The sensitivity of the ACS technique coronary MRA was found to be 96% (53/55), the specificity was 80% (4/5), the positive predictive value (PPV) was 98% (53/54), the negative predictive value (NPV) was 67% (4/6) and the accuracy was 95% (57/60) on the basis of each person. The area under the receiver operating characteristic curve (AUC-ROC) of the ACS technique MRA images was larger than that of the uCS technique group on a per-patient basis (0.824 vs. 0.796, P<0.001), a per-vessel basis (0.894 vs. 0.823, P<0.001) and a per-segment basis (0.897 vs. 0.836, P<0.001). CONCLUSIONS: The ACS technique shows excellent diagnostic performance at 3T imaging with shorter scan time and better image quality compared with the uCS technique.