Precision medicine arrives in RA: Biomarker test predicts TNF inhibitor response
By: Rob Volansky
If there was any question about whether rheumatologists would utilize an effective biomarker that can predict response to a commonly used class of drugs, it was answered by a recent study in Rheumatology International.
The biologic era has given rheumatologists more options than they have ever had before. However, a clear understanding of which patients may benefit from TNF inhibition vs. an interleukin (IL)-6 or janus kinase (JAK) inhibitor remains elusive in the specialty. The demand for biomarker tests to aid in these decisions is high.
The PrismRA (Scipher Medicine) test, which predicts response to TNF inhibition in biologic-naïve patients, can help meet that demand. The test offers a scale that shows a signal of no or low non-response, moderate non-response or high or very high non-response to these drugs.
James Mossell, DO, a rheumatologist in Tifton, Georgia, and colleagues surveyed 248 rheumatology providers as to the clinical utility of the test in daily practice.
Results of the 32-question decision-impact survey indicated that 80.2% of respondents agreed that the test would improve clinical decision-making, while 92.3% reported that it would improve confidence in prescribing decisions and 81.5% said it would be a useful tool when a TNF inhibitor was among the potential clinical choices for a given patient.
Some 89% of rheumatology providers would change from a TNF inhibitor to another class of drugs if the PrismRA score indicated a likely non-response.
When the test resulted in a signal of a high rate of non-response, selection of a TNF inhibitor declined by 81% (range, 79.8% to 15.3%) among respondents. When the result indicated a very high signal of non-response, this decline was 86% (range, 79.8% to 11.3%).
Conversely, 79.8% (range, 11.3%-25.4%) of respondents said they would be more likely to prescribe a TNF inhibitor if no signal of non-response was detected.
Respondents felt so confident in the test that 84.7% reported that payers should provide coverage for it, according to the findings.
Healio Rheumatology sat down with Mossell to discuss the need for biomarker tests in rheumatology, the utility and availability of PrismRA and the impact it can have on a shared decision-making treatment paradigm.
Q. What are your thoughts on the unmet need for biomarker tests in rheumatology?
A: Clearly, there has been an unmet need for a test like this for 25 years or more. We have been waiting for the ability to order a test to stratify which patients would respond to which medications for a long time. PrismRA has helped us stratify potential response to TNF inhibitors.
Q. How does PrismRA narrow the margin of error when selecting a treatment?
A. These days, we have a lot of different options when we are seeing an RA patient in clinic. When we are faced with making a decision in a patient who has failed on a conventional synthetic disease-modifying antirheumatic drug, we can use a TNF inhibitor, a JAK inhibitor, an IL-6 inhibitor, there are B-cell therapies, T-cell therapies.
Unfortunately, we are in a situation where commercial insurance essentially dictates to us what we can do. The tiers that commercial insurance push us through usually have us use an anti-TNF drug in the first line. The response to TNF inhibitors in the first line is only about 30% to 40%. If you are in the dark about which treatment would be most effective and you put them on an anti-TNF drug and it is not a good choice, you are losing time. It is critical to get a patient into remission or at least a place with low disease activity as soon as possible to save time and money.
Q. Could you talk about the potential cost savings as a result of preemptive tailored therapy?
A: While we did not cover the cost-benefit analysis in our paper, another paper, authored by Martin J. Bergman and colleagues, found that risk stratification using PrismRA would lead to a reduction of ineffective spend of $7,379 per patient per year, along with a reduction in overall and ineffective spending in biologic pharmacy.
Q. Does the accuracy of PrismRA change depending on whether the patient is treatment-naïve, or whether they have experience with a TNF inhibitor, or with another class of drugs?
A: As it stands right now, the PrismRA may be used only in biologic treatment-naïve patients. Now, it does, in fact, predict a response to TNF inhibition in those patients, but not in patients who have been treated with biologic therapy.
Q. Could you talk about the availability of PrismRA?
A: Well, it is available to us in southern Georgia. We are utilizing it now. The paper we published is really looking at how it is being used and perceived by various rheumatologists across the country. Our feedback is clearly showing us that the majority of practitioners felt this would be a useful process for them. There is clearly a need.
Q. You mentioned that the product was easy to use and interpret. Could you talk a little about that aspect of the test itself?
A: I want to stress that it is not a binary result, a yes or no. There are no reports reported as conclusive or inconclusive. It is a linear scale from one to 25, which can then be grouped as no signal, a moderate signal, a high signal or a very high signal. This can help a clinician choose a TNF inhibitor or not. It is easy to act on, and easy for patients to understand to aid in the clinical decision-making process.
Q. Shared decision-making is on the rise in rheumatology. Could you address how this test fits into that treatment paradigm?
A: Certainly, there are some patients who want to be told what to do. But an increasing number of patients want to be actively engaged in the conversation. This would help with those decisions. The choice is fairly clear if there is no signal or a low signal of non-response, and if there is a high or very high signal of non-response. But when the result is a moderate signal, some patients may choose another agent, while others are willing to try a TNF inhibitor.
Q. Apart from indicating that a TNF inhibitor will not work, does the test offer any insight as to which class of drugs will work?
A: Unfortunately, no. It only predicts response for TNF inhibitors.
Q. How accurate is the test, exactly?
A: PrismRA identified half of all true positive inadequate responders with a positive predictive value of 89.7% (95% CI, 79.5%-95.7%). This result had an odds ratio of 6.57 (95% CI, 2.75-15.7). PrismRA did not detect a molecular signal of inadequate response to TNF inhibitor therapies in 46 of the 53 patients who had an observed ACR50 response at 6 months, resulting in a specificity of 86.8% (95% CI, 72.4%-94.1%).
For more information:
James Mossell, DO, can be reached at 2227 Highway #41N, Tifton, GA 31794;