Point estimate and reference normality interval of MRI-derived myocardial extracellular volume in healthy subjects: a systematic review and meta-analysis

• Published in: European radiology 2019-12, Vol.29 (12), p.6620-6633
• Main Authors: Sardanelli, Francesco, Schiaffino, Simone, Zanardo, Moreno, Secchi, Francesco, Cannaò, Paola Maria, Ambrogi, Federico, Di Leo, Giovanni
• Format: Article
• Language: English
• Subjects: Age; Bias; Body mass; Body mass index; Body size; Confidence intervals; Diagnostic Radiology; Extracellular Space - diagnostic imaging; Field strength; Heart; Heterogeneity; Humans; Imaging; Internal Medicine; Interventional Radiology; Magnetic fields; Magnetic Resonance; Magnetic resonance imaging; Magnetic Resonance Imaging - methods; Males; Medicine; Medicine & Public Health; Meta-analysis; Myocardium - cytology; Neuroradiology; Normality; Predictive Value of Tests; Quality assessment; Radiology; Reference Values; Regression analysis; Reproducibility of Results; Statistical analysis; Subgroups; Systematic review; Ultrasound; Ventricle
• ISSN: 0938-7994; 1432-1084
• DOI: 10.1007/s00330-019-06185-w

Objectives To estimate the MRI-derived myocardial extracellular volume (ECV) in healthy subjects together with reference normality interval. Methods The study was registered on PROSPERO and reported according to PRISMA. In October 2017, a systematic search (MEDLINE/EMBASE) was performed for articles reporting MRI-derived ECV in healthy subjects. The pooled ECV (pECV) with 95% confidence interval (CI) was calculated using the random-effect model; the normality interval was calculated as pECV ± 2 root mean square of all study standard deviations. The Newcastle-Ottawa scale was used for assessing study quality, subgroup/meta-regression analyses for technical/biological covariates, and Egger test for publication bias risk. Results Of 282 articles, 56 were analyzed totaling 1851 subjects with age 16–68 years, body mass index 23–28 kg/m 2 , and left ventricular ejection fraction 58–74%. Contrast dose varied from 0.075 to 0.200 mmol/kg. Heterogeneity was high ( I 2  = 92%). The pECV was 25.6% (95% CI 25.2–26.0%) with a normality interval of 19.6–31.6%. pECV was slightly increasing with age ( β  = 0.03%, p  = 0.038) and slightly decreasing with the percentage of males ( β  = − 0.02%, p  = 0.053). Sequence type significantly ( p  = 0.003) impacted on pECV: the normal interval was 19.9–31.9% for MOLLI and 20.3–33.5% for ShMOLLI. Contrast type/dose, time of acquisition, and magnetic field strength did not significantly impact pECV ( p  > 0.093). Quality was moderate or high in 48/56 studies (86%). No risk of publication bias ( p  = 0.728). Conclusions Myocardial pECV in healthy subjects was 25.6%, increasing by 0.03% for each year of age. The ECV normality interval was 19.9–31.9% for MOLLI and 20.3–33.5% for ShMOLLI. Key Points • The pooled estimate of normal MRI-derived ECV based on 1851 subjects was 25.6%, slightly increasing with age and slightly decreasing with the percentage of males. • MRI-derived ECV was independent of contrast type/dose and field strength but dependent on the imaging sequence. • The modeled normality reference interval of MRI-derived ECV was 19.9–31.9% for the MOLLI sequence and 20.3–33.5% for the ShMOLLI sequence.