Melanoma stage as influenced by 31-GEP risk prediction.

e21604 Background: 31-gene expression profiling (31-GEP) has demonstrated the ability to predict the risk of recurrence (ROR) of cutaneous melanoma. The 31-GEP assay identifies high-risk patients (Class 2A/2B) who have recurrence rates higher than what is expected based on AJCC stage alone. This study details use of 31-GEP testing in a single center to identify patients at high ROR to guide risk-aligned management to improve patient outcomes. Methods: Patients with cutaneous melanoma who underwent 31-GEP testing on the primary tumor and were found to have Class 2A/2B results (n = 20 stage I, 52 stage II, and 28 stage III) were included in this study. They were offered follow-up quarterly physical examination and biannual cross-sectional imaging with yearly brain MRI and were followed by a single physician longitudinally. Oncology consultation for immunotherapy was routine. IRB approval was granted for review of patient outcomes. Results: Median patient age was 64 (range: 32-84); more than 50% of patients had more than 5 years follow-up. Median tumor depth was 2.9mm (0.4-12.0mm). Seventy-three patients were alive and 17 died from melanoma. The entire study population had a 41% recurrence rate. Seventy-one percent (22/31) of patients with metastatic recurrence were salvaged with immunotherapy. Two patients died of treatment complications and one of intracranial bleeding. All 7 patients with metastases not treated with immunotherapy died from their disease. Twenty percent (4/20) of stage I patients recurred, with 2 deaths and one patient alive with disease (AWD). Thirty-eight percent (20/52) of stage II patients recurred and 14/20 were disease-free (NED) or AWD after multidisciplinary treatment for a salvage rate of 70%. All 4 patients who received adjuvant immunotherapy for stage II, class 2B lesions were alive at time of last follow-up. Thirteen percent (7/52) died of metastatic disease. Twenty-eight patients had stage III disease. Forty-six percent (13/28) of patients recurred. Twenty-nine percent (n = 8) died of metastatic disease and 15 were NED. More patients who had regular clinical exams and imaging were NED at follow-up than those who refused imaging (55/73 vs 12/17, respectively). Conclusions: Patients with elevated 31-GEP melanoma have substantial ROR and death regardless of AJCC stage. Even "low risk, node-negative" patients had substantial ROR when found to have high 31-GEP scores. Nearly 75% of patients with recurrence were NED if followed closely and treated with immunotherapy. 31-GEP should augment AJCC staging to improve patient outcomes.