NEW YORK – Researchers from Ambry genetics and Mayo Clinic last month soft-launched an interactive mutation prevalence tool aimed at helping clinicians identify appropriate testing for their cancer patients.
According to one study, there are approximately 74,000 commercially available genetic tests on the market, with 14 new tests launched daily. The market expansion has particularly favored multi-gene panel tests.
However, the growth in test access and use has also made it more difficult for providers to know which multigene test they should select for patients based on their personal and family history and hereditary cancer testing guidelines. This is where Ambry is hoping its new tool could prove useful.
The soft-launch of this product coincided with the publication of a corresponding study on the preprint server medRxiv on mutation prevalence tables created from the data of 150,000 high-risk cancer patients who had been tested on an Ambry multi-gene panel. The paper is currently being submitted for peer review.
Genetic testing company Ambry has collaborated with researchers at Mayo Clinic since 2012. From March 2012 through December 2016, Ambry and Mayo researchers collected, deidentified, and organized clinical and genotype data from patients referred to Ambry for hereditary cancer multigene panel testing.
Steven Hart, a bioinformatician and a senior associate consultant at the Mayo Clinic, spends a lot of time working with big data and presenting it in a way that geneticists and genetic counselors can understand it.
"I kept getting these questions [such as] 'Well, what if we looked at the age less than 50 years old or 55? And how does that change the number of mutations that we see? And what about family history?" Hart said. "All these what-ifs got me thinking that there's probably a better way to interact with the data. That's where I got the idea of coming up with a web tool that would allow people to ask different questions given their own constraints. People want an easy way to go in and ask their particular question without having to wade through hundreds of pages of supplemental text."
The tool Ambry and Mayo developed could also be used to filter patients according to demographics such as age and ethnicity, as well as their personal or family history of cancer, and match them with the Ambry genetic test that best captures the range of relevant mutations that could contribute to their risk for cancer.
The application's utility also extends beyond Ambry's tests. The 'by gene' tab gives the "unbranded view" of what genes patients should expect to see. Patients can then look for other tests that contain a similar assembly of genes.
The multi-gene panel mutation prevalence tool was created to provide a "pretest probability for people who are having multi-gene panel testing for hereditary cancer," said Jill Dolinsky, director of clinical affairs at Ambry Genetics.
Even after doctors figure out which patients they should test according to the latest guidelines, those guidelines still don't provide much direction in terms of which test is most appropriate, Dolinsky pointed out. The interactive application can compare the chances of a patient getting a positive result on a five-gene panel versus a ten-gene panel, and provide information on what additional information an individual could glean from looking at more genes.
"There are people who know that they have a family history of various cancers, but they don't know how likely it is that they would get a positive result," said Hart. "Should they pay for [multi-gene testing from] one of these testing companies, because insurance companies don't always cover them? This [tool] empowers people to look and see how likely is it that given their situation that they would get back a positive result."
The tool could also be helpful in pretest counseling.
"A genetic counselor could be sitting with a person that is getting ready to undergo genetic testing [and] this [tool] can help facilitate that discussion [by providing] the likelihood of finding a positive result and opening the door to discussions about what exactly that means for them," said Hart.
This product, however, does not replace a full evaluation for hereditary cancer predisposition with a genetics expert and should not be used directly to make patient treatment decisions, Ambry advises.
As insurers become more interested in cutting back their genetic testing spending, a product like this could conceivably serve as a reference point for determining if a test is covered, or provide information that clinicians can use to make the argument to payors for why a test should be covered. But it's still early days and not clear yet if it will be useful for this purpose.
Clinicians have been using pretest probability models for decades to inform genetic decision-making. For example, BOADICEA, BRCAPRO and Myriad's prevalence tables have been used as a gold standard to assess the likelihood that an individual is a mutation carrier for the gene BRCA1.
Other similar models like the PREMM model or the MMRpro also exist for detecting mutations relevant to hereditary nonpolyposis colorectal cancer, or Lynch syndrome.
Although products like these have helped doctors figure out whether a patient should be tested or insurance firms determine whether to cover a performed test, Dolinsky noted that they also have limitations.
For example, she highlighted that most existing models are based on data from a small cohort of patients. Recently, Ambry competitor Color Genomics unveiled a similar application that provided cancer risk and prevalence information online for 30 hereditary risk-associated genes. But that information was based on 50,000 individuals, most without a history of cancer, Dolinsky noted.
Ambry aimed to create a dataset large enough to show strong, visible association, and easy for anyone to use. Some of the older programs and prior probability models are more complex, based on statistical algorithms that need to be downloaded or purchased, according to Dolinsky. With Ambry's tool, clinicians can go to its website, plug in values to the appropriate filters, and instantly get related gene information and test recommendations.
Additionally, a lot of the older models only take into account a limited set of genes, Dolinsky said. Ambry's tool contains information on 31 cancer predisposition genes that are commonly included in multigene panels for hereditary cancer risk assessment.
Another advantage she highlighted is that Ambry has tested a more diverse patient population, compared to earlier applications based largely on data from Caucasians and those with Ashkenazi Jewish ancestry. In Ambry's cohort of approximately 150,000 patients, around 19 percent self identified as African American, Asian, or Hispanic. This data allows users to tailor the mutation prevalence estimates according to their patients' ethnic background which may impact testing decisions, according to Dolinsky.
It can take two to three weeks to get results from a genetic test, and the mutation prevalence tool has only been available under a soft launch for a month. As such, Ambry hasn't yet evaluated what value it is adding to patient care in terms of picking the most appropriate tests.
While Ambry tests more patient samples, they will continue to feed that data into the application's algorithm, which in turn will improve prevalence estimates.
"This is definitely just the first iteration. We'll continue to work on the tool, expand the tool, improve the tool, get feedback from clinicians and others about its utility," said Dolinsky. "I think continued efforts to update this tool and others like it are going to be really beneficial for patients, providers in general, especially as panel testing is more prevalent."