NCCN Melanoma Guidelines Include Risk Assay to Inform SLNB Decision

The National Comprehensive Cancer Network (NCCN) has incorporated the Merlin clinicopathologic–gene expression profile (CP-GEP) assay into the latest version of its guidelines for cutaneous melanoma.^1,2

The recommendation focuses on patients with T1b and T2a cutaneous melanoma, a group in which the absolute risk of nodal metastasis is often intermediate and the decision to pursue SLNB can be clinically nuanced.

The updated guidelines cite prospective data from the MERLIN_001 trial (NCT04759781) that support the assay’s use in selected patients with early-stage disease to inform sentinel lymph node biopsy (SLNB) decision-making. The data published in JAMA Surgery in 2025 showed that patients categorized as high risk by the CP-GEP model had approximately a threefold higher likelihood of SLN positivity compared with those categorized as low risk.³

The update marks the first time a molecular assay has been specifically referenced as potentially appropriate for SLNB risk stratification outside of a clinical trial setting in this population.²

Trial Overview: MERLIN_001

MERLIN_001 was a prospective, multicenter, blinded validation study evaluating the ability of a CP-GEP model to predict SLN status in patients with primary cutaneous melanoma undergoing SLNB. The assay integrates clinicopathologic variables such as Breslow thickness and patient age with tumor gene expression data into a single binary output of high vs low risk.

The study enrolled patients with clinically node-negative melanoma who met standard criteria for SLNB. The primary end point was the association between CP-GEP classification and SLN positivity.

In 1761 patients who received CP-GEP testing and had a SLNB, 651 were considered low risk. SLNB was found to be positive in 7.1% in low risk and 23.8% in high risk.³ As the T category increased, the number of patients who were low risk by CP-GEP declined, with only 2.8% of patients with low risk being staged with T3 disease. In those with low-risk stage IB disease, SLN was positive in 6.5% of those with low risk and 18.3% with high risk. In patients aged 65 or older, 6.6% who had low risk and 20.3% who had high risk had positive SLN.

The authors reported that the model demonstrated clinically meaningful discrimination in T1b and T2 tumors, where SLNB decisions are often individualized.

The study was designed to address limitations of earlier retrospective GEP studies by incorporating prospective enrollment and blinded assessment of outcomes.² However, performance characteristics including sensitivity, specificity, and negative predictive value must be interpreted in the context of SLN positivity prevalence in the studied cohort.

Clinical Context: SLNB Decision-Making in Early-Stage Melanoma

SLNB remains a key staging procedure in clinically node-negative melanoma and provides prognostic information that can guide adjuvant therapy discussions. Current NCCN guidelines recommend discussing SLNB in patients with T1b lesions (≥0.8 mm thickness or ulcerated) and routinely recommending it in most T2 or higher tumors.¹

However, SLNB carries procedural risks, including seroma, infection, and lymphedema, and does not confer a melanoma-specific survival advantage in all subgroups.⁴ As a result, there has been sustained interest in tools that more precisely estimate nodal risk to avoid unnecessary procedures in low-risk patients while identifying those at higher risk.

The guidelines now state “CP-GEP testing may be used in select patients with T1b ± T2a melanoma to support shared decision-making, when a patient or provider would decide against SLNB if the true risk was < 10%.”

Prior NCCN guidance has generally advised against routine use of GEP testing to guide SLNB decisions outside of clinical trials due to limited prospective validation.¹ The 2026 update distinguishes the CP-GEP model evaluated in MERLIN_001 from other commercially available assays, noting that alternative GEP tests for SLNB risk prediction are not recommended outside of a clinical trial based on currently available evidence.

Assay Background and Regulatory Status

The Merlin CP-GEP test combines clinicopathologic features of age and tumor thickness with expression of 8 melanoma-associated genes to generate a binary metastatic risk classification.³ The model was developed through collaboration between investigators at the Mayo Clinic and SkylineDx and is commercially available in the United States as a laboratory-developed test.

Interpretation and Limitations

Although prospective validation represents a methodological strength, MERLIN_001 was not designed to test whether CP-GEP–guided decision-making improves patient-centered outcomes such as survival, quality of life, or cost-effectiveness.³ The study evaluated predictive performance for SLN status rather than downstream clinical utility.

The NCCN guideline indicates that the test may be considered in select patients as part of shared decision-making.¹ As with any risk stratification tool, results should be interpreted in conjunction with established clinicopathologic factors and patient preferences.

Longer-term outcome data from MERLIN_001 and other cohorts may further clarify whether molecular risk refinement translates into durable changes in management patterns or outcomes.

REFERENCES

1. NCCN. Clinical Practice Guidelines in Oncology. Cutaneous melanoma, version 1.2026. Accessed February 20, 2026. https://tinyurl.com/5b39f96p

2. Merlin CP-GEP Becomes the First and Only Gene Expression Profile (GEP) Test Recommended by NCCN® Cutaneous Melanoma Guidelines to Assess Metastatic Risk. News release. SkylineDX. February 19, 2026. Accessed February 20, 2026.

3. Hieken TJ, Egger ME, Angeles CV, et al. Gene expression profile–based test to predict melanoma sentinel node status: the MERLIN_001 study. JAMA Surg. 2025;160;(12):1358-1366. doi:10.1001/jamasurg.2025.4399

4. Morton DL, Thompson JF, Cochran AJ, et al. Final trial report of sentinel-node biopsy versus nodal observation in melanoma. N Engl J Med. 2014;370(7):599-609. doi:10.1056/NEJMoa1310460