NEW YORK – Pharmacogenomic testing has had a variety of barriers to implementation as laboratories work under complicated regulatory guidance and a lack of standardization for the testing and delivery of results.
Some of those barriers include a lack of provider awareness about PGx testing, non-standardized nomenclature, the availability of tests and timeliness of results, and laboratory preparedness to adopt PGx and informatics systems to aid in providing those tests.
A PGx program at the University of Chicago is working to address some of these issues and make PGx testing more widely available to a variety of patients. At the American Association of Clinical Chemistry's annual meeting this week, the university's KT Jerry Yeo presented on the school's PGx laboratory and its implementation of the Genomic Prescribing System. Yeo, a professor of pathology and medical director of the clinical chemistry, clinical pharmacogenomics, and clinical translational mass spectrometry laboratories at the university, also highlighted potential solutions to PGx implementation problems.
Those solutions, which the university's PGx lab have implemented, are developing and validating relevant PGx variants in a CLIA-certified clinical lab, creating a web portal for easy interpretation and action, preemptively genotyping all relevant PGx variants from patients, and demonstrating clinical usefulness guidelines, Yeo said.
UChicago's Center for Personalized Therapeutics, which opened in 2010, first developed a custom PGx panel analyzing different variants, Yeo said. They then created clinical summaries of pharmacogenomic literature that could be used to interpret tested genotypes, and developed the Genomic Prescribing System (GPS), a bioinformatic software portal "for the purpose of electronic dissemination and instantaneous availability of results for access at any clinical moment," Yeo said. Since implementing this system, the center has also measured the results of deploying this pharmacogenomic intervention for best prescribing, he continued.
The PGx panel developed at the University of Chicago Advanced Technology Clinical Laboratory uses the Thermo Fisher Scientific QuantStudio 12K Flex platform with OpenArray custom plates to assess 480 PGx variants per patient, as well as a Hologic Invader assay and Thermo Fisher's TaqMan Copy Number Assay to assess 30 CYP2D6 variants and copy number, respectively. The lab measures the CYP2D6 enzyme because it is responsible for metabolizing many clinically important drugs, Yeo said. Since its development, the panel has been updated four times in the past six years, he added.
Once DNA is extracted from a patient's blood sample, it either undergoes CYP2D6 analysis or OpenArray genotyping, which ultimately leads to a molecular report defining the correlation between genotype and phenotype, which is then inputted to the GPS, Yeo said. The GPS is consistently updated to reflect current available research and can determine how a patient will respond to certain medication.
The GPS has been used in multiple projects, including the 1200 Patients Project, the ACCOuNT Project, and the ImPreSS Trial, all of which are at the University of Chicago. Since 2011, the 1200 Patients Project at the university has used the GPS for patients undergoing pharmacogenomic testing and has analyzed more than 3,000 outpatient encounters from more than 1,200 patients.
The project includes adult outpatients who are receiving care from University of Chicago Medical Center physicians and who are taking at least one regularly used prescription medication, but not more than six medications, Yeo said. The physicians enrolled are from a range of disciplines, with most from primary care and some from oncology, cardiology, gastroenterology, and a variety of other areas, with an average of 20 years in practice.
The GPS system is based on a "traffic light signal" approach, where red means do not take the drug, yellow means explore other drug options, and green means "proceed to take the drug," Yeo said. Within the electronic system, clicking the yellow signal would provide a list of alternatives to a specific drug and whether they would be appropriate for the specific patient. The system also notes which drugs have had at least one study finding the medication appropriate, while a different study advises caution.
When the red light appears, it comes with a detailed explanation of why the drug isn't recommended for that genotype, what the US Food and Drug Administration label for the medication says, and links to further research on the drug and genotype. Levels of evidence are also listed for each drug.
Yeo noted that for warfarin, instead of the traffic light signal method, a prediction algorithm is used that combines genetic and non-genetic factors, such as age, weight, and race to calculate a stable warfarin dose. Warfarin dosages have been difficult to determine because they are influenced by both genetic variants and other comorbidities, such as age and body size.
Throughout the project, when PGx info was delivered it influenced the decision to prescribe a drug 56 percent of the time and influenced medication discontinuation 66 percent of the time. Yeo emphasized that "most importantly, no PGx high-risk medications were prescribed during the entire study when the GPS tool was consulted," reducing patient risk. In addition, there has been a 68 percent GPS accession rate and physicians have accessed red light signals 100 percent of the time, yellow signals 75 percent of the time, and green light signals 40 percent of the time, Yeo said. This implies that physicians were using GPS-generated signals, particularly the ones predicting adverse effects "to guide decision-making in drug prescription," he continued.
There is also an ongoing Phase 2 randomized study of preemptive genotyping where all enrolled patients are genotyped, but data is provided to physicians for only half of those patients, Yeo said. The researchers measure clinical outcomes, including genotype-associated adverse events. He added that other modalities will be integrated into the tool "to incorporate diagnostics testing, perhaps imaging studies, other environmental and social factors, to produce a more fine-tuned and accurate predictive model."