Abstract / Summary
Background: Autism spectrum disorder (ASD) is clinically and etiologically heterogeneous. Chromosomal microarray analysis (CMA) detects copy-number variants (CNVs), but diagnostic yield varies markedly with phenotype.
Objective: To assess CMA diagnostic yield in ASD, identify clinical features associated with higher yield, review recurrent clinically relevant CNVs, and distinguish diagnostic yield from downstream clinical utility.
Methods: We conducted a PRISMA 2020-informed focused update. A contemporary four-database systematic review (January 2021-June 26, 2026; 434 records, 13 studies, 11,535 children) served as the evidence backbone, followed by a targeted update through October 4, 2026, and inclusion of landmark earlier studies for historical and technical context.
Results: The contemporary review reported lower P/LP CNV yield in ASD (3.3%-15.2%) than in DD/ID phenotypes (17.1%-33.6%). Two 2026 ASD cohorts reinforced phenotype dependence: 17.1% in complex versus 4.7% in essential ASD, and 2.8% in a rigorously selected nonsyndromic cohort. In Tammimies et al., CMA yield was 24.5% in morphologically complex ASD versus 4.2% in essential ASD. Retrospective series support clinically actionable downstream evaluation, although prospective management-change evidence remains limited.
Conclusions: CMA remains useful for detecting CNVs and genomic syndromes, especially when ASD co-occurs with developmental/intellectual impairment, dysmorphism, congenital anomalies, epilepsy, or growth abnormalities. A negative CMA does not exclude a genetic etiology; ES/GS and targeted testing are complementary. Phenotypic complexity appears more informative for pre-test yield than an ASD diagnosis considered in isolation.