31-GEP May Refine Melanoma Risk Stratification and Surveillance

By Amara Osei

Thirty-one genes can add prognostic information to AJCC staging in stage I and stage II cutaneous melanoma, but current findings do not establish that testing itself improves survival. Gaurav Singh, MD, a Mohs surgeon at City of Hope Cancer Center in Chicago, discussed DecisionDx-Melanoma, a 31-gene expression profile (31-GEP) assay, in a Dermatology Times expert discussion.

What does AJCC staging leave unresolved?

American Joint Committee on Cancer (AJCC) staging remains central and incorporates established clinical and pathological factors, including Breslow thickness (tumor depth), ulceration, and nodal status. Singh cited tumor location, mitotic activity, and immunosuppression as additional influences that may contribute to differences in melanoma behavior.

Patients with stage I melanoma can still recur or metastasize despite standard-of-care treatment, while some patients with more advanced disease achieve favorable outcomes after surgery and immunotherapy. The concrete comparison is scope: 31-GEP measures expression across 31 genes, whereas AJCC staging integrates tumor depth, surface ulceration, and lymph-node status.

GEP therefore adds information about potentially aggressive tumor biology rather than replacing established staging variables. Singh characterized it as an additional source of prognostic information that conventional factors may miss.

What can a high-risk result change?

Singh uses DecisionDx-Melanoma to support risk stratification in stage I and stage II cutaneous melanoma. For stage I, a high-risk 31-GEP result may prompt additional consideration of sentinel lymph node biopsy (SLNB), sampling of the first draining lymph node(s), or imaging when conventional characteristics would not otherwise have led to either intervention.

That distinction matters because the discussion frames a high-risk result as a prompt to reconsider management, not an automatic surgical mandate. For stage II, Singh uses high-risk results, including Class 2A and 2B, to inform additional imaging and closer surveillance.

He also considers those results when sentinel-node findings are negative, because 31-GEP may identify aggressive biology that conventional staging misses. Clinical utility may be lower in established stage III melanoma with confirmed lymph-node involvement, depending on the specific management question.

What does the evidence establish?

The evidence cited spans population-based, prospective, and consensus analyses. Together, these studies provide the evidentiary framework Singh invoked, although the discussion did not reproduce their numerical results.

• A 2023 population-based Surveillance, Epidemiology, and End Results (SEER) collaboration in JCO Precision Oncology examined 31-GEP and survival outcomes.
• A 2025 prospective, multicenter DECIDE analysis in Cancer Medicine examined recurrence-free survival after SLNB decisions influenced by 31-GEP.
• A 2024 systematic review and meta-analysis assessed the prognostic value of 31-GEP; a separate 2024 review examined clinical utility and performance in a Mayo Clinic cohort.
• A 2025 Society of Surgical Oncology consensus statement assessed the evidence for and utility of GEP in primary cutaneous melanoma.

Singh said integrating molecular risk information may "help identify patients who could benefit from more intensive monitoring." He set a firm boundary: prognostic findings must not be interpreted as showing that 31-GEP-guided surveillance improves survival or displaces guideline-based management.

His formulation was that GEP should "supplement, rather than replace" established clinical assessments. The discussion reported no sample sizes, recurrence rates, hazard ratios, or confidence intervals, so the citations alone cannot establish a magnitude of clinical benefit.

How should a result influence individual care?

Singh cautioned against interpreting GEP independently of the patient’s overall clinical circumstances. Decisions about imaging, surgery, and surveillance should combine estimated risk with functional status, comorbidities, and patient preferences.

A low-risk result may provide reassurance even when it does not directly alter treatment. A high-risk result may justify closer observation, but it does not make disease progression inevitable.

For patients with confirmed stage III nodal disease, the test may answer fewer practical questions than in early-stage disease. Its value then depends on what management decision remains unresolved.

Can ctDNA extend surveillance?\n

A potential complement is circulating tumor DNA (ctDNA), which Singh described as a means of monitoring for molecular evidence of melanoma over time. GEP characterizes a tumor’s underlying prognostic profile; serial ctDNA testing tracks changes in circulating tumor DNA levels.

In his practice, Singh considers serial ctDNA testing for patients with high-risk GEP results and uses changes in ctDNA levels to help inform subsequent imaging decisions. He emphasized that ctDNA remains under investigation in melanoma, while combining it with GEP could support more individualized surveillance.

As evidence develops, integration will depend on demonstrated clinical utility, guideline recommendations, and individual patient needs, not risk estimates alone. The forward-looking question is whether GEP- and ctDNA-informed surveillance improves decisions and outcomes, rather than merely refining prognostic assessment.