Prognostic factors in localized invasive primary cutaneous malignant melanoma: results of a large population-based study.

Abstract

The prognostic impact of several histopathological prognostic features in cutaneous malignant melanoma (CMM) remains controversial. To assess the independent prognostic value of mitotic rate, regression, tumour-infiltrating lymphocytes (TILs) and growth phase in primary stage I and II CMMs. Clinicohistopathological data were obtained from the Stockholm-Gotland registry for 4237 patients diagnosed with an incident primary stage I or II CMM followed up to December 2011. The risk of CMM-specific death was evaluated by a Cox regression model. A mitotic rate of 1-10 mitoses per mm(2) [hazard ratio (HR) 1·69, 95% confidence interval (CI) 1·16-2·45] and > 10 mitoses per mm(2) (HR 2·27, 95% CI 1·46-3·52) were significant; TILs and regression were not. A more detailed analysis of data assessed between 1989 and 1995 confirmed significantly increased HRs for the presence vs. absence of mitoses (HR1-5/mm² 2·25, 95% CI 1·36-3·76; HR6-10/mm² 2·34, 95% CI 1·23-4·44; HR> 10/mm² 2·64, 95% CI 1·39-4·99). Other prognosticators were increasing T-stage vs. T1, presence of ulceration and presence of vertical growth phase (VGP). In T1 CMMs, an increasing tumour thickness vs. < 0·7 mm (HR0·7-0·8 mm 2·24, 95% CI 1·24-4·04; HR>0·8 mm  2·92, 95% CI 1·57-5·43) and presence of ulceration were significantly associated with higher HRs; mitotic rate, TILs, regression and growth phase were not. Determinants of increased risk of CMM death in stage I and II CMMs were increasing T-stage, presence of ulceration, presence of mitoses and VGP. This was not found for TILs or regression.