The statement hints at the Achilles’ heel of GEP in melanoma – that is, the lack of evidence that test results improve outcomes. This was the main concern of the consensus statement; the cart is before the horse.
One of the consensus authors, David Polsky, MD, PhD, professor of dermatologic oncology at New York University, New York City, said that “most of the data for this test come from retrospectively collected patient groups.” The prospective studies have been generally small, with no comparator group. “While they have shown some promise in intermediate thickness melanoma, they have not yet demonstrated utility for thin, stage I melanomas.”
First, do no harm
A new meta-analysis of over 800 patients with cutaneous melanoma tested by DecisionDx-Melanoma, published in JAMA Dermatology alongside the consensus statement, shows how the tests perform.
Among patients with a recurrence, DecisionDx-Melanoma correctly classified 82% with stage II disease but only 29% with stage I disease as high risk. Among those without recurrence, the test correctly classified 90% of stage I patients but only 44% with stage II disease as low risk.
Similar results were seen with the melanoma GEP test available in Europe, MelaGenix (NeraCare GmbH). This test was developed from a panel that was narrowed to seven protective genes and one high-risk gene using a training cohort of 125 cutaneous melanomas.
provided by MSKCC press office
Dr. Michael Marchetti
“The prognostic ability of GEP tests ... appeared to be poor at correctly identifying recurrence in patients with stage I disease, suggesting limited potential for clinical utility in these patients,” commented the meta-analysis authors, led by Michael Marchetti, MD, an assistant professor of dermatology at Weill Cornell Medical College in New York City.
“Unknown are the harms associated with a false-positive result, which were 10-fold more frequent than true-positive results in patients with stage I disease,” they pointed out.
“Further research is needed to define the incremental improvement in risk predictions provided by the test beyond ... all other known clinicopathologic factors,” which include patient sex, age, tumor location and thickness, ulceration, mitotic rate, lymphovascular invasion, microsatellites, and other factors proven to be linked to outcomes, they said.
Studies so far suggesting benefit have incorporated a few of those factors, but not all of them. For now, “it is not clear which patients should be tested or how to act on the results,” Dr. Marchetti and colleagues concluded.
Breast cancer standard of proof
Larger, prospective studies are needed to address whether GEP testing can replace SLNB to predict relapse “and [can identify] patients who could be spared surveillance imaging and/or benefit from adjuvant therapy,” wrote the consensus authors. Follow-up also needs to be long enough to detect delayed recurrence of thin melanomas, they added.
With more research, there is reason to hope that gene expression profiling will help in melanoma; it’s already standard of care in breast cancer, they pointed out.
On the hope front, one cohort study evaluated whether DecisionDx-Melanoma could identify patients at low risk for positive lymph nodes in T1/T2 disease who were eligible for biopsy. Only 1.6% of subjects who were aged 65 years or older and identified by the test as low risk had a positive node.
Dr. Carrie Kovarik
“This is a promising direction of investigation ... in a narrow, defined population,” noted authors led by Carrie Kovarik, MD, associate professor of dermatology at the University of Pennsylvania, Philadelphia, in an opinion piece last spring.
But still, until there’s “clear evidence that [DecisionDx-Melanoma] results affect patient outcomes, we should not use it to influence care decisions in patients with thin” melanomas. Dermatology “should expect the same standards” of proof as breast cancer, they wrote.
What to do right now?
Despite the marketing, “think twice before ordering GEP tests for” T1a melanomas is the message in an editorial that accompanies the consensus statement. The 5- and 10-year melanoma-specific survival rates are 99% and 98%, respectively. GEP tests are unlikely to change these estimates significantly. In fact, the new meta-analysis indicates “that there may be an approximately 12% misassignment rate in this population,” wrote editorialists Warren Chan, of Baylor College of Medicine, Houston and Hensin Tsao, MD, PhD, director of the melanoma genetics program at Massachusetts General Hospital, Boston.
“Even if you use GEP testing and discover a low-risk class assignment for a 2 mm thick melanoma, avoid the urge to bypass the sentinel lymph node discussion. ... Nodal sampling, for good reasons, remains part of all major guidelines and determines eligibility for adjuvant treatments. ... Many of us engaged in genomics research believe that accurate [melanoma] GEP will be developed in time, but better tools and greater tenacity are needed,” they wrote.
There was no industry funding for the consensus statement and meta-analysis. Authors on the consensus statement reported numerous ties to pharmaceutical and other companies, as listed in the paper.
A version of this article originally appeared on Medscape.com.