| Abstract: |
Background: Tumor ulceration (TU) is considered the second most important prognostic factor after Breslow thickness for localized cutaneous malignant melanoma (CMM). However, many studies have not included mitotic rate (MR) with TU in these analyses. When both TU and MR are included in the same analysis, MR appears to be the more important than TU and TU loses its significance as an independent prognostic factor. Methods: The relative importance of TU and MR as prognostic factors in localized CMM were compared in a population-based series of 650 consecutive invasive CMM cases ascertained from the Connecticut tumor registry and reviewed by a single dermatopathologist (RLB), during the period between January 15, 1987 and May 15, 1989. Seventeen clinical and histopathological variables including tumor thickness measured in mm, TU recorded as present or absent, and MR recorded as number per mm2 were included in an unconditional logistic regression model and selected for inclusion using a backward stepwise algorithm with death as an endpoint or at least five-years follow-up. Results: In the multivariate regression, the independent prognostic factors included: 1. tumor thickness in millimeters (OR = 1.5, 95% CI = 1.3-1.9) 2. moderate mitotic index (between 1 and 6): (OR = 8.3, 95% CI 2.4-28.7), 3. mitotic index (>6): (OR = 11.6, 95% CI = 3.0-44.6), 4. solar elastosis: (inversely associated with mortality) (OR = 0.4, 95% CI = 0.2-8). After adjustment for MR, TU lost its significance. When MR was left out of the analysis, ulceration then became an independent prognostic factor. The model with ulceration only (excluding MR) showed a relative risk (RR) of 2.4 (95% CI: 1.1-5.1). In the model with MR only, MR had a RR of 14.5 (95% CIS.9-53.7). Finally, regression analysis including both TU and MR yielded an RR of 11.6 for MR and 1.7 for TU. Conclusions: Our results suggest that MR as a proxy for tumor proliferation is a more important prognostic factor than TU. © Blackwell Munksgaard 2005. |
