31-Gene Expression Profile Test Improves Risk Assessment in Cutaneous Melanoma

Eddy C. Hsueh, MD,

Combining the 31-gene expression profile (GEP) test with current American Joint Committee on Cancer (AJCC) staging guidelines may improve risk assessment in patients with cutaneous melanoma, as well as determine which patients could be candidates for more intense follow-up, according to data from a multicenter, prospective cohort study published in JCO Precision Oncology.¹

The 31-GEP test proved to be significant for 3-year recurrence-free survival (RFS), distant metastasis-free survival (DMFS), and overall survival (OS) in a univariate analysis, as well as a significant independent predictor of RFS, DMFS, and OS in a multivariable analysis.

Additionally, GEP class 2 results, which indicated high risk for metastases, was significantly associated with lower 3-year RFS, DMFS, and OS in all patients and specifically those with stage I to IIA cutaneous melanoma. Moreover, patients with class 2 results who had stage I to IIA cutaneous disease had rates of recurrence, distant metastasis, and death that were similar to those observed in patients with stage IIB to III disease. Combining 31-GEP test results with current AJCC staging guidelines increased sensitivity vs each approach alone.

“The 31-GEP is an independent predictor of disease progression that adds prognostic information to clinical and pathologic staging criteria,” lead study author Eddy C. Hsueh, MD, of the Department of Surgery at St. Louis University, and colleagues, wrote. “This study confirms the clinical validity of the 31-GEP test in patients with stage I to IIA cutaneous melanoma with class 2 GEP results who may benefit from more intense follow-up.”

Although approximately 70% of patients with cutaneous melanoma are diagnosed with stage I to II disease, this population accounts for the most deaths in patients with melanoma who do not have distant metastases. Traditional staging factors have proven insufficient in identifying which patients are at risk for metastases and death. As such, improved prognostic indicators are needed to reduce treatment in patients who are less likely to progress and escalate treatment in those who have a higher likelihood of progression.

To evaluate tumor biology–based risk in patient with cutaneous melanoma, a 31-GEP test was developed. The test examines 28 discriminating genes and 3 control genes from formalin-fixed, paraffin-embedded tissue to stratify patient risk. Patients with low risk for metastases are identified as class 1; those who are class 1A are at lowest risk, while those who are class 1B are low risk. Those with high risk for metastases are identified as class 2; class 2A patients are at increased risk, while class 2B patients are considered to be highest risk.

While the efficacy of the 31-GEP test as a significant predictor of disease progression has been demonstrated in several studies, its use is not recommended under current staging systems and guidelines. In this study, investigators sought to evaluate the prognostic value of the 31-GEP test in patients with stage I to III cutaneous melanoma and examine whether patients with stage I to IIA disease and a 31-GEP class 2 result have metastatic risk that is similar to those with stage IIB to III disease.

Between 2013 and 2019, 372 patients were enrolled and evaluated on 2 multicenter registry studies, EXPAND (NCT02355587) and INTEGRATE (NCT02355574); a total of 323 patients met the eligibility criteria for this combined analysis. Patients needed to be at least 16 years of age, have stage I to III cutaneous melanoma, had successfully undergone 31-GEP testing, and have no prior history of melanoma.

The primary end points of this combined analysis were 3-year RFS, DMFS, and OS, which were evaluated using Kaplan-Meier analysis with log-rank test.

Patients enrolled had a median age of 58 years (range, 18-87), the majority were male (53%), and the median follow-up was 3.2 years (range, 0.1-7.0). The median follow-up in patients without an event proved to be similar between those classified as class 1 and class 2 , at 3.2 years (range, 0.1-7.0) and 3.4 years (range, 0.4-4.6), respectively (P = .209). Those who did experience an event and were classified as class 1 had a median follow-up of 3.3 years (range, 1.1-4.8), while the median follow-up was 3.1 years (range, 0.4-5.0) for those in class 2.

Tumors determined to be class 2 were associated with high-risk clinicopathologic factors compared with tumors identified to be class 1. Patients with class 2 results were older (P = .006), and a larger proportion were male (P = .011). Additionally, patients with class 2 tumors had a significantly increased median Breslow thickness (P < .001), presence of ulceration (P < .001), and proportion of positive sentinel lymph nodes (SLN; P <.001).

Of all patients at diagnosis, 79% (n = 256) had stage I to IIA disease, 76% (n = 246) had T1 to T2 melanomas, and 81% (n = 263) did not exhibit ulceration. Additionally, the majority, or 85% (n = 211/247), were SLN negative, which suggests that most patients in this cohort would be thought to be low risk and to have good prognosis based on traditional clinicopathologic risk factors.

To identify the utility of the 31-GEP test as a predictor of risk of progression in patients with cutaneous melanoma, investigators compared the 3-year RFS, DMFS, and OS in patients with class 1 disease (n = 252) with data from those with class 2 disease (n = 71).

Results indicated that patients with class 2 tumors had a significantly lower 3-year RFS compared with those who had class 1 tumors, at 66% (95% CI, 56-78) and 95% (95% CI, 92-98), respectively (P <.0001). Additionally, DFMS and OS were also found to be significantly lower in patients with class 2 disease vs those with class 1 disease, at 79% (95% CI, 70-90) vs 97% (95% CI, 94-99), respectively, and 81% (95% CI, 72-91) vs 97% (95% CI, 95-100), respectively.

When looking at subclasses, patients who were identified to be class 2B, or highest risk, were found to have a significantly lower 3-year RFS than those with class A disease, who were the lowest-risk subgroup, at 60% vs 96%, respectively (P <.0001); this was also true in terms of DMFS, at 78% vs 98%, respectively (P <.0001) and OS, at 74% vs 98%, respectively (P <.0001). Patients who were classified as class 1B or 2A were noted to have intermediate survival rates.

Twenty-five deaths were reported in this cohort. In 9 of the patients who died, distant metastasis was reported; however, melanoma was not determined to be the cause of death. Moreover, 7 of the 9 patients who died (78%) had class 2 disease; 6 of the 7 patients were categorized as class 2B.

“In a univariate analysis including class 2 (v class 1) 31- GEP results, continuous age, male (v female) sex, head and neck (v non–head and neck) tumor location, continuous mitotic rate, and high-risk stage IIB to III (v stage I to IIA) AJCC 8th edition staging, only GEP class 2 and high-risk AJCC stages reached a threshold of P <.01 for all 3 end points; whereas, mitotic rate was significant for DMFS, and age was significant for OS,” the study authors wrote.

Additionally, subclass analysis of the test showed that classes 1B, 2A, and 2B were significant predictors of risk of recurrence. Classes 2A and 2B were noted to be significant predictors of distant metastasis, while class 2B alone was reported to be a significant predictor of mortality.

Results from the multivariate analysis indicated that 31-GEP class 2 (HR, 4.34 [95% CI, 2.10-8.96], P <.001) and AJCC stage IIB to III (HR, 2.98 [95% CI, 1.48-6.02], P = .002) were significant, independent predictors of RFS, while GEP class 2 (HR, 5.45 [95% CI, 2.09-14.25], P <.001), AJCC stage IIB to III (HR, 2.81 [95% CI, 1.16-6.58], P = .023), and mitotic rate (HR, 1.04 [95% CI, 1.01-1.07], P = .007) were significant predictors of 3-year DFMS. Moreover, 31-GEP class 2 (HR, 3.13 [95% CI, 1.23-7.96], P = .016), AJCC stage II to III (HR, 3.89 [95% CI, 1.60-9.50], P = .003), and age (HR, 1.08 [95% CI, 1.04-1.13], P <.001) were significant predictors for OS.

Investigators also showed that class 2 results yielded by the 31-GEP test were more sensitive with regard to detecting recurrence, distant metastasis, and death compared with AJCC staging alone. 31-GEP class 1 results were noted to have a high negative predictive value for 3-year RFS, DMFS, and OS, at 94%, 97%, and 97%, respectively; this suggests favorable survival outcomes in patients who have a class 1 result.

Moreover, when identifying patients at high risk through the 31-GEP class 2 result or the AJCC high-risk category, enhanced sensitivity was noted in terms of 3-year RFS, DMFS, and OS, at 76%, 88%, and 76%, respectively, vs AJCC staging alone, which had sensitivities of 57%, 62%, and 60%, respectively, or the 31-GEP test alone, with sensitivities of 64%, 69%, and 68%, respectively.

Specifically, 31-GEP class 2 results identified AJCC stage I to IIA patients who were at increased risk of recurrence, distant metastasis, and death with sensitivity of 44%, 70%, and 40%, respectively. A 31-GEP class 1 result was noted to confirm a low risk of recurrence, distant metastasis, and death in these patients, with an negative predictive value of 92%, 98%, and 95%, respectively.

“These data provide a rationale for using the 31-GEP test in conjunction with AJCC staging to obtain an optimal prognosis for patients with cutaneous melanoma,” study authors concluded.

Reference

1. Hsueh E, DeBloom J, Lee J, et al. Long-term outcomes in a multicenter, prospective cohort evaluating the prognostic 31-gene expression profile for cutaneous melanoma. JCO Precis Oncol. Published online April 6, 2021. doi:10.1200/PO.20.00119