NEW YORK (GenomeWeb) – Medicare contractor Palmetto GBA issued a draft local coverage determination last week proposing that it end coverage of Myriad Genetics' rheumatoid arthritis test Vectra DA.
Explaining its reasoning for ending coverage of Vectra DA, Palmetto cited recent studies that raise questions as to whether the test is effective in assessing disease activity in RA patients. The draft LCD also noted the lack of prospective clinical utility data on the test and the fact that it is not included in the 2015 American College of Rheumatology treatment guidelines.
The comment period for the draft LCD begins on Feb. 6, 2017, and ends March 23, 2017. Medicare can issue a final decision at any point after the close of the comment period.
Myriad responded to the draft LCD in a statement, noting that the test's efficacy "has been demonstrated in more than 20 studies with more than 3,000 patients," and adding that the test has been ordered for roughly 300,000 patients in the US.
Company officials also challenged the findings of the main study cited in the draft LCD as cause for concern, an analysis of patients in the AMPLE (Abatacept versus Adalimimab Comparison in Biologic-Naïve RA Subjects with Background Methotrexate) trial looking at how well disease activity in these patients correlated with their Vectra DA scores.
The study, published in the September issue of Arthritis & Rheumatology, found that Vectra DA scores "did not reflect disease activity as assessed by radiographic non-progression or using the CDAI, SDAI, DAS28-CRP, or RAPID-3 in patients treated with abatacept or adalimumab in the AMPLE trial," the authors wrote, adding that their findings "indicate that [Vectra DA] should not be used to guide RA management decisions, particularly in patients treated with abatacept or adalimumab as a first biologic agent."
David Chernoff, senior vice president, medical affairs for Myriad subsidiary Crescendo Bioscience, which houses the company's Vectra DA business, said the company disputed these conclusions.
He noted, for instance, that some discordance between RA disease activity as measured by the Vectra test and existing clinical tests like CDAI, SDAI, DAS28-CRP, or RAPID-3 was to be expected and was, in fact, desirable, as it indicated that the test provided additional information beyond that provided by conventional clinical measures.
"The fact that it is discordance with clinical exams is exactly why we built the test," Chernoff said.
He suggested that the fact that the A&R study found little correlation between Vectra DA scores and disease activity as measured by radiographic progression stemmed from the study researchers' approach to their analysis. In the study, the researchers observed that a high percentage of non-progressing patients had high Vectra DA scores, indicating that the test could not effectively predict if a particular patient would exhibit radiographic progression. (Vectra DA returns a score of between 1 and 100, with higher scores indicating higher levels of disease activity.)
Chernoff noted, however, that the same data showed that only a small percentage of patients with low Vectra DA scores exhibited radiographic progression, demonstrating that the test is effective at ruling out patients who are unlikely to progress.
In the A&R study, "lower scores are associated with almost 100 percent non-progression," he said. "The negative predictive value of the test is extremely high for non-progression."
The study found that of 39 abatacept-treated patients with low Vectra DA scores at one year, one, or 3 percent, exhibited radiographic progression. For moderate and high scores, the percent progressing at one year were 5 percent (three of 60), and 18 percent (15 of 82), respectively. At two years, 4 percent (one of 26) of low-scoring patients, 21 percent (11 of 52) of moderate-scoring patients, and 24 percent (12 of 49) of high-scoring patients exhibited progression.
A similar pattern held for the adalimumab-treated patients. At one year, 4 percent (two of 45) of low-scoring patients progressed, 8 percent (seven of 91) of moderate-scoring patients, and 24 percent (12 of 50) of high-scoring patients progressed. At two years, the Vectra scores were less predictive of non-progression with 16 percent (five of 31 patients) with low scores progressing, 10 percent (five of 51) with moderate scores progressing, and 18 percent (eight of 45) with high scores progressing.
Crescendo and researchers associated with the company published a comment in A&R in November disputing the findings of the original study. Jeffrey Curtis, professor of rheumatology and immunology at the University of Alabama at Birmingham and first author on the comment, took particular issue with the suggestion that the A&R study results indicated that the Vectra DA test has no relationship to radiographic progression and the biology underlying this process.
"If the test measured nothing biologically that had anything to do with [radiographic] progression, you would be concerned that it's measuring something that frankly could be somewhat irrelevant, at least with respect to that outcome," said Curtis, who consults for Crescendo and is a member of its scientific advisory board. "In fact, that's not the case. There is a significant dose-response relationship. The higher the Vectra score category, the more likely you are to progress."
Even in the highest-risk category, however, only around 20 percent of patients showed radiographic progression, Curtis noted. And this, he suggested, led the study authors to dismiss the test, unreasonably, in his opinion.
"They just said, if all you know about the patient is this one score result, what is the likelihood of radiographic progression?" he said. "Oh, it's only 20 percent? Most people won't progress. That means the test is useless."
However, Vectra DA isn't intended to be used and hasn't been validated as a standalone test for predicting radiographic progression, Curtis said.
"I'm not going to use this score alone to decide who has [radiographic] damage," he said. "I'd really like a [radiographic progression] prediction score, and Vectra DA might be part of that score. But I don't have that [currently]. Neither Crescendo nor anybody else, and certainly not the authors of the [A&R] analysis, have actually tried to do that."
"To say, 'Well, the score itself is not a good single predictor in isolation,' is to me completely ridiculous," Curtis added. "It would be like saying, 'Well, cholesterol lab tests alone are not good predictors of heart attacks. Your blood pressure alone is not a good predictor of a heart attack.' You need to know age, and race, and sex, etc. You need to know a number of other things."
Roy Fleischmann, first author on the A&R study and chief of rheumatology at the University of Texas Southwestern Medical Center, declined to comment for the story, noting that he and his co-authors had said all they felt necessary in the original study and their reply to Crescendo's comment.
While Curtis said he believes the Vectra DA test has been thoroughly validated from a scientific standpoint, he noted that Crescendo has yet to establish the test's clinical utility in specific patient populations and patient types.
"We've established scientific validity, now [Crescendo] needs to prove its clinical utility," he said.
He acknowledged that because the company had not yet demonstrated the test's clinical utility, that might have factored into the recent draft LCD.
"I can certainly understand that opinion," he said of the notion that Medicare might have "jumped the gun by paying for a test whose clinical utility had not been precisely defined for a specific patient population and a given scenario."
However, he said, Medicare covers other tests for management of RA, like c-reactive protein, that similarly lack solid clinical utility and health economic data. "Show me the clinical utility study about why they should pay for a CRP," he said.
Curtis noted that one factor in Medicare's thinking might be the relative cost of the two tests. It reimburses the Vectra DA test at $587, while CRP tests typically cost around $50.
Crescendo is in the middle of clinical utility studies for the test, Curtis noted. Last week the company, in collaboration with a team of Swedish researchers, published a study in A&R that indicated the test could help guide second-line therapy in RA patients who had failed methotrexate treatment. Looking at 157 patients from the Swedish Farmacotherapy (SWEFOT) trial, the researchers found that of the 19 patients with low Vectra scores, 88 percent responded to triple therapy consisting of a combination of methotrexate, sulfasalazine, and hydroxychloroquine, while 18 percent responded to methotrexate plus infliximab. For the 88 patients with high Vectra DA scores, the response rates were 35 percent and 58 percent, respectively.
Aside from the AMPLE study and the lack of clinical utility data, several other factors likely figured into Medicare's decision to issue the draft LCD, said Bruce Quinn, a Medicare expert and principal of healthcare consulting firm Bruce Quinn Associates.
He said the lack of private payor coverage for the test might have been a red flag, particularly given that the test has been commercially available for around six years. He cited a figure presented during Myriad's Q4 2016 earnings report that showed that Medicare accounted for more than 95 percent of the patient lives covered for Vectra DA.
According to Myriad, Medicare-covered tests account for roughly one-third of total Vectra DA volume. The company declined to disclose what portion of Vectra DA tests are covered by private payors. In an investor note discussing the impact of a negative decision on the draft LCD, Piper Jaffray analysts said it would decrease Myriad revenues by $35 million to $40 million.
"I think the degree of private non-coverage is eye catching," Quinn said, noting it suggests that "a number of other insurers have looked at [the test] and written it off and not covered it. I think that is something that the Medicare policy makers pay attention to."
Another consideration cited in the draft LCD is that Vectra DA is not included in the current American College of Rheumatology treatment guidelines. Myriad announced last week that the test has been included in the United Rheumatology clinical practice guidelines. While this organization is not as prominent as ACR, Myriad said it expects the test will be evaluated for inclusion in the ACR guidelines.
Asked to assess the likelihood of a reversal of the draft LCD, Quinn put the odds at 50-50. "I think it could go either way," he said. "I'm sure that Medicare did not issue this without a lot of forethought, but I'm also sure that Myriad will work very hard to present its 20 [previous] studies and critiques as articulately as possible."