Competing-risks model to predict prognostic factors for cause-specific mortality in patients with skin verrucous carcinoma based on the SEER database

Abstract Background: Applying a competing-risks analysis to data from the Surveillance, Epidemiology, and End Results (SEER) database, we aimed to identify significant prognostic factors and evaluate the cumulative incidence of cause-specific (CS) death for skin verrucous carcinoma (SVC). The Cox proportional-hazards model, extensively employed in assessing survival trends and identifying prognostic indicators, has the potential to generate erroneous predictions. However, in the realm of clinical practice, there is still a lack of specific prognostic factors for cutaneous verrucous carcinoma, leading to disproportionate treatment. The insights derived from this analysis can serve as valuable guidance for clinical interventions Methods: The SEER database provided relevant data of patients with SVC. The reliability, precision, and logicality of estimations for cumulative incidence function (CIF) related to CS mortality and death from other causes at each time point were enhanced through the utilization of competing-risks analysis. In the univariate analyses, Gray's test and the CIF were used, while the multivariate analysis employed the Cox proportional-hazards model, CS, and the Fine-Gray model. Results: The study involved 656 eligible patients with SVC, with 332 deaths recorded: 115 attributed to SVC and 217 resulting from other causes. Univariate analyses revealed that variables such as differentiation grade, marital status, metastasis, American Joint Committee on Cancer (AJCC) stage, age, surgery, the status of radiotherapy, and chemotherapy significantly influenced the cumulative incidence rates for events of interest (P<0.05). Marital status, AJCC stage, race, age, and surgery status, emerged as independent risk factors in the multivariate Cox regression. Based on the multivariate Fine-Gray and CS model analysis, age, AJCC stage, differentiation grade, and surgery independently served as key determinants affecting the risk of specific outcomes in SVC patients (P<0.05). Conclusions: The novel competing-risks model increased the accuracy of predictions by examining the cumulative incidence rate of cancer-specific mortality. Moreover, this approach is high useful in research for obtaining data such as the prognostic variables for SVC.