Castle Biosciences (NASDAQ:CSTL) executives and clinical collaborators used a DecisionDx-Melanoma webcast to review results from the multicenter DECIDE study, which the company said was recently published in Future Oncology. The discussion focused on how the company’s 31-gene expression profile (31-GEP) test is intended to help clinicians refine two key questions in early-stage cutaneous melanoma: a patient’s risk of sentinel lymph node (SLN) positivity and a patient’s risk of recurrence.
Clinical decisions the company says the test is built to inform
Matthew Goldberg, MD, Senior Vice President of Medical at Castle Biosciences, framed sentinel lymph node biopsy (SLNB) and recurrence risk as the primary decision points that drive management intensity after melanoma diagnosis. He said lower-risk patients typically receive wide local excision, less intensive follow-up, and no advanced imaging, often remaining in a dermatology-led pathway. Higher-risk patients may be directed toward SLNB, closer follow-up, multidisciplinary oncology involvement, and potentially baseline or surveillance imaging.
- i31-SLNB, an individualized estimate of SLN positivity risk
- i31-ROR, an individualized estimate of recurrence risk
He emphasized that SLN positivity and recurrence are “related, but they’re not the same clinical question,” and said the test’s models incorporate different clinicopathologic factors accordingly. For i31-SLNB, he listed ulceration, Breslow thickness, age, and mitotic rate alongside the gene expression profile score. For i31-ROR, he said inputs include ulceration, age, Breslow thickness, mitotic rate, SLN status, and tumor location in addition to the gene expression profile score.
Positioning DECIDE within prior evidence and real-world data
Goldberg argued that DECIDE should be viewed as an additional “prospective, multi-center” contribution to an existing evidence base that includes retrospective and prospective studies as well as real-world evidence. He also highlighted Castle’s collaboration with the National Cancer Institute and the SEER Cancer Registry to evaluate outcomes in clinically tested patients, stating that the 31-GEP test stratified mortality risk within AJCC sub-stage groups and showed prognostic value for melanoma-specific survival beyond clinicopathologic factors.
During the webcast, Goldberg also said the company’s analysis found that clinically tested patients showed a 32% lower three-year melanoma-specific mortality rate than untested patients, and he described DecisionDx-Melanoma as the only melanoma GEP test shown to be associated with improved survival in that context.
DECIDE study design and key endpoints
Michael Guenther, MD, a surgical oncologist and lead author on the DECIDE trial, described the study as a prospective clinical utility effort assessing whether i31-SLNB can inform SLNB decisions in routine practice. Goldberg said the study’s primary objectives were to (1) prospectively confirm performance of i31-SLNB in predicting SLN positivity, (2) evaluate real-world use of i31-SLNB in guiding SLNB decisions, and (3) assess recurrence outcomes among patients predicted to have <5% risk of SLN positivity, including those who did not undergo SLNB.
Guenther said DECIDE followed an approach in which clinicians ordered the test in adults diagnosed within two months who were considering SLNB, then made a shared decision with the patient on whether to proceed. Outcomes tracked included SLN positivity, recurrence-free survival, distant metastasis-free survival, and melanoma-specific outcomes in those with <5% predicted nodal metastasis risk.
DECIDE enrollment and performance highlights discussed on the webcast
Guenther reported the study included 912 patients, with a median age of 65 (range 20–90). He said the median Breslow depth was 0.8 mm (range 0.1–12 mm) and noted that 42% of cases had a transected base, which he said reflects real-world shave biopsy patterns and was “reassuring” for clinicians.
Other study metrics he cited included:
- 47% underwent SLNB; among those, 10.2% were SLN-positive
- 52.9% did not undergo SLNB; Guenther said nearly 90% of these were T1a, T1b, or T2a primaries
- Ulceration was present in 9% of cases
- Risk distribution by i31-SLNB prediction: 52% <5%, 32.2% in the 5%–10% “discuss and consider” range, and 15.8% >10%
Guenther highlighted results in Stage IB (T1b/T2a) patients, calling that group a key area of clinical debate. He said the SLN positivity rate was 1.4% in those with low predicted nodal risk versus 18.5% in those with high predicted nodal risk, describing that as a 13.2-fold difference. Across all T1–T4 patients, he reported SLN positivity of 2.6% for <5% predicted risk, 7% for 5%–10%, and 21.4% for >10%.
Guenther also discussed false-negative metrics he attributed to pooled results and the DECIDE dataset. He reported a 2.6% false-negative rate for the <5% group among “all comers” with a 37-to-1 true-negative to false-negative ratio, and in the T1/T2a subgroup he reported a 1.8% false-negative rate with a 55-to-1 ratio. He contrasted these results with what he described as prior presentations regarding CP-GEP assays.
On recurrence outcomes, Guenther said patients with low-risk lesions who did not have SLNB had 97.8% three-year recurrence-free survival, and he said high-risk lesions had “nearly 7%” lower recurrence-free survival at three years. He also described broader recurrence stratification by gene expression score distribution, stating that lower-risk patients had 96% recurrence-free survival at five years, while higher-risk lesions had 61% five-year relapse-free survival.
Q&A: NCCN, adoption, and how clinicians use results
In the question-and-answer session, analysts asked about incremental takeaways versus prior publications and whether DECIDE addresses perceived evidence gaps for guideline inclusion. Goldberg and Guenther characterized DECIDE as consistent with prior retrospective performance data while adding prospective multicenter validation and clinical utility around the <5% and >10% thresholds used in SLNB decision-making.
Company executive Derek (last name not provided on the webcast) described how the sales team is using the study to communicate three messages: peers are using the test in practice, the test consistently identifies patients below the 5% threshold, and patients who avoid SLNB based on low-risk predictions “do extremely well.” He also discussed the competitive landscape, referencing other tests and stating that one U.K. firm’s test is no longer marketed in the U.S. and that another competitor has limited U.S. market share based on the company’s review of claims data.
Guenther also described real-world conversations with patients, saying most patients want prognostic information and that low-risk results can provide meaningful reassurance. He emphasized his view that identifying biologically aggressive disease among clinically “shallow” lesions is important, and provided a case example from his practice to illustrate how a high-risk DecisionDx-Melanoma result influenced management.
About Castle Biosciences (NASDAQ:CSTL)
Castle Biosciences, Inc is a molecular diagnostics company specializing in the development and commercialization of prognostic and diagnostic tests for patients with dermatologic conditions. The company’s proprietary portfolio of genomic assays is designed to improve risk assessment and guide clinical decision-making for individuals with skin cancers and other skin-related diseases. By combining genomic data with advanced statistical algorithms, Castle Biosciences seeks to provide actionable insights that help physicians tailor treatment plans and monitoring strategies.
The company’s flagship test, DecisionDx-Melanoma, evaluates the probability of metastasis in patients diagnosed with cutaneous melanoma, supporting more personalized surveillance and therapeutic approaches.
