A Twin Study of Autism Spectrum Disorder and MRI-Brain Incidental Findings

Background: Recent anatomic MRI-brain studies suggest the prevalence of asymptomatic “incidental” findings in autism spectrum disorder (ASD) is similar to that of typically developing controls. However, given the causes of ASD likely combine genetic and environmental factors, a study design using twins with ASD would facilitate controlling for genetic and environmental biases. To date, such a twin model has not been used to study incidental MRI-brain findings in ASD.

Objectives: To use multi-modal high-resolution 3T MRI to assess the prevalence of incidental findings in monozygotic (MZ) and dizygotic (DZ) twins with ASD, in comparison to typically developing control twins.  

Methods: MRI-brain scans of same-sex, ASD twins and typically developing twins between the ages of 6 and 15 years were obtained using a GE 3T scanner employing multiplanar T1 and T2 sequences. Incidental findings consisted of anatomic variants agreed upon by two neuroradiologists blinded to diagnosis.

Results: MRI data were available from 16 twin pairs (8 MZ; 8DZ) where both have ASD, 12 twin pairs (3 MZ; 9 DZ) where only one twin has ASD, and 17 control twin pairs (6 MZ; 11 DZ) where neither twin has ASD. In total there were 44 subjects with ASD, 12 typically developing siblings of ASD twins, and 34 controls. No significant difference was found between prevalence rates in subjects with ASD, typically developing siblings of ASD twins, and typically developing control twins (73%, 83%, and 65%, respectively, chi-square_­­= 1.61, p=0.4). McNemar’s test for correlated proportions of twins with ASD and their typically developing twin siblings showed no association between ASD and incidental MRI-brain findings (OR 2.0, 95% CI: 0.18-22.1, p=0.5). 

Conclusions: These preliminary findings are consistent with the available literature suggesting the absence of increased incidental findings associated with ASD.  Studies with large sample size would have the statistical power to detect smaller effect sizes while appropriately controlling for genetic and environmental biases. Additionally, the examination of the specific type of incidental finding would also be interesting, but larger number of participants will be needed.