Cutaneous malignant melanoma in West Yorkshire: I. A prospective study of variables, survival and prognosis.

A prospective study was made of 150 cases of primary invasive cutaneous malignant melanoma in clinical stage I diagnosed during the period 1966-1980. Thirty-six of the patients were male, the remaining 114 female, and thus the age-standardized male:female ratio was 2.9:1. One hundred and forty cases were part of the first author's personal series while the data for the remaining 10 patients were provided by colleagues and were subject to the same prospective approach. As a preliminary to multivariate analysis 19 clinical and pathological variables were subject to contingency table analysis to determine significant associations between pairs of variables. Sixty-six nominally significant associations were found of which 28 were highly significant (P less than or equal to 0.0001). Survival to 3, 5, and 7 years was examined by the life table method and a better survival was found in females than in males. Linear logistic regression analyses with dependent variables of 3, 5, and 7 years were carried out by 2 modifications of Cox's regression model, that with survival to 5 years as the dependent variable showing the best goodness of fit. In this study "level of microinvasion" and "patient's sex" emerged as the primary dominant variables in the 5-year regression model. Possible reasons for this and other apparent anomalies between different Cox's models are discussed.