Profound Discrepancies Seen Between DTC Genetic Results, Confirmatory Testing

Raw data from direct-to-consumer (DTC) genetic tests have an “alarmingly high” false-positive rate, according to a study published March 22 in Genetics in Medicine. Additionally, the study found that misinterpretation of variant risk, coming from both DTC companies and third-party interpretation services. Together, these inaccuracies could have potential clinical impact, highlighting the importance of confirmatory clinical-grade testing for DTC raw variants, as well as results interpretation by health care professionals with genetic training, the authors say.

“While having access to raw genotyping data can be informative and empowering for patients, this type of information can also be inaccurate and misinterpreted,” says lead author Stephany Tandy-Connor, from Ambry Genetics (Aliso Viejo, Calif.).

For years concerns have been raised about the accuracy and methods of DTC testing. Central to these concerns are DTC tests’ lack of comprehensive variant assessment. For example, 23andMe’s genetic health risk test reports on one variant in the LRRK2 and GBA.3 genes linked to Parkinson disease. However, experts say there are additional known pathogenic variants in these two genes as well as additional genes clinically associated with Parkinson disease (like SNCA and PARK2/PARKIN.4) that 23andMe does not report on. Additionally, some DTC companies provide customers their raw genotyping data that patients can provide to third-party interpretation services for a fee. The comprehensiveness of that analysis may not be known or understood by the patient. In contrast, the authors of the Genetics in Medicine study say, clinical diagnostic tests are generally comprehensive—examining the full coding sequences of all known genes associated with a disease.

“Many of the DTC genetic testing laboratories use a form of single-nucleotide polymorphism genotyping array for their assay. This particular methodology is analogous to spot checking an individual’s DNA with coverage at only specific preselected sites. This is not comprehensive full-gene sequencing nor does it include gross deletion or duplication analyses, which are both routinely part of clinical diagnostic testing.

The present study investigated the concordance between confirmatory clinical test results and variants identified in the raw data by DTC genetic testing of 49 patients referred to Ambry Genetics between January 2014 and December 2016. Medical geneticists/genetic counselors and oncologists accounted for the majority of confirmatory test orders (40.8 percent and 20.4 percent, respectively).

Testing of cancer genes comprised 87.8 percent of the orders, with submissions made for 26 unique variants, including four located within deep intronic regions that Ambry categorized as “well beyond the analytical range of most clinical laboratories.” Confirmatory testing was performed by Sanger sequencing for single-site analysis (44.9 percent) or next-generation sequencing analysis with Sanger confirmation for samples received for multigene panels.

The researchers found that 40 percent of variants reported in DTC raw data were false positives and 16 of 17 of these false positives were in cancer-related genes. These false positives were seen in a variety of genes, including BRCA1/2, CHEK2, TP53, ATM, MLH1, and COL3A1.

In addition, eight variants in five genes (ATM, BRCA1, BRCA2, COL3A1, and COL5A1) that were reported in the “increased risk” category in DTC raw data or by a third-party interpretation service were classified as benign, common variants at Ambry Genetics as well as several other clinical laboratories, based on their presence in in publicly available population frequency databases.

“The misinterpretation and potential inaccuracy of the raw data pose substantial risks to individuals who obtain this type of information from a DTC company,” writes Tandy-Connor and colleagues. “It is crucial that clinical confirmatory testing be performed on any variants reported in the raw data provided by a DTC company prior to any changes in medical management to confirm the presence of that variant in the individual as well as an accurate classification.”

Takeaway: This study provides some of the first evidence of the inaccuracies of DTC genetic testing based on raw genetic data and variant classification.