Expert Perspectives on Advances in Precision Medicine in Treating Rheumatoid Arthritis - Episode 4
An expert rheumatologist discusses the Accelerate Information of Molecular Signatures (AIMS) study and the comparative cohort study of the MSRC test.
Vibeke Strand, MD, MACR, FACP: The AIMS study, which is named Accelerate Information of Molecular Signatures, was to essentially develop a clinical database of longitudinal data from patients with RA [rheumatoid arthritis] treated by providers in real-world settings. In other words, in clinical practice, with the idea of looking at what happens when the MSRC [molecular signature response classifier] test is utilized or not utilized. It includes more than 70 private and academic rheumatology practices across the United States. The major response was looking at ACR50 [American College of Rheumatology criteria for 50% or more improvement] at 24 weeks.
They were looking at therapy, whether it was selected according to the MSRC test or not, and they had MSRC results for 73.5%, or 277 of 377 patients tested. Interestingly enough, there were those who didn’t follow the MSRC results, and about 60% of that was because they couldn’t due to the formulary restrictions or having insurance that would not cover an alternate mechanism of action therapy. When you look at the data, the primary data are at 24 weeks, ACR50 responses were much higher in those patients who were prescribed a therapy that was in alignment with their MSRC results. Overall, that was about 40%. Those who had a signature of nonresponse significantly improved if they didn’t get a TNF [tumor necrosis factor] inhibitor. And if they didn’t get a TNF inhibitor, their ACR50s were 34.8% vs those who were prescribed a TNF inhibitor, which was only 10.3%, which was a P value of .05. Again, showing the statistical significance of testing this prospectively. We also noticed significant changes with the CDAI [Clinical Disease Activity Index], ACR20, and ACR70 responses, and for responses earlier, at 12 weeks.
Now, the predicted TNF nonresponders who were treated with an alternative mechanism of action hadabout a 1.8-fold greater improvement in CDAI scores than those who received a TNF inhibitor. Those numbers at 12 weeks were 12.2% vs 8%, P value nonsignificant. But at 24 weeks, it was 14.2% vs 7.8%, which was a P value of .009. In other words, in predicted TNF nonresponders with high disease activity at baseline, using this test corresponds to a 43% relative improvement in achieving a lower CDAI disease activity level. Thus, it makes sense to follow the prediction of nonresponse to a TNF inhibitor and prescribe an alternative mechanism of action. This goes against formulary and insurance restrictions in at least some settings.
This second study was a comparison of 627 MSRC-tested patients from the AIMS study who were now compared to an external control arm from US electronic health records, which included 2721 patients between August of 2020 and August of 2022 who were starting a new biologic or targeted synthetic DMARD [disease-modifying antirheumatic drug] or a TNF. The idea was to compare the AIMS data with a broader cohort of patients starting therapy. What was interesting is that in the MSRC-tested cohort, the 627 patients, 58% had a signature for nonresponse to TNF inhibitors, and 70% received treatment according to their MSRC results.
Interestingly, in the TNF-treated patients, the MSRC had an 88% positive predictive value and 54% sensitivity. Those MSRC-guided treatment patients were significantly—P value less than .0001—more likely to respond to a biologic targeted synthetic than those treated according to standard of care, based on what we don’t know about how to treat. The difference was that the CDAI low disease activity and remission outcome was achieved in 36% vs 21.9%; that was an odds ratio of 2.1. And essentially, CDAI remission, an even more stringent outcome, was 10.4% vs 3.6%, and that was an odds ratio of 3.14. Finally, if it was just CDAI minimum improvement difference, it was 49.5% vs 32.8%. Again, an odds ratio of about 2.
These data clearly show again that the MSRC is proven to be predictive and that there is a significant difference between use of MSRC-guided therapy vs use of whatever therapy is chosen. What this means is that MSRC-aligned treatments are clearly better in terms of responses than what we obtain in regular clinical practice. You may remember that I told you the best published responses to TNF inhibitors outside of clinical trials is only about 30% to 40%.
Transcript edited for clarity