Nomograms for Cardiovascular Magnetic Resonance Measurements in the Pediatric Age Group: To Define the Normal and the Expected Abnormal Values in Corrected/Palliated Congenital Heart Disease: A Systematic Review

Our purpose is to provide an overview and to systematically review the strengths and limitations of studies on pediatric and adolescent normal values for cardiovascular MRI parameters. A literature search was performed within the National Library of Medicine using the following keywords: normal, reference values, cardiovascular magnetic resonance imaging, and children/pediatric. Eleven published studies evaluating cardiovascular MRI measurements in normal children were included in the present analysis. Our results revealed reasonable consistencies in the protocols employed for cardiovascular MRI. Inter- and intraobserver variability analyses were performed in most studies and generally showed acceptable reproducibility. However, several numerical and methodological limitations emerged. Besides small sample sizes (the largest study enrolled 114 subjects), data for some structures (pulmonary arteries, aortic arch) were limited, and neonates/infants were poorly represented (eg, only two studies). There was heterogeneity regarding measurement normalization (eg, for gender, age, or both), and data were mostly expressed as mean values, while z-scores (commonly used in pediatric echocardiography) were rarely employed. Theoretically, a z-score or a standard deviation of ±2 is considered pathological. Furthermore, differences among races and ethnic groups were not evaluated. In conclusion, our analyses revealed an important need for generation of pediatric and adolescent cardiovascular MRI nomograms built over a wide population of healthy children, using consistent methodologies and with consideration of potentially relevant confounders. More data on expected abnormal values in specific CHD populations (eg, univentricular hearts) also need to be defined. Level of Evidence: 2. Technical Efficacy Stage: 3. J. Magn. Reson. Imaging 2019;49:1222–1235.