Reference ranges for fetal brain structures using magnetic resonance imaging: systematic review

Objectives: To evaluate the methodology of studies reporting reference charts of fetal brain structures at magnetic resonance imaging (MRI). Methods: MEDLINE, EMBASE, CINAHL and the Web of Science databases were searched electronically up to December 31, 2020. The primary outcome was to evaluate the methodology of the studies investigating the biometry and growth of the fetal brain. A list of 26 quality criteria divided into two domains according to either "study design", "reporting and statistical methods" and "specific relevant MRI aspects" was developed to evaluate the methodological appropriateness of the included studies. The overall quality score was defined as the sum of risk of bias marks, with the range of possible scores being 0-26, and then expressed as a percentage (the lowest the percentage, the highest the risk of bias). This quality assessment was applied to each individual study reporting reference ranges of fetal brain structures at MRI. Results: Fifteen studies were included. The overall mean quality score of the studies evaluated in this review was 48.7%. When focusing on each domain, the mean quality score was 42% for "study design", 59.4% for "statistical and reporting methods" and 33.3% for "specific relevant MRI aspects". For the "study design" domain, the sample size calculation and the consecutive enrollment of women were the items found at the highest risk of bias. For the "statistical and reporting methods" domain, the presence of regression equations for mean and SD for each measurement, the number of measurements taken for each variable and the presence of postnatal assessment information were the items found at the highest risk of bias. For the "specific relevant MRI aspects" domain, a whole fetal brain assessment was performed in none of the included studies and therefore was considered as the item at the highest risk of bias. Conclusions: Most of the previously published charts evaluating fetal brain charts at MRI show a high heterogeneity and a low to moderate quality in terms of methodology, as already reported for ultrasound. This article is protected by copyright. All rights reserved.