Relationships between survival and real-world recurrence-free survival or distant metastasis-free survival among patients with completely resected stage IIB or IIC melanoma.

Long follow-up time is needed for overall survival (OS) data to mature for early-stage melanoma. This retrospective study aimed to describe the relationships between OS and two intermediate endpoints - real-world recurrence-free survival (rwRFS) and real-world distant metastasis-free survival (rwDMFS) - for patients with stage IIB or IIC melanoma that was completely resected from 1 January 2008 to 31 December 2017, with follow-up to 31 December 2020. We used three different approaches to describe the relationships: estimates of correlation using Kendall τ rank correlation; comparisons of all-cause survival with/without recurrence or distant metastasis using adjusted Cox proportional hazard models; and landmark analyses of all-cause survival stratified by recurrence status at 1-5 years. During a 39-month median follow-up from surgical resection, 223/567 patients (39%) experienced recurrence, among whom 171/567 patients (30%) developed distant metastasis. Median OS from surgical resection was 117.6 months [95% confidence interval (CI), 104.7-not reached], median rwRFS was 49.8 months (95% CI, 39.6-61.0), and median rwDMFS was 70.9 months (95% CI, 58.4-89.1). We observed strong correlations between rwRFS and OS, and between rwDMFS and OS (Kendall τ of 0.73 and 0.82, respectively). Risk of death was significantly greater after recurrence (all-cause survival adjusted hazard ratio [HR], 7.48; 95% CI, 4.55-12.29) or distant metastasis (adjusted HR, 11.00; 95% CI, 6.92-17.49). Risk of death remained significantly elevated with recurrence or distant metastasis by landmark years 1, 3, and 5 after surgical resection. These findings support the use of recurrence/rwRFS and distant metastasis/rwDMFS as surrogate endpoints for OS after complete resection of stage IIB or IIC melanoma.