Prognostic nomogram for early-onset vulvar cancer: a SEER-based study

The incidence of vulvar cancer has been increasing, particularly among younger women, yet prognostic tools specifically designed for early-onset vulvar cancer (EOVC) remain limited. Using the Surveillance, Epidemiology, and End Results (SEER) database, we identified 1386 patients with EOVC and developed prognostic models for overall survival (OS) and cancer-specific survival (CSS). Independent prognostic factors were identified using multivariable Cox regression, and nomograms were subsequently constructed and compared with five machine learning (ML) survival models, including random survival forest, gradient boosting machine, CoxBoost, LASSO-Cox, and survival support vector machine. Model performance was evaluated using the concordance index (C-index), time-dependent area under the curve (AUC), Brier score, calibration slope and intercept, and decision curve analysis (DCA). Age, marital status, histological type, tumor differentiation, lymph node status, SEER stage, primary tumor site, surgery, radiotherapy, and chemotherapy were independently associated with prognosis. The Cox-based nomograms demonstrated favorable discrimination, calibration, and greater net benefit than TNM staging. Although RSF and GBM showed competitive discriminative ability, no statistically significant improvement over the Cox-based nomogram was observed. In validation analyses, most ML models demonstrated reduced generalizability and less stable calibration, whereas the Cox-based model maintained more consistent predictive performance with lower Brier scores. In conclusion, we developed and validated clinically applicable nomograms for individualized survival prediction in EOVC. Compared with more complex ML approaches, the Cox-based nomogram showed robust performance, stability, and interpretability, supporting its potential clinical utility.