Individualized Predictive Nomograms for Overall Survival and Cancer-Specific Survival in Pediatric Medulloblastoma Patients: A Surveillance, Epidemiology and End Results (SEER) Database Analysis

Introduction: Medulloblastoma is the most common malignant brain tumor in children with an annual incidence of 0.7 per 100,000 children. The 5-year overall survival rate varies between 40% and 89%. We aim to develop and validate a nomogram that assists clinicians in visualizing overall survival (OS) and cancer-specific survival (CSS) based on various clinical and demographic factors.

Methods: We extracted data for patients diagnosed with medulloblastoma between 2000 and 2021 from the SEER database. Our study focused on 1842 pediatric cases, aged 0 to 18 years. The dataset was randomly split into training (n=1289) and validation (n=553) cohorts using a 7:3 ratio. We employed multivariate Cox regression on the training set to identify prognostic factors for OS and CSS. The predictive performance of the resulting nomograms was evaluated using C-indices, and area under the curve (AUC).

Results: Overall, patients had a mean ± SD age of 7.04 ± 4.68 years, and 63.4% were males. Median OS and CSS were not reached in the cohort. Multivariate analysis revealed that chemotherapy, radiation, and age were independent predictors of CSS, while OS was influenced by these same factors plus type of surgery. Children aged 0-2 years had significantly lower CSS and OS rates compared to those aged 3-18 years (log-rank test, P < 0.0001). In the training cohort, the C-indices for OS and CSS nomograms were 0.707 (95%: 0.599-0.791) and 0.713 (95% CI: 0.665-0.755). In the validation cohort, the C-indices for OS and CSS nomograms were 0.703 (95%: 0.659-0.7511) and 0.709 (95% CI: 0.607-0.796). The AUC values for CSS at 6, 12, 24, and 36 months are 0.877, 0.759, 0.680, and 0.631, while for OS, they are 0.877, 0.751, 0.686, and 0.637, respectively.

Conclusion : This nomogram effectively predicts survival in pediatric medulloblastoma, with strong validation, aiding clinicians in personalized risk assessment and informed treatment decisions.