| Autor: |
Verkouteren JAC; Department of Dermatology, Erasmus MC University Medical Center, Rotterdam, The Netherlands., Smedinga H; Department of Public Health, Erasmus MC University Medical Center, Rotterdam, The Netherlands., Steyerberg EW; Department of Public Health, Erasmus MC University Medical Center, Rotterdam, The Netherlands., Hofman A; Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, The Netherlands., Nijsten T; Department of Dermatology, Erasmus MC University Medical Center, Rotterdam, The Netherlands. Electronic address: t.nijsten@erasmusmc.nl. |
| Abstrakt: |
A third of basal cell carcinoma (BCC) patients will develop subsequent BCCs. We aimed to develop a simple model to predict the absolute risk of a second BCC. We observed 14,628 participants of Northern European ancestry from a prospective population-based cohort study. BCCs were identified using a linkage with the Dutch Pathology Registry (Pathological Anatomy National Automated Archive). Predictors for a second BCC included 13 phenotypic, lifestyle, and tumor-specific characteristics. The prediction model was based on the Fine and Gray regression model to account for the competing risk of death from other causes. Among 1,077 participants with at least one BCC, 293 developed a second BCC at a median of 3 years. Several well-known risk factors for a first BCC were not prognostic for a second BCC, whereas having more than one initial BCC was the strongest predictor. Discriminative ability at 3 years was reasonable (bootstrap validated c-index=0.65). Three groups were created, with 7, 12, and 28% risk of a second BCC within 3 years. We conclude that a combination of readily available clinical characteristics can reasonably identify patients at high risk of a second BCC. External validation and extension with stronger predictors is desirable to further improve risk prediction. |
