NEW YORK (GenomeWeb) – Several expert groups have recommended ways in which labs should standardize pharmacogenetic testing methodology, terminology, and reporting of results, which they believe are critical for broad adoption of these tests for personalizing therapies.
In the journal Pharmacogenomics, Kelly Caudle from St. Jude Children's Research Hospital and several other PGx experts recently catalogued existing recommendations for standardizing aspects of testing, such as the Association for Molecular Pathology's (AMP) recommendations for a minimum set of alleles that should be included in CYP2C19 genotyping panels. Caudle and colleagues in their paper also identified gaps in test standardization, for example, bringing more consistency to genotype-to-phenotype translations, something the Clinical Pharmacogenetics Implementation Consortium (CPIC) is currently working on.
Broad adoption of these and similar recommendations could go a long way toward encouraging physician adoption of PGx testing, incorporating results in electronic health records and decision support systems, and improving reimbursement by payors, the experts said. "Standardization across the implementation process will enable efficient and successful integration of pharmacogenomics into routine clinical care at a larger scale," Caudle and colleagues wrote in their paper, but they also noted that more needs to be done to disseminate and encourage acceptance of existing standards.
Importantly, standardization is necessary not just for test adoption but also for ensuring that patients are getting appropriate care, the authors noted.
One area where the lack of standardization in PGx testing could impact patient care is genotype-to-phenotype translations. CPIC is currently working with the Dutch Pharmacogenetics Working Group (DTWG) and around 30 experts to try to align genotype-to-phenotype translations for CYP2D6, one of the most widely tested PGx genes.
CPIC publishes recommendations for drugs metabolized by certain PGx genes, which doctors can use to figure out how a patient's genotype might impact their ability to metabolize that treatment. "That process in translating that genotype to phenotype is very important," Caudle said. "All our recommendations are based on that phenotype."
These recommendations are based on the activity score calculated for each genotype. When an allele has no function it gets an activity score of 0, an allele with decreased function gets an activity score of 0.5 and an allele with normal function gets an activity score of 1. That single-allele activity score is multiplied by the number of gene copies of that allele. Then, the sum of both of the alleles that comprise the genotype become the final activity score.
There is general agreement among experts that a final activity score of 0 is a poor metabolizer, 0.5 is intermediate metabolizer, 1.5 and 2 scores indicate a normal metabolizer, and 3 or greater is an ultrarapid metabolizer. However, there is disagreement among experts about how to translate an activity score of 1. For example, the DPWG considers a CYP2D6 activity score of 1 as an intermediate metabolizer, while CPIC recommends translating this score as a normal metabolizer.
"You can imagine how confusing that was to laboratories or institutions doing the testing. Which recommendation is correct?" Caudle said. "For patients, depending on what lab they are going to, their results could be different. They could be called a normal metabolizer by some labs and intermediate by others."
As a workaround, some laboratories have come up with their own system and created a new phenotype group — a normal-to-intermediate metabolizer — whereas other labs either use the CPIC guidelines or the DPWG's system. "This makes patients and physicians not trust pharmacogenetic testing when one lab is calling [one score] a normal metabolizer and another lab is calling [the same score an] intermediate metabolizer," Caudle said.
To resolve this discrepancy, CPIC decided to work with DPWG and other experts to come up with a consensus system. After six surveys, Caudle said the group has reached consensus (defined as agreement among at least 70 percent of experts), and the experts will soon issue a statement to detail what they've agreed on.
After a consensus statement is published, CPIC and DPWG will incorporate this in their guidelines, present on this at meetings, and inform some 250 CPIC members. CPIC will also reach out to lab accrediting bodies, such as the College of American Pathologists and the lab industry to let them know of the changes. "There is no governing body to make sure everyone implements this," Caudle acknowledged, but noted that the lack of standardization around genotype-to-phenotype translations in PGx testing is a widely recognized problem in the field, so a lot of labs will appreciate a consensus position on this front.
Achieving consistency in genotype-to-phenotype translations could also make it easier to educate physicians, which Caudle noted is challenging when there is disagreement as with the CYP2D6 activity score of 1. "We have to educate them not only on what that phenotype means but also explain to them how we derive that phenotype a bit more, and that there might be discrepancies in the way a lab might report that back," she said.
Another point of confusion for doctors arises from the lack of concordance in the alleles gauged by PGx panels. One study conducted by the Centers for Disease Control and Prevention's Genetic Testing Reference Material Coordination Program gave blinded patient samples to nine research and commercial labs who evaluated them using a variety of methods, including SNP genotyping, copy number assessments, and sequencing. The study revealed that although the results were consistent among labs based on the variants tested, the labs tested for different sets of variants in PGx genes, which resulted in different haplotype calls for the same allele.
Caudle and colleagues noted in their paper that the lack of standardization in this regard is important because it could lead a doctor down the wrong clinical care path for a patient.
This is something that AMP has been trying to tackle. Earlier this year, for example, AMP published a two-tier categorization of alleles to help labs design their CYP2C19 genotyping panels: tier 1 includes a minimum set of alleles and defining variants that should be included in all marketed tests, and tier 2 lists optional alleles that don't meet the criteria for inclusion in tier 1.
"We started with these CYP2C19 genotyping panel recommendations due to the widespread adoption of these tests and our desire to help physicians, pharmacists, researchers, and other stakeholders better understand what these panels include and what the test results mean," said Victoria Pratt, associate professor of medical and molecular genetics at Indiana University School of Medicine and AMP PGx Working Group Chair.
Since these recommendations were published in the Journal of Molecular Diagnostics in May, the feedback has been positive, according to Pratt, though she said it was too early to discuss adoption rates. AMP plans to publish a series of recommendations in this vein for additional PGx genes, she noted.
Meanwhile, CPIC's recommendations to standardize terms for PGx test results have seen some uptake. So far, these recommendations have been endorsed by AMP and are being used by variant classification projects ClinGen and ClinVar, as well as within CAP's proficiency testing programs.
In comparison, a set of recommendations that haven't been adopted as widely come from an international group of experts convened by the CDC. A few years ago, this group issued nine recommendations for how labs should name sequence variants and how they should report results, including listing the variants detected for a patient, all the variants that the test can detect, and the limitations in terms of types of variants the test cannot detect.
An evaluation published last year, by Bousman et al., looked at whether PGx tests for psychiatry were complying with these recommendations. After comparing 20 tests for CYP2D6 and CYP2C19 alleles, the authors found a lot of variability. Although most of the test panels included the major alleles in these genes, no two panels gauge the same combination of alleles and none met all the CDC's recommendations. "Consensus has yet to be reached on which CYP2D6 and CYP2C19 star alleles to include on pharmacogenetic testing panels and pharmacogenetic results reporting could be considerably improved," they wrote.
"It's hard to standardize things when there is no penalty when you don't use these standardized methods," Caudle said. Noting the positive acceptance of CPIC's recommendations for standardized PGx terms, Caudle and colleagues pointed out in their paper that these types of recommendations will remain voluntary best practices until an accrediting organization like CAP endorses them.