Prognostic value of diffusion-weighted magnetic resonance imaging of brain in fetal growth restriction: results of prospective multicenter study

Abstract

Objective: To measure prospectively apparent diffusion coefficient (ADC) values between 28 and 32 weeks of gestation in different cerebral territories of fetuses with estimated fetal weight (EFW) ≤ 5th centile, and analyze their association with adverse perinatal outcome. Methods: This was a prospective study involving six tertiary-level perinatal centers. In the period 22 November 2016 to 11 September 2017, we included singleton, small-for-gestational-age (SGA) fetuses with EFW ≤ 5th percentile, between 28 and 32 weeks of gestation, regardless of the umbilical artery Doppler and maternal uterine artery Doppler findings. A fetal magnetic resonance imaging (MRI) examination with diffusion-weighted sequences (DWI) was performed within 14 days following inclusion and before 32 weeks. ADC values were calculated in the frontal and occipital white matter, basal ganglia and cerebellar hemispheres. An ultrasound examination was performed within 1 week prior to the MRI examination. The primary outcome was a composite measure of adverse perinatal outcome, defined as any of the following: perinatal death; admission to neonatal intensive care unit with mechanical ventilation > 48 h; necrotizing enterocolitis; Grade III–IV intraventricular hemorrhage; periventricular leukomalacia. A univariate comparison of median ADC values in all cerebral territories between fetuses with and those without adverse perinatal outcome was performed. The association between ADC values and adverse perinatal outcome was then analyzed using multilevel logistic regression models to adjust for other common prognostic factors for growth-restricted fetuses. Results: MRI was performed in 64 patients, of whom five were excluded owing to fetal movement artifacts on DWI and two were excluded for termination of pregnancy with no link to fetal growth restriction (FGR). One intrauterine death occurred secondary to severe FGR. Among the 56 liveborn neonates, delivered at a mean ± SD gestational age of 33.6 ± 3.0 weeks, with a mean birth weight of 1441 ± 566 g, four neonatal deaths occurred. In addition, two neonates required prolonged mechanical ventilation, one of whom also developed necrotizing enterocolitis. Overall, therefore, seven out of 57 (12.3%) cases had an adverse perinatal outcome (95% CI, 3.8–20.8%). The ADC values in the frontal region were significantly lower in the group with adverse perinatal outcome vs those in the group with favorable outcome (mean values of both hemispheres, 1.68 vs 1.78 × 10–3 mm2/s; P = 0.04). No significant difference in ADC values was observed between the two groups in any other cerebral territory. A cut-off value of 1.70 × 10–3 mm2/s was associated with a sensitivity of 57% (95% CI, 18–90%), a specificity of 78% (95% CI, 63–88%), a positive predictive value of 27% (95% CI, 8–55%) and a negative predictive value of 93% (95% CI, 80–98%) for the prediction of adverse perinatal outcome. A mean frontal ADC value < 1.70 × 10–3 mm2/s was not associated significantly with an increased risk of adverse perinatal outcome, either in the univariate analysis (P = 0.07), or when adjusting for gestational age at MRI and fetal sex (odds ratio (OR), 6.06 (95% CI, 0.9–37.1), P = 0.051) or for umbilical artery Doppler (OR, 6.08 (95% CI, 0.89–41.44)). Conclusion: This first prospective, multicenter, cohort study using DWI in the setting of SGA found lower ADC values in the frontal white-matter territory in fetuses with, compared with those without, adverse perinatal outcome. To determine the prognostic value of these changes, further standardized evaluation of the neurodevelopment of children born with growth restriction is required. Copyright © 2019 ISUOG. Published by John Wiley & Sons Ltd.