BACKGROUND: The American Joint Committee on Cancer eighth edition substaging might be suboptimal for predicting melanoma progression. Using it to select stage II patients for adjuvant immunotherapy risks overtreating low-risk stage IIB/IIC patients and undertreating high-risk stage IIA patients. Prognostic capability of tumor-infiltrating lymphocytes (TILs) is unclear in stage II melanoma. OBJECTIVE: To evaluate the American Joint Committee on Cancer eighth edition substaging and TIL scoring as predictors of progression in stage II melanoma. METHODS: Retrospective cohort study of 366 sentinel lymph node negative stage II melanoma patients from 4 UK hospitals (2004-2017), with long-term follow-up. RESULTS: Twenty-three percent of melanomas progressed (median 9.5-year follow-up). Among those, 41.5% were stage IIA, 41.5% IIB, and 17.1% IIC. TIL scoring independently predicted progression risk (brisk vs non-brisk: odds ratio: 0.298, P = .009; absent vs non-brisk: odds ratio: 0.436, P = .049) and progression-free survival. Nonbrisk TILs, present in 80% of progressing tumors, denoted high risk. TIL scoring split patients into high and low risk across substages: stage IIA patients with non-brisk TILs had similar 5-year progression-free survival to stage IIB/IIC patients with absent/brisk TILs. LIMITATIONS: Retrospective study design and unknown generalizability. CONCLUSION: Stage II melanoma progression is poorly predicted by the American Joint Committee on Cancer eighth edition substage. TIL scoring offers improved risk stratification across substages and could serve as a cost-effective method to better identify patients who may benefit from adjuvant immunotherapies.