Development and validation of a nomogram for predicting suicide risk factors in thyroid cancer patients following diagnosis: a population-based retrospective study

Objectives To develop and validate a user-oriented nomogram of suicide risk in thyroid cancer patients to enable clinicians to identify and intervene in a timely manner with high-risk subgroups. Methods This was a retrospective, population-based cohort study in which patients with thyroid cancer diagnosed from the Surveillance, Epidemiology, and End Results (SEER) database between 2000 and 2020 were include. Optimized features were screened by the least absolute shrinkage and selection operator (LASSO) regression model. Subsequently, we selected variables with nonzero coefficients, entered them into a Cox proportional hazards regression model and constructed a visualized nomogram model predicting suicide. We implemented receiver operating characteristic curves (ROC), calibration curves, decision curve analysis (DCA), and internal validation to assess the discrimination, calibration, clinical applicability, and generalizability of the nomogram. Results To our knowledge, this is the first nomogram specifically designed for thyroid cancer patients, integrating histopathological, therapeutic, and socioeconomic predictors. Furthermore, the calibration curves for this nomogram fit well with the diagonal, and the C-indexes for the training and testing sets were 0.760 and 0.724, respectively, and the decision curve analysis indicated clinical benefit. Conclusion This study successfully identified risk factors for suicide in patients with thyroid cancer and developed a nomogram that provides patients with an individualized, quantifiable assessment of suicide risk and assists clinicians in identifying and intervening in potential suicides.