Castle Biosciences Release: New Publication Highlights First Prospective Performance Assessment Of The Decisiondx-Melanoma Gene Expression Profile Test

Published: Sep 05, 2017

Multi-center study of 322 patients confirms prognostic accuracy and performance

FRIENDSWOOD, Texas--(BUSINESS WIRE)--Castle Biosciences, Inc., a provider of molecular diagnostics to improve cancer treatment decisions, today announced the publication of the first prospective assessment of prognostic performance for the DecisionDx®-Melanoma test in a multi-center population of patients with cutaneous (skin) melanoma. The gene expression profile (GEP) test uses tumor biology to provide an individual risk of recurrence in cutaneous melanoma patients.

The paper titled, “Interim Analysis of Survival in a Prospective, Multi-center Registry Cohort of Cutaneous Melanoma Tested with a Prognostic 31-Gene Expression Profile Test,” was recently published in the Journal of Hematology & Oncology. In the study of 322 prospectively tested patients, the DecisionDx-Melanoma test demonstrated a robust ability to predict recurrence, distant metastasis, and death at year 3 of an anticipated 5-year prospective, multi-center registry study.

Since two-thirds of all melanoma deaths occur in patients diagnosed with early-stage disease, distinguishing patients with high-risk tumor biology from those patients who are at a lower risk of recurrence is a clinically important goal to drive appropriate management decisions,” commented study co-investigator and senior author Kelly M. McMasters, M.D., Ph.D., Chair of the Department of Surgery, University of Louisville. “The strong association of test results with outcomes across multiple practice settings in this prospective, multi-center study aligns with results from previous studies using archived tumor samples and provides assurance of the test’s prognostic value.”

Of the 322 Stage I-IV patients from 11 dermatologic and surgical centers assessed in this analysis, 88% were Stage I or II and 74% had a sentinel lymph node biopsy (SLNB). Endpoints of recurrence-free survival (RFS), distant metastasis-free survival (DMFS) and overall survival (OS) were assessed using Kaplan-Meier and Cox regression analyses. Median follow-up time for this analysis was 1.5 years for event-free patients.

Key Study Findings

  • For all endpoints assessed, the DecisionDx-Melanoma test was the most sensitive prognostic factor including SLNB.
    • A high-risk (Class 2) result identified 80% of patients who experienced a recurrence compared to 40% for a positive SLNB result.
    • Similarly, the GEP test identified 83% of those who had a distant metastases compared to 50% for a positive SLNB result.
    • The GEP test also identified 73% of those who died compared to 9% for SLNB positive (see table).

Outcomes and correlation with DecisionDx-Melanoma class and SLN status:



Distant metastasis




Class 1 (n=248) 5 (2%) 2 (0.4%) 3 (2%)
Class 2 (n=74) 20 (27%) 10 (14%) 8 (11%)


<0.0001 <0.0001 0.0001
SLN negative (n=286) 15 (5%) 6 (2%) 10 (3%)
SLN positive (n=36) 10 (28%) 6 (17%) 1 (3%)
p-value <0.0001 <0.001 1
  • Consistent with prior studies, the DecisionDx-Melanoma test was significantly associated with recurrence, distant metastasis, and death due to melanoma (p<0.001 for all). Among patients with Stage I or II disease (n=282), the GEP test was also significantly associated with each endpoint analyzed.
  • Significantly different RFS was observed for patients with a Class 1 result (97%) compared to those with a Class 2 result (77%; p<0.0001). Similarly, significantly different DMFS (99% versus 87%; p<0.0001) and OS (99% versus 92%; p<0.0001) were observed among all patients in the analysis.
  • Multivariate analysis demonstrated that DecisionDx-Melanoma class is an independent predictor of RFS outcome, along with Breslow thickness and mitotic rate (p<0.01).
  • High negative predictive values (NPV) of 98% for recurrence and 99% for distant metastasis demonstrate that a Class 1 outcome was associated with a low risk of metastasis during the study.
  • Results from this interim prospective analysis are comparable to recently reported results from Castle’s cumulative retrospective cohort (n=782). While the median non-event follow-up time of 1.5 years in this prospective analysis is shorter than that reported in the cumulative retrospective studies (6.9 years), comparable 1.5-year RFS rates were observed in both studies (97% and 95% for Class 1; 77% and 67% for Class 2).

The results from this prospective, multicenter study further support the robust body of evidence demonstrating that the DecisionDx-Melanoma test provides consistent, clinically meaningful results,” noted study co-investigator Eddy C. Hsueh, M.D., Professor and Director, Division of Surgical Oncology, St. Louis University Hospital. “These prospective results show that the GEP test for melanoma, in combination with standard clinicopathologic factors, can strengthen risk determination and improve patient management.”

The publication is open access and may be accessed at:

About DecisionDx-Melanoma

The DecisionDx-Melanoma test uses tumor biology to provide a prediction of individual risk of melanoma recurrence beyond traditional factors. Using tissue from the primary melanoma, the test measures the expression of 31 genes. The test has been validated in three multicenter studies that have included 690 patients and have demonstrated consistent results. Performance has also been confirmed in three independent, prospective studies including 510 patients. The consistent high performance and accuracy demonstrated in these studies, which combined have included 1200 patients, provides confidence in disease management plans that incorporate DecisionDx-Melanoma test results. Clinical impact has been demonstrated in a multicenter and single-center study showing that test results add additional patient-specific prognostic information to complement traditional staging tools. More information about the test and disease can be found at

About Castle Biosciences

Castle Biosciences is a molecular diagnostics company dedicated to helping patients and their physicians make the best possible treatment and follow-up care decisions based on the individual molecular signature of their tumor. The Company currently offers tests for patients with cutaneous melanoma (DecisionDx®-Melanoma; and uveal melanoma (DecisionDx®-UM and DecisionDx®-PRAME;, with development programs in other underserved cancers. Castle Biosciences is based in Friendswood, TX (Houston), and has laboratory operations in Phoenix, AZ. More information can be found at

DecisionDx-UM, DecisionDx-Melanoma and DecisionDx-PRAME are the trademarks of Castle Biosciences, Inc. Any other trademarks are the property of their respective owners.

Castle Biosciences, Inc.
Derek Maetzold, 866-788-9007
President and CEO

Back to news