Clinical correlates of Breslow thickness of malignant melanoma.

Abstract

Breslow thickness is a major predictor of prognosis in cutaneous malignant melanoma (MM) and patients continue to present with thick lesions, which have a poorer prognosis. To investigate correlations of Breslow thickness with demographic variables, tumour site, clinical features, false negative diagnostic rate and clinic of primary referral. Data were obtained from the records of 738 patients with histologically diagnosed cutaneous MM. Tumours included were in situ and invasive MM and lentigo maligna melanoma. In view of the skewed distribution of MM thickness, categories of MM thickness were used for analysis, using the commonly applied cut-offs of 0.75, 1.5 and 3.5 mm. The variables investigated were particularly associated with changes in the proportion of the thickest group, 'thick' MMs. The proportion of this thickness category is proposed as an appropriate and sensitive variable for the investigation of correlations with MM thickness. Thickness >/= 1.5 mm has also been considered in view of its prognostic significance. Results were similar for the two thickness groups, but more significant for the thick group. The previously described correlations of tumour thickness and increasing age (P < 0.00001) and location on head and neck (P = 0.0002), together with the independence of these variables, have been confirmed. The correlation with male gender was also confirmed but this was weak (P = 0.05). Novel findings were correlations of Breslow thickness with all features of the seven-point checklist (P varying from P = 0.01 to P < 0.00001) and tumour elevation (P < 0.00001). In general in the seven-point checklist, the absence of the major features (except variation in size) (P < 0.00001) and presence of minor features (except altered sensation) (P varying from P = 0.004 to P < 0.00001) were strongly associated with thick MMs. These results for tumour thickness are a further argument for retention of the minor features of the seven-point checklist. False negative diagnosis was found to be correlated with tumour thickness (P < 0.02) but this relationship was lost on multivariate analysis with inclusion of the clinical features. MM thickness was highly correlated with primary referral clinic (P < 0.00001). Only approximately 8% of MMs presenting to the Pigmented Lesion Clinic (PLC) were thick, while the proportion presenting to general dermatology was about three times greater and to plastic surgery about five times greater. This was maintained on multivariate analysis, including all other variables and, therefore, there must be other determining factors of referral not examined in the study. Conclusion MM thickness is associated with increasing age, male gender, location on the head and neck, all features of the seven-point checklist, tumour elevation and referral to the PLC. It is important to investigate further the reasons for general practitioner referral of different thickness MM to different types of clinic, as referral to clinics other than the PLC results in delay in the first hospital appointment, and it is now apparent that thicker lesions are disproportionately affected.