Chromosomal microarray or whole exome sequencing: choosing the right diagnostic path

Chromosomal microarray produced a pooled yield of approximately 19% in the same evidence base. The comparison is not marginal: in studies evaluating both methods within comparable samples, the odds of a diagnosis were higher with exome sequencing.

This does not make chromosomal microarray obsolete. The two assays interrogate different classes of genomic variation. CMA is optimized for copy number variants. WES is optimized for sequence-level changes in coding regions. The correct diagnostic path depends on the phenotype, the suspected variant type, laboratory capability, interpretation quality, cost, and the consequences of a positive result.

For pediatric patients with unexplained global developmental delay or intellectual disability, current clinical guidance has moved toward WES or whole genome sequencing as first-tier tests, either alongside or in place of CMA. That change reflects diagnostic yield, not a wholesale replacement of one technology by another.

The diagnostic hierarchy has changed

For many years, pediatric genetic testing followed a predominantly stepwise model. Chromosomal microarray was commonly placed near the beginning because it could identify clinically significant deletions and duplications across the genome without requiring a narrowly defined clinical hypothesis. Additional testing was then selected according to the phenotype.

This pathway remains clinically useful in selected cases. It is particularly relevant when the working diagnosis involves a copy number disorder, a recognizable microdeletion or microduplication syndrome, or a phenotype in which dosage imbalance is strongly suspected.

The evidence base has shifted the default position of sequencing. The 2021 American College of Medical Genetics and Genomics guideline recommended exome or genome sequencing as a first- or second-tier test for pediatric patients with congenital anomalies, global developmental delay, or intellectual disability. An updated American Academy of Pediatrics clinical report subsequently recommended WES or WGS as first-tier genetic tests for children with unexplained global developmental delay or intellectual disability.

The practical implication is a change in the initial decision point:

• CMA remains a high-value assay for genome-wide copy number analysis.
• WES provides broader access to pathogenic single-nucleotide variants and small insertions or deletions in coding genes.
• WGS extends analysis beyond the exome and may be selected when noncoding, structural, or technically difficult variants are clinically relevant, although its availability and cost vary.
• Targeted panels retain value when the phenotype maps closely to a defined set of genes or when the laboratory has strong validated coverage for a specific disorder group.

The phrase “first-tier” does not mean that one test is appropriate for every child. It means that the test can reasonably be placed at the beginning of the diagnostic pathway instead of being reserved only for cases that remain unresolved after multiple lower-yield investigations.

The central distinction is technical: CMA measures genomic dosage, while WES reads coding sequence. A higher aggregate yield does not eliminate the need to match the assay to the suspected variant class.

Diagnostic yield: quantifying the 37% versus 19% gap

Diagnostic yield is the proportion of tested patients for whom the assay identifies a result judged to explain the clinical presentation. It is not the proportion with any genetic finding. It is not the proportion with a variant of uncertain significance. It is not a guarantee for an individual patient.

The 2025 meta-analysis reported:

In same-sample comparisons, the reported odds ratio for diagnosis with WES versus CMA was 2.27, with a 95% confidence interval of 1.08–4.78. The confidence interval is broad, indicating variation across cohorts and study designs. It supports a higher diagnostic rate for WES but does not justify treating the result as universal across every developmental phenotype.

The cohort composition matters. Children with global developmental delay, intellectual disability, congenital anomalies, dysmorphic features, seizures, or multisystem involvement generally have a higher pretest probability of a monogenic or syndromic condition than children with isolated, mild, and nonspecific developmental differences. Diagnostic yield from a complex syndromic cohort cannot be transferred directly to isolated mild speech delay.

This is one of the main limitations of headline comparisons. A yield of 37% describes the pooled performance of WES across selected clinical populations. It does not mean that WES will identify the cause in 37 out of every 100 children referred for any developmental concern.

The same principle applies to the approximately 19% yield reported for CMA. The result is not a measure of poor assay performance. CMA detects a different substrate. A child whose disorder is caused by a pathogenic single-nucleotide variant may have a normal CMA. That negative result reflects the assay’s analytical scope, not necessarily the absence of a genetic cause.

What counts as a clinically useful result

A positive result has clinical utility when it changes one or more aspects of care. Potential effects include:

• establishing or refining the diagnosis;
• clarifying recurrence risk for the family;
• directing surveillance for known complications;
• changing the need for additional diagnostic procedures;
• identifying relatives who may benefit from testing;
• improving interpretation of future symptoms;
• supporting access to condition-specific services.

A molecular result without a clear phenotype correlation has lower immediate utility. A variant of uncertain significance should not be treated as a confirmed diagnosis. Its presence can add interpretive complexity without resolving the clinical question.

The relevant endpoint is therefore not simply analytical detection. It is the chain from assay signal to pathogenic classification to phenotype fit to management consequence.

Technical capabilities: CNVs versus pathogenic sequence variants

Chromosomal microarray

Chromosomal microarray measures gains and losses of genomic material. Its principal targets are copy number variants, including microdeletions and microduplications. These alterations can affect one gene, multiple genes, or regulatory regions, depending on their size and genomic location.

CMA has several operational advantages:

• genome-wide coverage for copy number changes;
• no requirement for a narrowly defined candidate gene;
• established use in pediatric developmental disorders;
• comparatively lower estimated cost than WES or WGS;
• a direct fit for suspected dosage-sensitive conditions.

Its limitations are equally important. CMA does not function as a general sequence-reading assay. It will not reliably identify the majority of single-nucleotide variants responsible for monogenic disease. It is therefore poorly matched to a phenotype caused by a coding-region variant without a copy number change.

Interpretation also depends on genomic resolution, platform design, laboratory databases, and the ability to distinguish pathogenic changes from benign population variation. A detected CNV may require parental testing or additional clinical correlation to establish significance.

Whole exome sequencing

Whole exome sequencing analyzes the protein-coding portions of genes. These regions represent a minority of the genome but contain a substantial proportion of known disease-causing variants. WES can identify single-nucleotide variants and small insertions or deletions. Some laboratory workflows also evaluate selected copy number changes from exome data, although performance is not equivalent to a dedicated microarray in every region or variant type.

WES is particularly informative when:

• the phenotype is clinically heterogeneous;
• several genetic syndromes remain plausible;
• the child has a multisystem presentation;
• congenital anomalies and developmental delay occur together;
• a monogenic disorder is suspected without a single obvious candidate gene;
• previous cytogenetic testing was nondiagnostic.

The assay is not complete genome analysis. Coverage is uneven across exons. Some coding regions are difficult to sequence or align. Variants in noncoding regions, deep intronic regions, repeat expansions, certain structural variants, and some mosaic changes may be missed or require separate methods. A negative WES result therefore reduces, but does not eliminate, the probability of a genetic diagnosis.

Trio sequencing, in which the child and both biological parents are analyzed, can improve interpretation by showing whether a variant is de novo, inherited, or absent in the parents. The supplied evidence supports WES as a high-yield first-tier option, but the exact performance of trio versus singleton testing depends on laboratory protocol and case selection. Those distinctions should be documented in the laboratory report rather than assumed from the test name alone.

Whole genome sequencing

Whole genome sequencing analyzes a broader genomic territory than WES. It may offer improved access to noncoding regions, structural variation, and some variant classes that are incompletely represented by exome testing. The clinical value is greatest when the suspected mechanism extends beyond conventional coding variants or when prior testing has been negative.

The trade-off is not simply price. WGS generates a larger interpretive space, requires validated pipelines, and may produce findings whose clinical relevance is difficult to establish. Estimated costs range from approximately $1,000 to more than $5,000, depending on the laboratory and depth of analysis. Coverage, reporting policy, variant classification, and data interpretation remain decisive components of performance.

Targeted panels

Targeted sequencing panels remain appropriate when the clinical phenotype is sufficiently specific. A panel can provide strong analytical coverage of selected genes and may reduce the volume of incidental or unrelated findings. It is less useful when the phenotype is broad, atypical, or not yet clinically resolved.

Panels also age quickly. New disease-gene associations and revised genotype–phenotype relationships can change the adequacy of a previously designed panel. A negative panel result does not necessarily exclude a genetic disorder outside the selected gene list.

Choosing between CMA and WES in practice

The comparison between chromosomal microarray and whole exome sequencing for pediatric developmental delay should be based on the suspected biology rather than on a universal sequence of tests.

A practical decision structure is:

1. Define the phenotype at syndromic resolution.

Global developmental delay, intellectual disability, congenital anomalies, seizures, growth abnormalities, dysmorphic features, and organ-system involvement should be recorded as a combined phenotype. Isolated symptoms produce a less specific diagnostic signal.

2. Estimate the likelihood of a copy number mechanism.

A recognizable microdeletion or microduplication pattern, a strong dosage-sensitive syndrome hypothesis, or a prior clinical indication for genome-wide CNV analysis supports CMA.

3. Assess the likelihood of a monogenic disorder.

Multisystem disease, a strong family history, consanguinity, recurrent unexplained findings, or a phenotype compatible with many single-gene disorders supports WES or WGS.

4. Review prior testing instead of repeating it by label.

A previous “genetic test” may have been a limited panel, a low-resolution assay, or a test with restricted variant interpretation. The exact platform, coverage, variant classes, and reporting criteria determine what has actually been excluded.

5. Select a test that can change management.

The result should be capable of affecting diagnosis, surveillance, family counseling, treatment selection, or the need for further investigations.

6. Plan the interpretation pathway before sequencing.

The clinical team should define how variants of uncertain significance, secondary findings, parental testing, and reanalysis will be handled. Sequencing without an interpretation plan increases report volume without guaranteeing clinical clarity.

For many children with unexplained global developmental delay or intellectual disability, WES or WGS is now reasonable at the beginning of testing. CMA may be ordered concurrently, used selectively, or retained for cases in which copy number pathology is a primary concern. The optimal arrangement varies by laboratory access, insurance policy, phenotype, and local clinical practice.

Cost, turnaround, and the economics of first-line testing

Estimated clinical or out-of-pocket costs provide a broad range rather than a fixed price:

• CMA: approximately $300–$1,000;
• WES: approximately $500–$2,500;
• WGS: approximately $1,000–$5,000 or more.

These figures are dependent on the laboratory, test design, depth of analysis, trio status, interpretation package, and regional reimbursement structure. They should not be presented as universal patient charges. Coverage criteria also vary among private insurers and healthcare systems.

The lowest unit price is not necessarily the lowest pathway cost. A lower-yield assay followed by sequential testing can create additional blood draws, referrals, interpretation events, and delays. Conversely, a broader test can produce unresolved findings that require parental studies or specialist review.

Clinical utility therefore has at least three components:

Turnaround time is not specified in the available evidence and should not be inferred from the assay category. A laboratory with high sequencing throughput may report WES efficiently, while another may have a longer workflow due to confirmatory testing or multidisciplinary interpretation. The operational metric is the laboratory’s documented turnaround for the exact test configuration.

A useful referral process records:

• the precise assay name;
• whether analysis is singleton, duo, or trio;
• the genomic regions and variant classes covered;
• the policy for secondary findings;
• whether CNV calling is included in WES;
• the conditions for parental or orthogonal confirmation;
• the availability of periodic data reanalysis.

These details determine the real scope of the test more reliably than the label “genetic sequencing.”

Reanalysis and the unresolved diagnostic pathway

A negative initial WES result is not equivalent to a completed genetic evaluation. Variant classification changes as disease-gene relationships expand, phenotype descriptions improve, and previously uncertain variants accumulate evidence. The available evidence indicates that periodic reanalysis of exome data can increase diagnostic yield by approximately 10% to 16%.

This gain does not arise from changing the original sequence data. It arises from applying updated knowledge and improved interpretation to data that were already generated. Reanalysis may reconsider:

• variants previously classified as uncertain;
• genes newly associated with developmental disorders;
• inheritance patterns clarified by parental testing;
• phenotype terms that were incomplete in the original referral;
• technical findings that become relevant after a revised diagnosis;
• candidate variants that were not reportable under earlier laboratory criteria.

Reanalysis should be treated as a defined clinical service, not as an automatic assumption. Laboratories differ in whether they offer scheduled reinterpretation, clinician-requested review, or analysis only after a phenotype update. The interval and trigger for reanalysis should be documented.

CMA data can also be revisited when classification databases and phenotype interpretation change, although its principal value remains the identification and interpretation of copy number variation. A negative CMA does not become a negative exome by passage of time.

A nondiagnostic result is an assay-specific statement. It means that the tested method did not identify a reportable cause within its validated scope. It does not mean that the child has no genetic disorder.

Clinical utility beyond the diagnostic percentage

The 37% versus 19% comparison is useful for selecting a first-line strategy, but yield alone is insufficient. A test with a higher yield can still have limited clinical value if results are poorly classified, weakly correlated with the phenotype, or disconnected from care.

The strongest pathway combines analytical performance with disciplined interpretation:

• phenotype documentation should be specific enough to guide variant review;
• the laboratory should state the assay’s coverage and limitations;
• positive findings should be evaluated for pathogenicity and phenotype concordance;
• variants of uncertain significance should remain separate from confirmed diagnoses;
• parental testing should be used when it materially clarifies inheritance;
• negative results should be interpreted according to the assay’s blind spots;
• reanalysis should remain available when the initial result is nondiagnostic.

This framework also prevents an unhelpful binary conclusion that WES has replaced CMA. WES has a higher pooled diagnostic yield in the cited GDD/ID evidence and is now recommended as a first-tier option by major professional organizations. CMA remains analytically valuable for CNVs and remains relevant when the suspected disease mechanism is genomic dosage imbalance.

The choice is not between an old test and a new test. It is between variant detection strategies.

Rigid assessment of the two pathways

For unexplained global developmental delay or intellectual disability, WES generally provides the stronger first-line diagnostic yield. The pooled estimate of approximately 37% versus approximately 19% for CMA supports that position. The 2021 ACMG guideline and updated AAP guidance are consistent with this shift.

CMA remains appropriate when copy number variation is a central diagnostic hypothesis, when the clinical presentation is compatible with a dosage disorder, or when laboratory and pathway constraints favor genome-wide CNV analysis. A normal CMA does not exclude a sequence-level disorder. A negative WES does not exclude a CNV, noncoding variant, structural variant, or poorly covered coding-region alteration.

The technically defensible approach is phenotype-led:

• use CMA for a primary CNV question;
• use WES for broad or heterogeneous monogenic disease assessment;
• consider WGS when the suspected mechanism extends beyond exome space or prior testing remains nondiagnostic;
• use targeted panels when the phenotype maps convincingly to a defined gene set;
• preserve and periodically reassess negative sequencing data.

For pediatric developmental delay, WES has become the more efficient general diagnostic entry point in many cases. CMA retains clinical utility because its analytical target is different. The correct test is determined by the variant class that must be found, not by the popularity of the platform.