The 31-gene expression profile for identification of real-world node-negative patients with stage I-IIA cutaneous melanoma who have a higher risk of death: A SEER collaboration.

Abstract

e21593 Background: Most patients with stage I-IIA (tumor stage T1-T3a) cutaneous melanoma (CM) have good overall outcomes, and current guidelines do not recommend increased surveillance or systemic therapies for these patients with early-stage disease. However, a significant subset (~16%) of sentinel lymph node negative patients who are ineligible for advanced imaging experience metastasis and/or melanoma-specific death. In collaboration with the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program, we analyzed whether the 31-gene expression profile (31-GEP) identifies early-stage patients at increased risk of melanoma-specific mortality (MSM) or all-cause mortality (ACM) who may benefit most from more risk-aligned disease management plans. Methods: We analyzed data from node-negative patients with stage I-IIA CM who were clinically tested with the 31-GEP (2013-2019) and linked to the SEER registry database (n = 8,896). Five-year melanoma-specific survival (MSS) and overall survival (OS) were estimated using Kaplan-Meier analysis and compared between groups using the log-rank test. Multivariable Cox regression analysis was used to evaluate significant predictors of MSM and ACM (*p < 0.05, **p < 0.001). Results: The 31-GEP significantly stratified risk of death in node-negative patients with stage I-IIA CM (p < 0.001). The 5-year MSS for a Class 1A result was 98.8% compared with 96.1% for Class 1B/2A and 93.4% for Class 2B results. The 31-GEP also significantly stratified 5-year OS (p < 0.001) for Class 1A (93.5%), Class 1B/2A (88.7%), and Class 2B (81.1%). In multivariable analysis, 31-GEP Class 1B/2A (HR = 2.34*), Class 2B (HR = 2.94*), and age (HR = 1.04**) were significant predictors of MSM. The 31-GEP Class 2B (HR = 1.68*), Breslow thickness (HR = 1.23*), and age (HR = 1.09**) were significant predictors of ACM. Conclusions: In a large, real-world cohort of patients with early-stage, node-negative CM, the 31-GEP identified patients at increased risk of death, including melanoma-specific death. These patients with lower-stage disease and a high-risk 31-GEP result may benefit from more intensive surveillance and follow-up schedules. No variable was more predictive of death from melanoma than Class 2B results, even Breslow thickness. Critically, these findings suggest that a reconsideration of current indications for adjuvant therapies, including the highest risk 31-GEP signature, may be warranted.