NEW YORK – After publishing positive results from a clinical validation study of its PrismRA assay, Scipher Medicine is gearing up to launch the test by the end of the summer for clinical use in guiding drug treatment decisions for patients with rheumatoid arthritis (RA).
Company researchers and academic collaborators reported the retrospective validation results last week in the journal Network and Systems Medicine.
The authors reported that the company's RNA-based molecular signature, called PrismRA, accurately discriminated RA patients who failed to respond adequately to treatment with tumor necrosis factor inhibitor therapies, and for whom physicians might want to consider an alternative treatment.
PrismRA comes out of a signature development method that Scipher created in which it characterizes protein interaction networks in order to identify putative predictive gene expression markers.
Jeffrey Curtis, professor of Medicine in the division of clinical immunology and rheumatology at the University of Alabama at Birmingham and a lead author of the study, said that TNF inhibitor therapies fail to help nearly 70 percent of the overall unstratified patient cohort. Based on the data he and his colleagues saw in their validation, PrismRA could help prevent a significant number of these individuals from taking a drug that is unlikely to ameliorate their disease and move immediately to one that could work better for them.
"If somebody has been on methotrexate and stopped responding, you want to start a biologic or a JAK [inhibitor]. Right now, the default, mostly for reasons of inertia and insurance [requirements], is you start a TNF because we've never had any way in the past to do anything better or, frankly, more erudite than that," Curtis said.
"Insurance companies know that on average [response rates are] roughly all about the same so with no companion diagnostic pre-Scipher … they're going to make you start with the cheapest one," he added. "But now there's potentially a diagnostic that would help you pick out the people for whom a TNF is probably not the best choice."
It's a narrow indication amid what can be a much more complex, iterative decision-making process as RA patients progress through different disease states and stages, Curtis argued. However, for these specific patients at this one treatment decision time point, the benefit is pretty clear, he said.
In their validation study, the researchers used samples and outcome data from a group of biologic therapy-naive patients from the Comparative Effectiveness Registry to Study Therapies for Arthritis and Inflammatory Conditions (CERTAIN) trial, in which individuals with RA were treated with adalimumab, certolizumab pegol, etanercept, golimumab, or infliximab at the discretion of their physician and followed longitudinally for at least 6 months.
Investigators used an initial training cohort of 143 CERTAIN subjects to develop a predictor incorporating gene expression indicative of drug response, RA-associated SNPs, and clinical indicators in a 70-biomaker panel.
According to the authors, the training cohort population had a response rate to anti-TNF therapies of 30.8 percent overall, which is representative of the general population and reflects the real-world prospective collection approach of the CERTAIN study.
In the end, the optimal response discrimination algorithm didn't require all 70 candidate biomarkers and ended up including 10 SNPs, eight transcripts, two laboratory tests (CRP and anti-CCP), and three clinical metrics (sex, body mass index, and patient disease assessment).
Applying this final PrismRA algorithm to an independent group of samples from 175 biologic-naive RA patients, the investigators were able to identify patients who had a sub-adequate response to anti-TNF treatment with a positive predictive value of 89.7 percent and specificity of 86.8 percent.
Patients predicted to be nonresponders had an observed response rate of just 10.3 percent on anti-TNF therapies, far below the average seen in the general RA population. In contrast, predicted responders had a response rate of 43 percent, a significant improvement from what has been observed among unstratified patients.
Overall, the data suggest that a patient identified by PrismRA to be a non-responder is about 6.6 times more likely to inadequately respond to a TNF inhibitor therapy compared to someone lacking that molecular signature, Scipher said.
Curtis said that he considers the current retrospective validation data "foundational and important," but he said that it isn't yet the "end all, be all with a large-enough sample size that lets us have complete confidence that we're done."
"I don't think Scipher thinks that either," he added. "This is going to need to prove itself with no additional replication studies. I think that that's an additional step that, frankly, they're working on feverishly right now."
Scipher CEO Alif Saleh said that the company has about 260 patients already enrolled in a prospective validation following up on the newly published data and expects this study to read out in October.
Other companies have tried to create tools to guide treatment for autoimmune disease, specifically RA. A longstanding example is the Vectra DA test, which was developed by Crescendo Biosciences (acquired by Myriad Genetics) and offers clinicians a tool to more accurately track patients' disease activity.
Curtis was also involved in validation studies of VectraDA and said this disease tracking application is also valuable to the rheumatology community but agreed that the therapy prediction PrismRA offers could make a much more dramatic impact on patient outcomes. "With Vectra, it's really [proposing] a better way to track," he said. "That may be helpful, but that's not going to help you pick drug A versus B or class A versus B."
According to Curtis, the first step for Scipher now is to replicate and extend the validity data for PrismRA, which its current prospective follow-up should help do.
The second piece, he added, is to prove its clinical utility and health economic value. "Presumably, it feels intuitive to me at least, that you should save money because putting somebody on a drug that doesn't work all that well is costly," he said. "Poorly managed disease is just expensive, so if you can put people on a drug that works better and you can get them closer to remission or at least low disease activity, you'll save money."
With plans to launch the assay by the end of this summer, Saleh said Scipher is in "heavy launch mode."
"We have a pretty substantive pipeline, both on the rheumatology practice side and on the payor side to roll out the test in 2020, albeit to a targeted population," he added. "Then 2021, we're going to go more broadly."
Curtis stressed that for this type of testing to offer a more comprehensive benefit to RA patients, Scipher or others will hopefully be able to develop additional molecular signatures that can inform other clinical decision.
"Biologic-naive is frankly probably what the market needed most," he said. "But you're only ever that once, and the question of how to choose drug number two or three, that's going to be important, too."
Most likely that will require new unique classifiers. The same PrismRA signature probably won't answer all these questions in one.
"It's very possible that there's the Scipher test for biologic-naive patients. And then there's [another] for test for people that are currently or have been on a TNF that are looking for a [second] drug," Curtis said. "There may ultimately be a family of tests that helps you pick what's the next drug or at least what's the next mechanism of action we should go to, or if not … [at least] tell us who shouldn't get the thing that we would otherwise do."
Scipher hasn't detailed plans in this vein, but Curtis said that the fact that the company was successful in creating a strong differentiator for the biologic-naive population suggests it could work similarly for other points of decision in RA treatment.
In the meantime, the company has already applied the same approach to ulcerative colitis, with a PrismUC test in development. Saleh said the firm has also begun work on doing the same for multiple sclerosis.