Background: Women with stage I malignant melanoma (MM) have a survival advantage over men as judged by univariate analysis. However, on multivariate analysis, gender was found to be an independent predictor of survival in only 8 of 14 published studies.

Objective: This study attempts to explain the disparate findings for gender as a prognostic factor in different multivariate analyses.

Methods: Univariate and multivariate analyses were performed on 832 patients with stage I MM in the New York University Melanoma Cooperative Group (NYU-MCG) data base. The results were compared with those of 14 similar studies.

Results: In the NYU-MCG data base, gender, age of the patient, and number of mitoses per square millimeter were not independent factors on multivariate analysis, whereas thickness, anatomic site, and presence of ulceration were. The statistically significant difference in survival by gender on univariate analysis, in the NYU-MCG data base, could be explained by the differences in thickness and anatomic site of the MMs in the sexes. Comparison of these results with the reviewed reports from the literature consistently shows thickness and ulceration to be independent prognosticators of survival. Likewise, most authors agree that age is not an independent predictor. However, there is no consensus with respect to gender and site, each of which was found to be an independent predictor of survival in only about half the studies reviewed.

Conclusion: The disparate findings for gender in different multivariate analyses are explained by a gender-related difference in anatomic distribution of MM. Gender and site appear to have a similar influence in multivariate analysis and thus either one or the other is a dominant factor in different multivariate analyses.