NEW YORK – New data stemming from a collaboration between Castle Biosciences and the US National Cancer Institute could bolster the case for gene-expression profiling to guide treatment decisions for early-stage melanoma patients.
In a study published in JCO Precision Oncology in late June, researchers, including Castle Biosciences data scientist Christine Bailey and scientists from the NCI's Surveillance, Epidemiology, and End Results (SEER) program, tapped into a large, real-world cohort in the SEER database to assess if doctors' use of Castle's 31-gene expression profiling test, DecisionDx-Melanoma, improved patients' survival outcomes.
For the study, the researchers considered 3,258 patients with stages I to III cutaneous melanoma who'd undergone testing with the DecisionDx-Melanoma test between 2016 and 2018 and 9,773 patients who did not receive this testing. Bailey and coauthors considered patients who'd been diagnosed with melanoma after 2016 "to generate a contemporary population with potential access to effective US Food and Drug Administration-approved therapy options." Notably, immune checkpoint inhibitors became commercially available in this setting that year.
The data on patients who underwent gene-expression profiling came from Castle Biosciences, and the researchers worked with a third-party firm, Information Management Services, to link the biomarker data to other medical and outcomes information in 17 SEER registries. They then turned to the SEER database again to create a cohort of 9,773 patients who had not been profiled using DecisionDx-Melanoma but were similar to tested patients in terms of melanoma stage, age, sex, race, and socioeconomic and other factors.
The real-time PCR 31-gene expression profiling test categorizes patients into one of three risk profiles: patients who have Class 1A tumors are at lowest risk, Class 1A/2A tumors are at increased risk, and Class 2B tumors have the highest risk of disease recurrence or metastasis within five years.
In the JCO Precision Oncology study, researchers looked into survival outcomes and found that patients who'd gotten the DecisionDx-Melanoma test had improved survival outcomes versus patients who did not. Patients who had gotten the test had a 29 percent lower risk of dying from melanoma within three years and a 17 percent lower risk of dying from any cause within three years, compared to those who didn't get tested, the researchers concluded.
Within the tested cohort, 96.6 percent of patients with Class 1A cancers were alive after three years, as were 90.2 percent of patients with Class 1B/2A tumors and 79.4 percent of patients with Class 2B tumors. Collectively, 5.2 percent of patients who received testing had died after three years, versus 6.2 percent of patients who had not.
"Validation of DecisionDx-Melanoma in a large, unselected and representative population provides greater confidence in the test's reliability," said Robert Cook, Castle's senior VP of R&D. "In addition, the demonstrated association with a survival benefit suggests that management changes based on the additional risk information could provide long-term value."
A key limitation of the study, according to Bailey and colleagues, was that the SEER data don't include the specific treatment regimens patients received after testing. Accordingly, they can't say for certain whether the DecisionDx-Melanoma test's results regarding patients' risk of dying from melanoma was the main factor behind how aggressively oncologists treated them. Still, the authors believe that treatment decisions following the test likely played a significant role in the improved survival.
For instance, Cook shared the example of patients with early-stage melanoma with a high-risk GEP score. These patients, Cook said, might receive more frequent imaging, which has been shown to guide effective immunotherapy post-surgery. Patients with early-stage melanoma and a low-risk GEP score, on the other hand, might be able to avoid a sentinel lymph node biopsy to determine whether their melanoma has metastasized beyond the tumor site.
For those patients with high-risk scores on DecisionDx-Melanoma, "implementing more intense management strategies … can improve outcomes," wrote Bailey and coauthors, highlighting recent research finding that physicians using the test for clinical management changed their treatment plans 50 percent of the time after receiving the results.
Although in the JCO Precision Oncology study researchers weren't able to consider whether DecisionDx-Melanoma predicted patients' outcomes on specific treatments, past studies have hinted at the test's potential utility in this regard. "The SEER database doesn't include specific treatments [but] other research efforts have studied how gene-expression profiling testing informs changes in decision-making that lead to improved clinical outcomes," said Cook.
For example, in a paper published in the journal Archives of Dermatological Research this past March, researchers from Northwestern University and elsewhere showed that melanoma patients without lymph node involvement who received routine imaging based on their high-risk DecisionDx-Melanoma scores had improved outcomes versus patients who didn't undergo testing. Their recurrences were detected an average of 10 months earlier, and more of these tested patients went on to receive immunotherapy.
After surgical resection, melanoma patients can receive checkpoint inhibitors such as Merck's Keytruda (pembrolizumab) or Bristol Myers Squibb's Opdivo (nivolumab) and Yervoy (ipilimumab). In order to receive the BMS drugs, patients' cancers must have spread to the lymph nodes, while stage IIB through III melanoma patients can receive Keytruda. DecisionDx-Melanoma can help with treatment decision-making, Cook noted, since patients who are eligible for these immunotherapies but have a low risk of metastasis according to the gene-expression test could potentially do without these drugs, which comes with clinical and financial toxicities.
"Physicians who are managing early-stage patients with a high-risk gene-expression profiling score may consider management options that will identify metastatic disease earlier, when treatments are more effective," Cook said.
As of now, patients with stages I to III cutaneous melanoma are eligible to receive the DecisionDx-Melanoma test, and any healthcare professional, including a dermatologist, physician assistant, nurse practitioner, or oncologist can order the test, said Cook.
The firm declined to share the list price of the test, but Cook said it is working with government and commercial payors to secure reimbursement. Medicare contractor Novitas recently listed DecisionDx-Melanoma among several tests that it said it wouldn't cover because they were not medically necessary, but implementation of that local coverage determination has since been halted, and Novitas has said it will publish a new draft LCD and seek public input on its coverage policy.
According to a Castle Biosciences spokesperson, the firm is "pleased with Novitas' decision to delay implementation of the LCD." Castle declined to comment on whether it would be sharing these new data with Novitas, citing the fact that the Medicare contractor's revised position on DecisionDx-Melanoma is not yet public.
Castle Biosciences has a supply deal in place with the US Veterans Health Administration to provide DecisionDx-Melanoma to veterans treated through the VHA as well as active-duty service members and their families treated through the Military Health System.