Reducing Contrast Agent Dose in Cardiovascular MR Angiography with Deep Learning

Abstract

Background Contrast‐enhanced magnetic resonance angiography (MRA) is used to assess various cardiovascular conditions. However, gadolinium‐based contrast agents (GBCAs) carry a risk of dose‐related adverse effects. Purpose To develop a deep learning method to reduce GBCA dose by 80%. Study Type Retrospective and prospective. Population A total of 1157 retrospective and 40 prospective congenital heart disease patients for training/validation and testing, respectively. Field Strength/Sequence A 1.5 T, T1‐weighted three‐dimensional (3D) gradient echo. Assessment A neural network was trained to enhance low‐dose (LD) 3D MRA using retrospective synthetic data and tested with prospective LD data. Image quality for LD (LD‐MRA), enhanced LD (ELD‐MRA), and high‐dose (HD‐MRA) was assessed in terms of signal‐to‐noise ratio (SNR), contrast‐to‐noise ratio (CNR), and a quantitative measure of edge sharpness and scored for perceptual sharpness and contrast on a 1–5 scale. Diagnostic confidence was assessed on a 1–3 scale. LD‐ and ELD‐MRA were assessed against HD‐MRA for sensitivity/specificity and agreement of vessel diameter measurements (aorta and pulmonary arteries). Statistical Tests SNR, CNR, edge sharpness, and vessel diameters were compared between LD‐, ELD‐, and HD‐MRA using one‐way repeated measures analysis of variance with post‐hoc t‐tests. Perceptual quality and diagnostic confidence were compared using Friedman's test with post‐hoc Wilcoxon signed‐rank tests. Sensitivity/specificity was compared using McNemar's test. Agreement of vessel diameters was assessed using Bland–Altman analysis. Results SNR, CNR, edge sharpness, perceptual sharpness, and perceptual contrast were lower (P < 0.05) for LD‐MRA compared to ELD‐MRA and HD‐MRA. SNR, CNR, edge sharpness, and perceptual contrast were comparable between ELD and HD‐MRA, but perceptual sharpness was significantly lower. Sensitivity/specificity was 0.824/0.921 for LD‐MRA and 0.882/0.960 for ELD‐MRA. Diagnostic confidence was 2.72, 2.85, and 2.92 for LD, ELD, and HD‐MRA, respectively (PLD‐ELD, PLD‐HD < 0.05). Vessel diameter measurements were comparable, with biases of 0.238 (LD‐MRA) and 0.278 mm (ELD‐MRA). Data Conclusion Deep learning can improve contrast in LD cardiovascular MRA. Level of Evidence Level 2 Technical Efficacy Stage 2

Editorial For “Reduction of Contrast Agent Dose in Cardiovascular MR Angiography Using Deep Learning”

“…The variant of the U‐Net and trained neural network provides excellent blood pool enhancement, vascular edge sharpness, and diameters and ensured sufficient diagnostic accuracy and confidence, which were comparable with the standard dose of MR angiogram. This study confirms that deep learning method can reduce GBCAs dose by 80%, thereby mitigate the risk of adverse Gd 3+ effects and reduce contrast media costs 6 …”

“…It also remains to be seen whether the deployment of deep learning method and low dose of GBCAs could be used in demonstrating pathologic lesions of soft tissues, such as tumors or necrosis. The study by Montalt‐Torddera et al 6 is an important step in evaluating cardiovascular diseases.…”

“…Deep learning methods have shown remarkable improvements in MR image reconstruction, image quality improvement, parameter mapping, image contrast conversion, and image segmentation 4 and have been successfully used in the brain. 5 In this issue of JMRI, an article by Montalt-Torddera et al 6 implemented deep learning method to substantially reduce the dose of GBCAs (from 0.2 mmol/kg to 0.04 mmol/kg) in MRA. The dosage of clinically approved GBCAs ranges from 0.1 mmol/kg to 0.3 mmol/kg, which translates to about 1 g of Gd 3+ per person per injection.…”

“…This study confirms that deep learning method can reduce GBCAs dose by 80%, thereby mitigate the risk of adverse Gd 3+ effects and reduce contrast media costs. 6 Interestingly, recent inaugural consensus statements developed and endorsed by the American College of Radiology and National Kidney Foundation indicated that the risk of nephrogenic systemic fibrosis from group II GBCAs in patients with advanced kidney disease is thought to be very low (zero events following 4931 administrations to patients with estimated glomerular filtration rate < 30 mL/min per 1.73 m 2 ; upper bounds of the 95% confidence intervals: 0.07% overall, 0.2% for stage 5D chronic kidney disease (CKD), 0.5% for stage 5 CKD, and no dialysis). 8 The contradiction between the studies on the toxicity of Gd3+ may be attributed to the types/doses of GBCAs.…”

“…In this issue of JMRI , an article by Montalt‐Torddera et al 6 implemented deep learning method to substantially reduce the dose of GBCAs (from 0.2 mmol/kg to 0.04 mmol/kg) in MRA. The dosage of clinically approved GBCAs ranges from 0.1 mmol/kg to 0.3 mmol/kg, which translates to about 1 g of Gd 3+ per person per injection 7 .…”