Clinical Risk Factors and Outcomes of Spontaneous Hepatoblastoma Rupture in Children: A 10-Year Propensity Score-Matched Study

Background: Spontaneous rupture of hepatoblastoma is uncommon but potentially fatal in children. Published cohorts often combine spontaneous, traumatic, biopsy-related, and treatment-associated events. We aimed to identify clinical risk factors for spontaneous rupture and evaluate its prognostic impact. Methods: We retrospectively reviewed 106 children with pathologically confirmed hepatoblastoma treated at a tertiary center from July 2010 to July 2020. After excluding 10 with incomplete data or follow-up, 96 were eligible. Fourteen children who presented with spontaneous rupture were included in the rupture group. Eighty-two children without rupture were eligible as controls, and 42 were selected after 1:3 propensity score matching for age, sex, height, and weight. Subsequent analyses were restricted to the 1:3 propensity score-matched cohort. However, the matched-set identifier was not incorporated into the group comparisons, regression models, or survival analyses. Rupture was diagnosed by contrast-enhanced imaging and/or intraoperative findings. Logistic regression was used to identify independent risk factors. Exploratory two-predictor ridge regression and 2000-bootstrap internal validation were used to examine maximum tumor diameter and macrovascular invasion. Overall survival (OS) and event-free survival (EFS) were estimated using Kaplan–Meier analysis. Results: The rupture group had lower hemoglobin levels, larger tumors, more frequent bilobar disease, higher PRETEXT stage, and substantially more vascular invasion. Exploratory multivariable analysis suggested associations of maximum tumor diameter (OR 3.078, 95% CI 1.62–5.55) and macrovascular invasion (OR 13.521, 95% CI 1.16–148.23) with rupture. Given the small number of rupture events and the wide confidence interval for macrovascular invasion, these estimates should be interpreted cautiously and not as confirmation of independent risk factors. Macrovascular invasion was associated with rupture in crude analysis (odds ratio 8.00, 95% confidence interval 2.05–31.16) and the exploratory ridge model (adjusted odds ratio 2.89, 95% bootstrap confidence interval 1.27–7.93). The combined model incorporating maximum tumor diameter and macrovascular invasion showed an apparent area under the curve of 0.930 and an optimism-corrected area under the curve of 0.924. Children with ruptures had significantly worse OS and EFS than matched controls. Conclusions: Spontaneous rupture was associated with a clinically aggressive presentation characterized by substantially larger tumors, advanced local disease, and more frequent macrovascular invasion, and affected children had markedly lower OS and EFS estimates than matched controls. The results support tumor burden and macrovascular invasion as clinically relevant warning features that may help identify children requiring closer surveillance and early multidisciplinary planning.