NEW YORK (GenomeWeb) – SomaLogic has developed a nine-protein panel for assessing the risk of adverse cardiovascular events in patients with coronary heart disease.
The company published a paper describing the panel in detail this week in the Journal of the American Medical Association and plans to launch it as a commercial test performed by its CLIA lab this fall, Stephen Williams, senior author of the JAMA study and Somalogic's chief medical officer, told GenomeWeb.
The firm developed the CHD panel on its SomaScan platform, which uses its Slow Off-rate Modified Aptamer, or Somamer, affinity reagents to measure proteins of interest. These Somamers – essentially short strands of DNA – bind to protein targets in a sample of interest. They can then be quantified via microarrays, with the quantity of a given Somamer corresponding to the quantity of its target protein.
The version of the SomaScan platform used in the study quantified 1,130 different proteins. The current version of the platform can measure 1,310 proteins, and the company maintains for internal use a custom platform it developed in collaboration with Novartis that can measure more than 4,000 proteins.
In the JAMA study, the researchers analyzed 938 samples from CHD patients to develop and validate the nine-protein panel for predicting the four-year probability of these subjects experiencing myocardial infarction, stroke, heart failure, or death.
In terms of discriminating between patients who would and would not go on to have an adverse event, the panel performed with a C statistic of .70, providing a slight improvement over the commonly used Framingham secondary event risk score, which was .64.
University of California, San Francisco Professor Peter Ganz, first author on the JAMA paper and an unpaid member of SomaLogic's advisory board, suggested, however, that this score did not fully account for the test's potential value.
More indicative of its potential clinical use, he told GenomeWeb, was its ability to stratify the patient population by risk level. For instance, he said, patients in the top risk decile had four-year event rates of 60 percent to 80 percent, while patients in the lowest risk decile had event rates of less than 10 percent.
"From the standpoint of being a clinician and thinking about personalizing healthcare, that is crucial," he said. "Because somebody whose risk is six percent will very likely choose to be treated differently from somebody whose risk is 80 percent. It is the risk stratification that is really important, and these nine proteins are very good at risk stratification."
An accompanying JAMA editorial by Brigham and Women's Hospital researcher Marc Sabatine also made this point, noting that while the Somalogic panel provided only "modest discriminatory accuracy," a focus on that "is misplaced in studies of biomarkers for long-term prognosis."
Such discriminatory accuracy is essential in tests intended for making an immediate diagnosis of a condition, Sabatine noted. But for tests intended for long-term prognosis, "clinicians are typically less interested in discriminating risk between two patients and more interested in being able to better calibrate a particular patient's risk of adverse events," he wrote, adding that the predictive data generated by the Somalogic panel in the JAMA study "would be important to physicians and patients alike."
Ganz said that one potential use for the test would be identifying particularly high-risk patients who could be candidates for a variety of more powerful treatments that have either recently come to market or are slated for launch in the near future.
"Until now, if you have stable coronary heart disease, it is kind of looked at as a homogeneous population in terms of risk, and everyone says, 'well just treat everyone the same'," he said, adding that this treatment typically consists of aspirin and statins and possibly beta blockers for symptomatic patients.
But recently, a new class of drugs, PCSK9 inhibitors, has come to market. These agents, which include Sanofi's Praluent (alirocumab) and Amgen's Repatha (evolocumab) can have a dramatic effect on patients' cholesterol levels, lowering them an additional 50 percent to 60 percent beyond what can be achieved with statins, Ganz said.
However, these drugs, which are injectable biologics, are expensive, running more than $10,000 a year. Therefore, it might be desirable to target them to high-risk patients as opposed to the overall CHD patient population, he said.
Using a test like the SomaLogic panel "may offer a rational way of deciding who should be on [PCSK9 inhibitors] and who should not," he said.
The test could also be useful for guiding therapy with several other drugs currently in clinical trials for CHD that have relatively significant side effects, Ganz said, for example low-dose methotrexate and Novartis' Ilaris (canakinumab), an interleukin-1β blocker.
SomaLogic's Williams said, in fact, that the company plans to initially target pharma companies as customers for the panel for use in drug development. He cited as one potential application the enrichment of trial cohorts for patients likely to suffer an adverse cardiac event.
"We have lots of relationships with different pharma companies, and so we think an initial use of the test will be to improve the efficiency of drug development," he said.
At the same time, the company will begin piloting use of the panel in clinical practice this fall, Williams said. Echoing Ganz, he suggested it could be useful in identifying patients most likely to benefit from expensive drug treatments.
"People have argued about whether at $14,000 a year, PCSK9 inhibitors really should be applied to everybody the label suggests," he said. "The cost of doing that is high. So you can use this kind of test to move medical resources to the people who need them the most."
He added that the test might also prove useful in improving patient compliance with existing drug treatments and lifestyle recommendations.
"These risks are kind of deniable, and the question is, if they become a bit less deniable, will that enable some fraction of the population to make better choices?" he said.
Earlier this year, SomaLogic announced a collaboration with the University of Oxford, and Williams said that the company is in discussions with the Oxford Academic Health Sciences network about working together to look at the panel's clinical utility.
The company has also begun looking at patients without coronary heart disease but with various risk factors to see if it can identify proteins that would similarly help to stratify these patients by risk, Williams said.
Ganz also said patient populations like chronic kidney disease and HIV sufferers would be interesting to look at. While these patients are at elevated risk for CHD, traditional risk factors aren't very informative for them, he said. Another line of research might be testing if the panel can differentiate responders from non-responders to various treatments, he added.
Ganz said he has been involved in targeted biomarker research for years, working with pharma firms, including Pfizer, on using markers in clinical trials for statins. This work, he said, showed him the limitations of a targeted approach and piqued his interest in large-scale omics approaches.
"Basically, we're limited by our knowledge of biology [in targeted work], and we know that our understanding of human biology is very incomplete," Ganz said. "So I always wanted to do something on a large scale that didn't necessarily rely on knowledge, an agnostic approach."
SomaLogic's platform intrigued him in that it offers higher throughput and sensitivity than mass spec and higher multiplexing than immunoassays, he said. "The fact that we could measure 1,130 proteins was, I would say, a major step in the right direction."