Factors associated with a high tumour thickness in patients with melanoma

Prognosis of patients with melanoma is strongly associated with tumour thickness at time of diagnosis. Therefore, knowledge of patient characteristics and behaviour associated with a high tumour thickness is essential for the development and improvement of melanoma prevention campaigns.

The present study aimed to identify sociodemographic, clinical and behavioural factors associated with high tumour thickness according to Breslow.

The study population consisted of 217 patients with histologically proven primary invasive cutaneous melanomas seen at the Department of Dermatology and Allergology at the Ludwig-Maximilian-University Munich, Germany, between January 1999 and January 2001. Personal interviews were conducted by two physicians to obtain information on sociodemographic characteristics and on patients' knowledge of melanoma symptoms, sun behaviour, delay in diagnosis and related factors. Multivariate linear and logistic regression analysis with stepwise variable selection was used to identify risk groups with a high tumour thickness. To assess possible effect modifications, interaction terms were included in the regression analysis.

The median tumour thickness was 0.8 mm (interquartile range 0.5-1.6). Fifty-seven patients (26%) had tumour thickness >1.5 mm. In a multivariate linear regression analysis, patients living alone and patients with a low educational level showed a significantly greater tumour thickness. The relation of melanoma knowledge to tumour thickness was modified by the melanoma subtype: whereas lack of melanoma knowledge led to an increased tumour thickness for the subtypes superficial spreading melanoma, lentigo maligna melanoma and unspecified malignant melanoma, no significant effect was estimated for the subtypes nodular melanoma (NM) and acrolentiginous melanoma (ALM). Sex, age, self-detection of melanoma, patient delay and professional delay were not significantly associated with the tumour thickness in multivariate linear regression. Similar results were found in multivariate logistic regression.

An increased tumour thickness was found in subjects living alone and having a low educational level. These subjects should be targeted in future prevention campaigns in a more focused way. Further efforts are necessary to improve knowledge and earlier detection of melanoma subtypes NM and ALM.