Background: The prognostic significance of regression in melanoma is debated. We present a large multicenter study correlating regression with sentinel lymph node metastasis and melanoma-specific survival. Methods: The Sentinel Lymph Node Working Group database was reviewed from 1993 to 2018. Patients with known regression and sentinel lymph node status were included. Clinicopathologic factors were correlated with regression, sentinel lymph node status, and melanoma-specific survival. Results: There were 4,790 patients; median follow-up was 39.6 months. Regression was present in 1,081 (22.6%) cases, and 798 (16.7%) patients had sentinel lymph node metastases. On multivariable analysis, male sex, truncal tumors, and decreasing thickness were significantly associated with regression (P < .05), whereas head/neck or leg tumors had lower rates of regression (P < .05). Regression was significantly correlated with a decreased risk of sentinel lymph node disease on multivariable analysis (odds ratio 0.68, 95% confidence interval 0.54–0.85; P = .0008). Multivariable analysis also showed that increasing age, male sex, increasing thickness, ulceration, lymphovascular invasion, microsatellitosis, and sentinel lymph node metastasis were significantly (P < .05) associated with worse melanoma-specific survival, while regression was significantly associated with better melanoma-specific survival (hazard ratio 0.75, 95% confidence interval 0.57–0.99; P = .043). Conclusion: This large study shows that regression is significantly associated with better outcomes in patients with melanoma and is correlated with a lower risk of sentinel lymph node metastasis and a better melanoma-specific survival.
ASJC Scopus subject areas