Clinical features and a prognostic nomogram based on the SEER database for hepatoblastoma, hepatocellular carcinoma, and embryonal sarcoma among children and adolescents

BACKGROUND: Hepatoblastoma (HB), hepatocellular carcinoma (HCC), and embryonal sarcoma (ES) are the three main types of liver tumors in children and adolescents. At present, epidemiological knowledge and predictors of these three liver tumor types in multi-ethnic populations are limited. This study aimed to outline the clinical features and construct a prognostic nomogram for these tumors, which can contribute to the prediction of dynamic overall survival probability during the follow-up period. METHODS: A total of 1,122 patients liver tumor patients between 2000 to 2019 in Surveillance, Epidemiology, and End Results (SEER) database were enrolled for the current study, and separated into 824 HB, 219 HCC, and 79 ES according to the type of pathology. Independent prognostic factors were screened by univariate and multivariate Cox regression analysis, and a prognostic nomogram was constructed for overall survival. The accuracy and discriminative abilities of the nomogram were evaluated by concordance index as well as time-dependent receiver operating characteristic curves and calibration curves. RESULTS: Race (P=0.0016), surgery [hazard ratio (HR): 0.1021, P<0.001], and chemotherapy (HR: 0.27, P=0.00018) are independent prognostic factors for hepatoblastoma. Pathological tissue grading (P=0.00043), tumor node metastasis (TNM) staging (P=0.00061), and surgery are independent prognostic factors for hepatocellular carcinoma. Household income and surgery (HR: 0.1906, P<0.001) are independent prognostic factors for embryonal sarcoma. All of these prognostic factors are significantly associated with prognosis. A nomogram consisting of these variables was established, which showed a good concordance index (0.747, 0.775, and 0.828 in hepatoblastoma, hepatocellular carcinoma, and embryonal sarcoma, respectively). Also, the 5-year area under curve (AUC) of the nomogram were 0.738, 0.812, and 0.839 in hepatoblastoma, hepatocellular carcinoma, and embryonal sarcoma, respectively. In the calibration diagram, an optimal agreement between the nomogram-predicted and actual observed survival was evident. CONCLUSIONS: We developed an effective prognostic nomogram for overall survival prediction in hepatoblastoma, hepatocellular carcinoma, and embryonal sarcoma in children and adolescent patients, which will further benefit the assessment of long-term outcomes.