MRI findings in 77 children with non-syndromic autistic disorder

Abstract

Background: The clinical relevance of MR scanning in children with autism is still an open question and must be considered in light of the evolution of this technology. MRI was judged to be of insufficient value to be included in the standard clinical evaluation of autism according to the guidelines of the American Academy of Neurology and Child Neurology Society in 2000. However, this statement was based on results obtained from small samples of patients and, more importantly, included mostly insufficient MRI sequences. Our main objective was to evaluate the prevalence of brain abnormalities in a large group of children with a non-syndromic autistic disorder (AD) using T1, T2 and FLAIR MRI sequences.

Methodology: MRI inspection of 77 children and adolescents with non-syndromic AD (mean age 7.4+/-3.6) was performed. All met the DSM-IV and ADI -R criteria for autism. Based on recommended clinical and biological screenings, we excluded patients with infectious, metabolic or genetic diseases, seizures or any other neurological symptoms. Identical MRI inspections of 77 children (mean age 7.0+/-4.2) without AD, developmental or neurological disorders were also performed. All MRIs were acquired with a 1.5-T Signa GE (3-D T1-FSPGR, T2, FLAIR coronal and axial sequences). Two neuroradiologists independently inspected cortical and sub-cortical regions. MRIs were reported to be normal, abnormal or uninterpretable.

Principal findings: MRIs were judged as uninterpretable in 10% (8/77) of the cases. In 48% of the children (33/69 patients), abnormalities were reported. Three predominant abnormalities were observed, including white matter signal abnormalities (19/69), major dilated Virchow-Robin spaces (12/69) and temporal lobe abnormalities (20/69). In all, 52% of the MRIs were interpreted as normal (36/69 patients).

Conclusions: An unexpectedly high rate of MRI abnormalities was found in the first large series of clinical MRI investigations in non-syndromic autism. These results could contribute to further etiopathogenetic research into autism.

Figure 1. White matter abnormalities in autism.

Two children illustrating the principal categories of white matter signal abnormalities. Figure 1A. Punctate T2 Hyperintensity: Abnormal findings were placed in this category when small (<2 mm) rounded abnormalities were found scattered bilaterally in the white matter (white arrow). They were asymmetric and homogeneous, and no findings suggest that necrosis was present. They were very intense compared with adjacent white matter on T2 and FLAIR sequences, and did not involve the basal ganglia, the periventricular white matter fibers or the sub-cortical U fibers. These abnormalities were generally found in association with other supratentorial abnormalities. Figure 1B. Posterior T2 Hyperintensity. Abnormalities placed in this category were “plaque-like areas” of mild white matter hyperintensity relatively symmetrical bilaterally at the posterior horns of the lateral ventricles (black arrow). There was no deformation of the lateral ventricular contour adjacent to these lesions. No abnormality of the sub-cortical U fibers was observed.

Figure 2. Dilated Virchow-Robin spaces in autism.

Virchow-Robin (VR) spaces are fluid-containing dilatations of the perivascular space that surrounds penetrating arteries in the brain (white arrow). We defined abnormal VR spaces when the spaces were >3 mm using the classification system developed by Heier et al .

Figure 3. Temporal pole abnormalities in a child with autism.

Example of a typical sub-cortical hyperintensities on T2-weighted coronal images localized in the temporal poles observed in children with autism (red arrows) and a normal image of a control child without autism.