The American College of Medical Genetics this month published recommendations for how best to design and interpret chromosomal microarrays used in clinical labs.
The recommendations are presented in a pair of papers in the current issue of Genetics in Medicine, the non-profit's official journal, and outline standards for reporting copy-number variation in the postnatal setting.
Both papers were authored by a working committee of the organization. The first, directed toward manufacturers, concerns design and performance characteristics of arrays designed to detect constitutional abnormalities; while the second aims to help clinical labs evaluate copy-number variants and to "promote consistency" in their ability to interpret and report genomic microarray results.
The papers follow ACMG's September 2010 recommendation that chromosomal microarrays should be used as the first-tier test for detecting postnatal constitutional genetic abnormalities (BAN 9/28/2010).
They also fall in line with the desire by the medical genetics community to create guidelines for the clinical use of cytogenetic arrays, and to set standards for interpreting CNVs in light of the US Food and Drug Administration's potential regulation of such arrays (BAN 7/6/2010).
In particular, both papers argue that interpreting CNVs detected by arrays should be part of the practice of medicine, and that such arrays should be assessed based on their ability to detect CNVs.
Kathleen Rao, director of the clinical cytogenetics laboratory at UNC Hospitals in Chapel Hill, NC, and a coauthor of the first paper, told BioArray News this week that the recommendations were drawn up by a committee appointed by the ACMG board following a June 2010 meeting attended by representatives of the FDA, array manufacturers, clinical laboratory directors, the ACMG, and consumers.
She said that drafts of the document were sent out to "experts in the clinical application of whole genome arrays, array manufacturers, and to the entire ACMG membership for comment before submission of the document for approval," and that the authors "incorporated many of the suggestions submitted."
According to Rao, the recommendations were drawn up "primarily to inform the FDA and manufacturers of what clinical laboratory directors want and need in the design of a whole genome array that will be used for clinical testing of postnatal samples for constitutional abnormalities."
She noted that the majority of clinical labs are purchasing arrays from manufacturers and validating them in their laboratories, and that the FDA "has indicated that they are moving towards requiring that these arrays go through an FDA approval process before they can be applied to clinical samples." Given clinical labs' familiarity with the technology, the ACMG "wanted to share the recommendations of this very experienced group of laboratory directors with both the FDA and the manufacturers," Rao said.
While labs and manufacturers are not obligated to follow the recommendations, Rao said that ACMG guidelines are "often used as a standard for accreditation agencies when clinical laboratories are inspected" under the Clinical Laboratory Improvement Amendments, and that "many of the ACMG laboratory guidelines" have become part of the College of American Pathologists' inspection criteria for genetic testing laboratories. The FDA also regulates reagents and devices used in all clinical laboratories, she added.
It is unclear to what extent vendors will consider or abide by the recommendations when making next-generation chromosomal arrays for geneticists. Rao said that some array manufacturers were present at the June 2010 meeting and contributed to the discussion at that time. In addition, drafts of the recommendations were sent out to several of the known manufacturers to review and comment on at the same time that it was sent to the ACMG membership.
At least one firm, Affymetrix, is already marketing a new product, called CytoScan HD, which it claims was developed with the new guidelines in mind (BAN 5/10/2011).
Affy, Agilent Technologies, Illumina, and Roche NimbleGen did not comment in time for this publication.
Rao said that her hope as a lab director is that the guidance will "result in the availability of FDA-approved devices …for clinical use that meet the needs of the clinical genetics community for the accurate, reliable, high resolution detection of postnatal constitutional abnormalities in our patients."
[ pagebreak ]
The First Paper
According to the first paper, it is "not feasible" to use the same method to validate microarray technology that targets the entire genome as an assay that targets a specific gene or syndromic region. This "new paradigm" requires a different approach to validating and regulating the technology.
ACMG suggests that arrays "be evaluated and manufacturers regulated for the ability to accurately measure copy number gains or losses in DNA. Afterwards, medical professionals with "appropriate training and certification" should interpret the findings and assign clinical significance.
The paper breaks down the guidelines into 11 recommendations that include overseeing array manufacturing and use, chip design, probe selection, and genomic coverage. It also describes the best way for vendors to validate their arrays and for end users to analyze and interpret their findings.
The first six of the 11 recommendations concern array design. According to ACMG, probes should be placed throughout the genome at regular intervals to enable the array to detect CNVs at least 400 kilobases in length.
This span will allow for a "broad genomic screen for novel imbalances and reliably detect all currently described syndromic microdeletions and microduplications mediated by segmental duplication architecture," according to the paper.
It adds that platform choice will influence the number of probes necessary to achieve this resolution. The paper also stressed that its purpose is not to limit the resolution of clinical genomic microarrays but to encourage manufacturers to build arrays that can exceed the minimum detection size of 400 kb "when feasible."
Chromosomal arrays should contain probes targeting dose-sensitive genes known to cause phenotypes consistent with common indications for a postnatal genomic screen, the paper states. These include general intellectual disability, developmental delays, autism, and congenital anomalies.
Dose-sensitive genes include those for which a deviation from the normal copy number has clinical implications for the patient.
The paper also cautions that "regardless of probe enrichment, no microarray platform will detect all mutations associated with a given syndrome."
Because of this, array manufacturers and laboratories must "clearly state that failure to detect a copy number alteration at any locus does not exclude the diagnosis of any of the disorders targeted on the microarray."
If the microarray platform is meant to replace alternative clinically validated technologies to detect gains or losses of particular regions of the genome, the manufacturer and performing lab should ensure appropriate probe coverage in those genomic regions, the paper recommends.
Additionally, it states that including SNP probes on an array is "desirable" and will make such chips more appealing to cytogeneticists.
Unlike with CNVs, labs can use SNP probes to identify copy-neutral loss of heterozygosity and uniparental disomy in samples. This fact has prompted companies like Agilent Technologies and Roche NimbleGen (BAN 3/29/2011), which have traditionally sold SNP-absent comparative genomic-hybridization arrays, to add SNP content to newer generations of their chips.
The first paper also states that the ability to analyze SNPs is not currently "a requirement" for arrays. Rather, the "primary goal of these analyses is to reliably detect copy number alterations."
Among other facets of array design, ACMG also recommends that labs should avoid probes that target repetitive sequences or "show spurious calls as gains or loss that do not represent true copy number variation in the sample."
Before a design is finalized, manufacturers should also challenge the array with a comprehensive set of abnormal samples that, in combination, survey all regions of the genome represented on the microarray.
Probes that fail to demonstrate appropriate and reliable copy-number response should be excluded or permanently suppressed from analysis, ACMG said.
With manufacturers still in mind, the organization made two recommendations specifically for them. The paper notes that manufacturers should detail an assay's performance, its quality parameters, and the algorithm on which it relies to prove it yields 99 percent analytical sensitivity of CNVs greater than 400 kb in the regions covered by the platform.
[ pagebreak ]
Manufacturers can establish their arrays' 99 percent analytical sensitivity for such CNVs in a study that tests more than 300 well-characterized cases, the paper said.
They can calculate false-positive rates per CNV call by using the same samples from the analytical-sensitivity study. Lower-confidence calls "may be flagged for review and correlation with the patient's clinical indication, but should be confirmed by an independent methodology if reported," the paper noted.
As for analyzing and interpreting array findings, ACMG made three specific recommendations. First, it "strongly discourage[s]" the use of in silico filters to "blind" the user to all data beyond certain, targeted regions.
"We believe that blinding the interpreting geneticist to parts of the genome can lead to oversimplification and misinterpretation of complex and atypical abnormalities, and may in turn result in missed or incorrect diagnoses," according to the paper.
ACMG encourages manufacturers to provide software that enables labs to customize their analysis.
Finally, as arrays provide increased genomic coverage, there is an expected increase in their ability to distinguish between CNVs that are pathogenic, benign, or whose clinical significance is uncertain.
The interpretation and appropriate clinical reporting of these findings is "complex and is the practice of medicine," the paper notes.
Specifically, the paper recommends that individuals who report clinical microarray findings must have "appropriate professional training and certification," including American Board of Medical Genetics-certified clinical cytogeneticists, ABMG-certified clinical molecular geneticists, or ABMG- or American Board of Pathology-certified molecular genetics pathologists.
The Second Paper
The second paper covers in greater detail ACMG's position on evaluating and clinically interpreting CNVs.
It also provides guidelines for reporting findings from chromosomal arrays, including unanticipated but clinically significant findings unrelated to the reason for the referral, and includes recommendations for reevaluating the significance of a CNV based on data from members of a patient's family.
Beyond recommending that the interpretation of array results be conducted by certified professionals, the second paper argues that such professionals should be familiar with well-established contiguous gene syndromes, and that any associated low-copy repeat sequences and critical regions should be recognized and mapped before offering clinical interpretations of microarray data.
Just as the size of a CNV may prompt a clinician to include or withhold its discovery in a clinical report, the second paper argues that the lab's experience with the array used, together with available clinical follow-up data, should be taken into account while assessing clinical significance.
For clinicians, "by far the most relevant interpretive consideration" is the genomic content in CNV intervals, the paper states. "One should consider whether the CNV contains unique, gene-rich sequence or is void of genes or is primarily comprised of repetitive elements or pseudogenes. The gene content should be carefully scrutinized for documented and relevant clinical association."
Here, comparing the CNV with internal and external databases comes into play. ACMG recommends labs track their experience and document CNVs that are pathogenic, benign, or whose clinical significance is uncertain.
For CNVs with which the lab has no prior experience, clinicians should compare such findings with publicly available databases of CNVs in the general population, such as the Database of Genomic Variants, the paper recommends.
ACMG also set standards for reporting findings. For instance, a lab report should include a description of the criteria used to review the data and to include a CNV in the report, if that is the case. Information contained should include a description of position, size, dosage, and relative gain or loss for each CNV detected.
Each reported CNV should also be accompanied by a statement of significance. This means that evidence supporting the interpretation should be summarized and supplemented with appropriate references.
Reports should also include recommendations for a clinical follow-up, which may include genetic counseling, any cytogenetic characterization of the CNV, and evaluation of relevant family members.
[ pagebreak ]
In some cases, CNVs may be found unintentionally in a patient who is found to be a carrier for a recessive condition; are diagnostic or predictive of a certain condition; or are associated with an increased risk of neoplasia. According to the paper, such findings are unavoidable.
"It is not possible to construct a whole-genome microarray platform that purposefully avoids interrogation of any loci associated with the aforementioned cases, especially as many of the findings will be part of a large CNV involving multiple contiguous genes," according to the paper.
ACMG therefore urges clinicians to understand the potential for such discoveries and inform patients and their families before ordering a test.
The group also made recommendations about including family-member data in order to help assess the clinical significance of a discovered CNV. The paper said that such studies should expand to include a larger family analysis.
"It is important to stress that it is difficult and often imprudent to attribute clinical significance based on the inheritance pattern of a CNV in a single family," the paper states. "It is only through ascertainment of significantly large families with multiple affected and unaffected family members segregating a given CNV or ascertainment of multiple individuals with the same CNV that a true measure of clinical significance can be confidently assessed."
One issue that the cytogenetics community has had to grapple with following the advent of chromosomal microarrays is the identification of consanguinity. Earlier this year, geneticists at Baylor College of Medicine recommended that healthcare institutions establish a committee to draft guidelines that deal with issues of consent and reporting of incestuous parental relationships identified with arrays (BAN 2/11/2011).
In that statement, they also said that such guidelines could be drafted by existing ethics committees associated with ACMG, the American Society of Human Genetics, and the European Society of Human Genetics.
Have topics you'd like to see covered in BioArray News? Contact the editor at jpetrone [at] genomeweb [.] com