Abstract Background Melanoma is diagnosed as either in-situ or invasive disease. The relationship between the two diseases is unclear. If every in-situ is an early stage of invasive melanoma, diagnosis and removal of in-situ lesions should reduce the incidence of invasive melanoma. If the association is more complex, the excision of in-situ lesions might not effectively prevent invasive disease and may represent overdiagnosis. Methods A population-based cohort study involved all patients diagnosed with either in-situ or invasive melanoma from the New Zealand Cancer Registry between 2001 and 2017. The pattern of in-situ and invasive melanoma was compared by incidence, trends, and key patient characteristics (age at diagnosis, sex, body site, and ethnicity). Results The incidence of in-situ melanoma increased annually by 3.77% whereas that of invasive melanoma was relatively stable (annual increase 0.04%) over the study period. The pattern of in-situ and invasive melanoma was similar by sex and ethnicity but differed by body site. Since the distribution of melanoma for age at diagnosis was highly influenced by body site and sex, it was difficult to compare between the two diseases. The observed risk of invasive melanoma among in-situ cohort was four times higher than that expected among general population. Conclusions Not every in-situ was a precursor of invasive melanoma, but some did progress to an invasive lesion. Key-messages Plans should be considered to compare the potential harms and benefits of the screening and excision of in-situ melanoma.