Incidence trends and predictive model of hepatic malignant tumors in children: a population-based study

Objective: To analyze the incidence trend and establish a model to predict the prognosis of hepatic malignant tumors in children (CHMTs). Methods: We analyzed the incidence data of CHMTs from 1975 to 2018 from the Surveillance, Epidemiology, and End Results (SEER) database, and evaluated the incidence trends based on different demographic and pathological features. We also analyzed clinicopathologic data from 2000 to 2018 from the SEER database. Univariate and multivariate Cox regression analyses were performed to explore prognostic factors related to overall survival (OS). Then, we established nomograms based on independent predictors and verified them using receiver operating characteristic curves, calibration plots, and decision curve analysis plots. Results: The incidence of CHMTs increased significantly, from 0.1 per 100,000 in 1975 to 0.4 per 100,000 in 2018. Incidences among different races and genders were increasing and converging. The incidence of hepatoblastoma (HB) increased, while that of hepatocellular carcinoma (HCC) was relatively stable. The 1-, 3-, 5-, and 10-year OS rates were 86.2%, 77.5%, 74.2%, and 70.2%, respectively. Being Spanish-Hispanic-Latino, HB, surgery, and systemic therapy were independent predictors of longer OS, whereas regional and distant stages were independent predictors of shorter OS. Nomograms with good predictive ability and clinical utility were established to evaluate the prognosis of children with HB or HCC. Conclusion: The incidence of CHMTs is increasing, especially for HB and in younger children. This study identified independent predictors and developed nomograms that could provide a personalized and accurate prognosis for CHMTs.