摘要Background:Prostate cancer(PCa)patients are at risk of developing second primary malignancies(SPMs),which can significantly shorten their survival.Understanding the risk of SPMs and associated factors is crucial to the optimization of patient follow-up.Methods:This study focuses on PCa patients who were later diagnosed with SPMs using data from the Surveillance,Epidemiology,and End Results(SEER)database.Variables were carefully selected,and the data were analyzed using machine learning techniques combined with mul-tivariate Cox proportional hazards modeling.Subsequently,a nomogram was generated to predict the 1-,3-,and 5-year survival rates for SPMs patients.Additionally,a two-sample Mendelian randomization(TSMR)analysis was conducted to investigate the causal relationships between PCa and its top ten SPMs.Results:Among the variables,age,marital status,SPM site,M stage,American Joint Committee on Cancer(AJCC)stage,PCa surgery,and prostate-specific antigen(PSA)levels were identified as key prognostic factors through least absolute shrinkage and selection operator(LASSO)and backward stepwise regression.Based on these factors,a nomogram was developed to visually represent survival predictions,complemented by a web-based calculator for easy application.This nomogram,which serves as a supplement to traditional AJCC staging,demonstrated strong predictive power for 1-,3-,and 5-year survival,with area under the curve(AUC)values exceeding 0.85.Additionally,TSMR analysis revealed a causal link between PCa and urothelial carcinoma(UC).Conclusion:This study developed a nomogram for predicting survival in prostate cancer patients with secondary primary malignancies,enhancing prognosis accuracy.TSMR identified a causal link between PCa and UC.