High Diagnostic YIELD and Low Therapeutic IMPACT of Array-CGH in the Clinical Management of Autistic Patients.

Background:  Array-CGH has become a first-tier clinical diagnostic test in the medical management of children with autism spectrum disorder (ASD). However, the full medical implications of this approach beyond diagnosis have not been explored.

Objectives:  (1) To propose specific criteria for the definition of a “positive” or “negative” array-CGH outcome; (2) Based on these criteria, to define the diagnostic yield of array-CGH in a sample of Italian ASD patients; (3) To define the follow-up diagnostic and therapeutic yield, whereby further positive diagnostic assessments and therapeutic interventions were driven by array-CGH results.

Methods:  Array-CGH was performed using the Human Genome CGH SurePrint G3 Microarray 4x180K Kit (Agilent Technologies). CNVs were classified into “rare” or “common” using the last release of Database of Genomic Variants (DGV) setting the threshold for “rare” at ≤4 gains or losses in DGV. All genes spanned in rare CNVs were sought in the autism candidate gene lists present on the Simons Foundation web site (https://id.sfari.org/) and the AutismKB database (http://autismkb.cbi.pku.edu.cn/).

Results:  Array-CGH outcomes were blindly classified by four authors (AMP, CL, SG, CP) into five categories, based on a stringent set of criteria: 1-“Positive - certainly causal”, 2-“Positive - probably causal or with relevant functional consequences”, 3-“Uncertain causality”, 4-“Negative - common variant with modulatory effects”, 5-“Negative - No functional role”. Array-CGH results were obtained for 182 families (159 simplex and 23 multiplex), including 581 individuals. A positive yield was obtained from 21 (10.0%) and 39 (18.8%) of 209 ASD patients for outcome cat. n. 1 and 2, respectively. Based on array-CGH results, further medical testing was undertaken in 43/209 (20.1%) ASD patients and 10/372 (2.7%) first-degree relatives. These exams include: brain MRI (with/without 31P spectroscopy); urinary and blood levels of Mg++ or proline; EKG; cardiac or abdominal sonogram; consultations with cardiology, immunology, pneumology, ophtalmology; auditory evoked potentials; capillary fragility testing; blood liver parameters and lipids; blood amino acids; glucose tolerance test; IQ and memory (in family members). These tests turned out positive in 8/209 (3.8%) ASD patients and in 3/372 (0.8%) first-degree relatives. Pharmacological or supplement therapies driven by array-CGH results were prescribed in 2/209 (1.0%) of ASD patients and clinical improvement was recorded in 1/209 (0.5%) case.

Conclusions:  The high diagnostic yield obtained by array-CGH in our sample confirms the usefulness of this test in the autism clinic. Array-CGH are also able to successfully guide further medical testing in a small percentage of ASD cases, who would not have been further assessed otherwise. To this date, the therapeutic impact of positive array-CGH results remains elusive, due to the lack of ASD-specific drugs. Nonetheless, array-CGH appears a strong candidate to contribute, in conjunction with whole-exome sequencing and with other exams and specific biomarkers, to the molecular characterization of ASD patients in view of future targeted personalized pharmacological therapies.