NEW YORK – Castle Biosciences said this week that it is making progress on several research fronts, including building evidence for the survival benefit afforded to patients by its cutaneous melanoma test, DecisionDx-Melanoma.
The firm is also pushing forward its newest product in the dermatological space, a test for inflammatory skin conditions, presenting its first proof-of-concept data at a scientific meeting last month.
During a call this week discussing the firm's first quarter financial results, Castle CEO Derek Maetzold discussed the company's presentation of results from a recently expanded collaboration with the National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) program, shared in a poster at the European Association of Dermato Oncology Congress in April.
"This expanded real-world data showed improved survival for patients who have the benefit of the DecisionDx-Melanoma test, in addition to traditional clinical and pathological data, compared to untested patients who only had access to traditional [measures] to determine their treatment and follow-up plan," Maetzold said.
Castle and its academic collaborators analyzed records from the SEER database for 3,261 stage I-III melanoma patients tested with DecisionDx-Melanoma, matching them to a group of about 10,000 controls who did not receive the test.
The matching process incorporated 11 clinicopathologic and socioeconomic variables. The untested patients and their clinicians had only the traditional clinicopathologic features upon which to develop a melanoma treatment plan.
Overall, patients whose doctors ordered DecisionDx-Melanoma had improved survival compared to untested patients, with a 27 percent improvement in melanoma-specific survival, and 21 percent better overall survival.
"This suggests that DecisionDx-Melanoma test results can aid in providing more risk-aligned treatment plans for improved patient outcomes," the group wrote.
According to Castle, the results also reflect what it has seen in previously published retrospective and prospective studies — that DecisionDx-Melanoma offers independent risk stratification for both melanoma-specific and overall survival.
In the SEER cohort, a DecisionDx-Melanoma Class 2B result conferred the highest risk of all clinicopathologic factors included in multivariable analyses, including age, ulceration status, Breslow thickness, and sentinel lymph node status.
During the company's call this week, Maetzold said that he believes data like this could prompt new recognition of DecisionDx-Melanoma by guidelines bodies like the National Comprehensive Cancer Network.
"[This] is a very significant real-world prospective data set that shows that when clinicians have an opportunity to incorporate the results of our test with the other standard of care information … they are able to make … decisions that result in their patients living longer, and that's pretty significant [especially] given that it's not a Castle study, but a third-party large NCI study," he said.
The same would probably not be true, though, of other bodies like the American Joint Commission on Cancer, which is more concerned with diagnostic procedures. "It wasn't until the last version [in 2018] when you saw AJCC including even Oncotype Dx for breast cancer in the diagnostic workup guidelines, so I think impacting AJCC diagnostic criteria is a ways off for us still," Maetzold said.
Castle also saw progress recently in its development of a new test for non-cancer skin disorders, initially announced last May.
The assay is intended to harness genomic signals that can predict systemic therapy response in patients with moderate to severe psoriasis, atopic dermatitis, and related conditions.
Researchers presented new proof-of-concept data at the Revolutionizing Atopic Dermatitis 2022 Conference last month, demonstrating that a noninvasive skin scraping technique Castle has adopted for the test produces enough RNA to reproducibly assess gene expression signatures.
"We were looking to find a sample collection method for Castle's test that would be simple for clinicians to use yet would provide enough RNA to adequately examine the expression of genes in a lab setting; and we believe we succeeded," Aaron Farberg, a study investigator and a dermatologist affiliated with Baylor Scott & White Health System in Dallas, said in a statement.
"I am excited by the possibility to better care for my patients, as I have seen how debilitating psoriasis and atopic dermatitis can be," he added.
In the study, Castle and its collaborators obtained skin scraping samples from 20 patients with atopic dermatitis and 20 patients with psoriasis from two dermatology practices — Farberg's in Dallas, and another at Washington, D.C.'s George Washington University School of Medicine.
Investigators used quantitative real-time PCR to evaluate expression of a standard set of genes in scrapings from the 40 patients, concluding that the noninvasive sampling technique produced sufficient RNA to support measurement of Castle's gene expression profile for predicting treatment response.
Maetzold said Castle is on track to see its first development and validation data next year, and to launch the test by the end of 2025. The firm initiated a planned 4,800-patient, prospective study last year. As of last month, the effort has signed on 52 committed sites, with approximately 145 patients enrolled thus far.