Unexplained mental retardation: is brain MRI useful?

Background: Mental retardation (MR), defined as an IQ below 70, is a frequent cause of consultation in paediatrics. Objective: To evaluate the yield of brain MRI in the diagnostic work-up of unexplained MR in children. Patients and methods: The MRI features and clinical data of 100 patients (age 1-18 years) affected with non-progressive MR of unknown origin were compared to an age-matched control group (n=100). Two radiologists conducted an independent review of the MRI scans. Results: Univariate and multivariate analyses showed a higher incidence of brain anomalies in the MR group than in the control group (53 vs 17, OR=5.7 [2.9-11.1]), for signal abnormalities within the periventricular white matter (OR=20.3 [2.6-155.3]), lateral ventricular dilatation (OR=15.6 [2.0-124]), mild corpus callosum abnormalities (shortness, atrophy) (OR=6.8 [1.8-25.6]) and subtle cerebellar abnormalities, including fissure enlargement (OR=5.2 [1.1-26.2]). The diagnostic value of MRI abnormalities was considered good in 5% of patients (Alexander disease n=1, diffuse cortical malformation n=1, leukomalacia n=1, vermian agenesis n=1, commissural agenesis n=1), and weak in 48% of patients, in whom non-specific abnormalities did not lead to a diagnosis. Some clinical features resulted in a significantly higher percentage of abnormal MRI scans: abnormal neurological examination (82% vs 47%, P=0.008), abnormal skull circumference (66% vs 49%, P=0.04). Motor delay was associated with cerebellar abnormalities (P=0.01). Conclusions: This study confirms the weak diagnostic yield of MRI in mentally retarded children. The use of a control group has enabled us to identify the neuroimaging markers frequently associated with MR. Subgrouping patients according to neuroimaging markers and clinical signs should help identify those who would benefit from molecular studies.