NEW YORK (GenomeWeb) – University of British Columbia spinout GenXys recently shared results from a feasibility study of its pharmacogenomic testing and decision support platform, showing that physicians and pharmacists across a group of rural and urban practices could and did use the system when it was available to them.
The study, published in the Canadian Medical Association's online journal CMAJ Open late last month, provides some early evidence that GenXys' approach could be attractive and easy enough to use to spur broader adoption by clinicians who have so far failed to widely adopt PGx approaches to drug prescription.
Martin Dawes, president and co-founder of the company and head of the department of family practice at UBC, told GenomeWeb this week that company's next steps are to build out its commercial customer base and to start studies that will be able to measure the clinical utility of its system and its impact on patient outcomes compared to standard prescribing.
The use of genetic information to help make medications safer for patients — seemingly perpetually on the cusp of wider adoption — has struggled to achieve broad use. However, numerous clinical research programs, medical centers, and commercial companies have still emerged hoping to optimize PGx and push it forward.
For example, the Mayo Clinic has implemented its own system integrating pharmacogenomic testing into patients' electronic medical records in such a way as to provide real-time alerts to physicians who are prescribing medications.
Other groups have come out with tools that they think can help improve implementation and clinical adoption of PGx, for example the Stanford University and Geisinger Health System-led Pharmacogenomics Clinical Annotation Tool (PharmCAT).
Earlier this year healthcare software company Epic Systems agreed to integrate Genelex's PGx clinical decision support tool into its electronic health records (EHRs).
In addition to Genelex there are other players in the field, including newcomer Kailos, which offers a low-cost service that allows patients to seek out testing, although not to actually receive results, without consulting a physician.
Admera Health (formerly GeneWiz) has been marketing its PGxOne test since 2014, and the Mayo Clinic also spun out a commercial company in 2014 called Oneome.
According to Dawes, GenXys believes that its system, called TreatGx, may have a leg up compared to some others because it is designed to function in a different place in the timeline of a clinical visit or appointment.
TreatGx is based on a series of algorithms that Dawes and colleagues built using databases of drug-drug interactions, disease-drug interactions, genotype interactions — even data from clinical trials on drug effectiveness and harm. According to GenXys, the algorithms include over 14,000 rules covering the 24 conditions so far included.
Health Technology investigators from Emily Carr University developed TreatGx's software interface, working with clinicians to make the interface fit into the right part of the primary care workflow.
Unlike schemes that link drugs to genetics — and which send prescribers an alert if they are trying to give a patient a drug that is contraindicated by their genotype — TreatGx is primarily about connecting a particular disease with genetics, and painting a comprehensive picture of what drugs could work.
"We realized a few years ago that trying to put PGx data into EMRs alone just wouldn't work. The EMR alert system is ignored as much as 96 percent of the time, so adding another alert [based on genetics] isn’t going to do it," Dawes said.
"Most of an exam, which is about 15 minutes in Canda, is making a diagnosis, doing a physical exam, collecting symptoms. You usually have about a minute at the end to identify a drug and make the prescription," he added.
This means that a physician is likely to ignore an alert that pops up right at the end of an appointment and could prolong their work significantly.
TreatGx works upstream and provides a list of medications after a physician inputs all of the patient's relvant information, including genotype, disease state or progression, other medications, et cetera.
"We talked to doctors and they felt that having it upstream where it manages everything about the condition, [is] better for patients," Dawes said.
The UBC team's initial development of TreatGx, as well as of a PCR-based PGx testing panel to supply it with the required genetic data, was supported by a grant from Genome British Columbia. The same money also supported the group's initial feasibility research published last month.
In that study, the UBC and GenXys researchers recruited a cohort of six primary care providers (five family practices in both rural and urban areas, and one pharmacy). They picked physicians and pharmacists who had some sort of teaching connection with the department of family practice at UBC and showed an interest in pharmacogenomics.
These participating clinicians then recruited patients opportunistically, or by looking through their EHRs. In all, the six groups enrolled 191 adults with at least one of 10 common diseases — gout, chronic obstructive pulmonary disease, migraine, depression, osteoarthritis, hypertension, hyperlipidemia, atrial fibrillation, osteoporosis, and/or epilepsy.
When they identified a patient for the study, the participating offices then took saliva samples, and sent them to GenXys' lab, where the company tested them using a multi-gene PCR panel covering about 33 variants.
The group chose TaqMan allelic discrimination quantitativePCR assays for the study for use on Thermo Fisher Scientific's QuantStudio 12K Flex platform.
According to Dawes, because some of the genetic targets included in its analyses are difficult to test, the group really wanted separated PCR assays for each, to make sure it could be confident that tricky markers were being picked up appropriately.
"Some of these are really difficult to do on an array. CYP2D6, for example, is a nasty area, and HLA is difficult, so we felt we needed to understand the nuts and bolts of the tests — and the grey areas," he said.
"Mass spec would also obviously be a good alternative," he added, and comparing PCR and mass spec-based analyses directly is a goal for the UBC group's continuing research.
In its initial feasibility study, GenXys was able to successfully genotype 189 patients out of 191 recruited. Of these, about 97 percent had at least one actionable result linked to one of the medications included in the company's TreatGx decision support system.
Physicians and pharmacists participating in the study then used TreatGx 236 times over three months as part of their care of patients.
Obviously, Dawes said, this initial study speaks mainly to the basic feasibility of getting clinicians to adopt GenXys's support tool and to collect and send in samples for the required genotyping. To make real strides in growing the use of this, or any PGx strategy, will require evidence of clinical or therapeutic utility.
To that end, the company has applied for funding for a 4,000-patient randomized study that should help to answer questions about clinical impact and cost effectiveness.
The group is also working with partners in UK and Australia to collect an even larger amount of data. "They won't be doing exactly the same intervention," Dawes said, "but we will hopefully be able to measure the same outcomes."
"With 4,000 we think we will have enough power to answer the question of cost effectiveness, and with 10,000 we'll definitely nail it," he added.
In terms of GenXys's commercial operations, Dawes said the company is currently in a beta period. The company has a few early users in British Columbia. It has also made its testing and its decision support software available directly to individuals with an adapted patient portal that is more accessible to non-medical professionals.
According to Dawes, the company should be able to avoid regulatory challenges because TreatGx doesn't recommend therapies, it only displays them based on its algorithms for determining risk. Moreover, most of the drugs that the system deals with require a prescription from a physician, so acting on the information provided still requires a physician's input.
"We prefer this to be done with a physician," he said. But anyone with a chronic disease probably has three or four health professionals looking after them, plus a pharmacist … so [by opening it up to patients directly] we have a system that allows a patient to share this with any professional they want."
Though the group's feasibility study used a smaller panel, GenXys has since expanded its actual genetic testing platform, which it calls myPGx, to include about 60 tests covering variants with relevance to 74 drugs.
The combination of access to the TreatDx clinical support tool and testing using GenXys' 60-SNP panel currently costs $399.
Dawes said that users can also plug existing genetic data into TreatGx. So far, the panel of genetic variants that TreatDx incorporates has remained fairly stable, but Dawes said that he and his colleagues recognize that in the future, the ability to add new associations without having to test patients with additional single-marker tests may make a more comprehensive assay tool like next-gen sequencing more attractive.
"Right now … you want a cheap proposition with a really valid result, which means PCR and mass spec, but maybe in a few years when the costs of sequencing have gone down [that will make more sense,]" he said.