Background/aim: People diagnosed with thick melanomas have poorer survival than those diagnosed with thin melanomas, prompting efforts to diagnose melanomas early. The incidence of in situ and thin melanomas has risen in recent decades, likely through early detection efforts, but so has the incidence of thick melanomas. The rise in thick melanomas could reflect biology (i.e., aggressive tumours that arise rapidly) or health service use (i.e., limited access or late presentation). To address this question, we compared people who presented with melanomas of different thickness with those who did not develop melanoma.
Methods: We examined associations between sociodemographic, phenotypic, genetic, and health service factors and melanoma thickness in the QSkin cohort (n=38,660). We compared people who did not develop melanoma with those who had a first diagnosis of in situ (T0), thin invasive (T1), and thick invasive (T2-T4) melanoma. We used cause-specific Cox proportional hazards models and compared hazard ratios across outcome groups. Covariates were selected a priori using directed acyclic graphs.
Results: During a median 12.4 years of follow-up, 1773 people developed an incident in situ melanoma, and 833 an invasive melanoma (633 T1, 200 T2-T4). T2-T4 melanomas were more likely to be nodular and to occur on the lower limbs compared with T0 or T1 melanomas. Older age, red hair colour, higher nevus density, and atypical nevi were associated with increased risk across all thickness categories, with stronger associations for thick melanoma. Higher socio-economic advantage was inversely associated with thick melanoma, but positively associated with T0 and T1 melanoma. A history of ≥2 doctor skin checks was associated with increased risk across all thickness categories, with stronger associations for T0 and T1 melanoma. Genetic risk score for melanoma and measures of healthcare utilisation (e.g., GP visits, biopsies, excisions) were not differentially associated with melanoma thickness at diagnosis.
Conclusions: We found reasonably strong evidence that biology might explain thick melanomas. Associations with health service factors were complex, with some evidence that thin melanomas were associated with greater levels of access.