为了确保黑色素瘤患者的最佳治疗和监测，需要了解美国联合委员会第八版（AJCC8）亚分期的临床分期特定复发风险、死亡率和复发模式。为了估计阶段特定的复发率和黑色素瘤特异性死亡率，评估阶段特定的复发和死亡风险，并描述阶段特定的复发模式，包括条件比率。回顾性队列研究，利用前瞻性收集的全国人群登记数据。全国范围的人群队列研究。纳入了从2008年1月1日至2019年12月31日期间首次诊断为IA至IV期皮肤黑色素瘤的丹麦患者，年龄大于或等于18岁，并从主要治疗开始直至2021年12月31日进行随访。首次诊断的IA至IV期皮肤黑色素瘤。分期特定的复发累积发生率和黑色素瘤特异性死亡率，黑色素瘤特异性无复发生存率和评估阶段特定的复发和黑色素瘤特异性死亡风险。次要结果是阶段特定的复发模式，包括条件比率，以及黑色素瘤特异性生存率。我们对25,720名患者进行了中位随访5.9年（95% CI, 58.9-59.3年）。平均年龄为59.1岁（95% CI, 58.9-59.3岁）。IIB至IIC期黑色素瘤患者年龄较大，在诊断时合并疾病较多，并且通过前哨淋巴结活检的病理分期比例最低（81.6%-87.4%）。总共有10.6%的患者发生复发；首次复发中，56.6%的患者出现远处复发，单独或同时伴有局部区域复发。我们发现，IIIA和IIB阶段的复发风险相当（29.7% vs 33.2%），IIIB和IIC阶段的复发风险也相当（35.9% vs 36.8%）。黑色素瘤特异性死亡率在IIIA和IIA阶段（13.0% vs 13.6%）以及IIIB和IIB阶段（18.4% vs 22.0%）之间相当。这些风险模式在因果特异危险模型中持续存在。这项全国范围的人群队列研究发现，当前AJCC8分期系统的不断增加阶段并不能准确反映黑色素瘤的复发和死亡风险增加。远程复发的高比例表明血液转移是一种比以前认为更常见的转移途径，应考虑在II级B至IV期进行例行功能性/断层成像进行监测。未来的工作应致力于开发新的风险分层工具，并确定例行成像监测对生存的影响。

To ensure optimal treatment and surveillance of patients with melanoma, knowledge of the clinical stage-specific risk of recurrence, mortality, and recurrence patterns across the American Joint Committee on Cancer Eighth Edition (AJCC8) substages is needed.To estimate stage-specific recurrence and melanoma-specific mortality rates, assess absolute stage-specific risks of recurrence and mortality, and describe stage-specific recurrence patterns, including conditional rates.Retrospective cohort study of prospectively collected nationwide population-based registry data.Nationwide, population-based cohort study.The 25 720 Danish patients, 18 years or older, diagnosed with first-time stage IA to IV cutaneous melanoma between January 1, 2008, and December 31, 2019, were included and followed up from time of primary treatment until December 31, 2021.First diagnosis of stage IA to IV cutaneous melanoma.Stage-specific cumulative incidence of recurrence and melanoma-specific mortality, melanoma-specific recurrence-free survival, and assessed absolute stage-specific risks of recurrence and melanoma-specific mortality. Secondary outcomes were stage-specific recurrence patterns, including conditional rates, and melanoma-specific survival.We followed up 25 720 patients for a median of 5.9 years (95% CI, 58.9-59.3 years). Mean age was 59.1 years (95% CI, 58.9-59.3 years). Patients with stage IIB to IIC melanoma were older, had more comorbidities at diagnosis, and had the lowest rate of pathologic staging by sentinel node biopsy (81.6%-87.4%). A total of 10.6% of patients developed recurrence; first recurrence included distant recurrence, alone or with synchronous locoregional recurrence, in 56.6% of patients. We found a comparable risk of recurrence in stages IIIA and IIB (29.7% vs 33.2%) and in stages IIIB and IIC (35.9% vs 36.8%), respectively. Melanoma-specific mortality was comparable between stages IIIA and IIA (13.0% vs 13.6%) and between stages IIIB and IIB (18.4% vs 22.0%), respectively. These risk patterns persisted in cause-specific hazards models.This nationwide, population-based cohort study found that the increasing stages of the current AJCC8 staging system do not accurately reflect an increasing risk of recurrence and mortality in melanoma. The high proportion of distant recurrences suggests that hematogenous spread is a more common metastatic pathway than previously assumed, and surveillance with routine functional/cross-sectional imaging should be considered for stages IIB to IV. Future efforts should be put toward developing new tools for risk stratification and determining the survival effect of routine imaging in surveillance.