Background Pediatric oncology patients with sepsis are at high risk of morbidity and mortality due to immunosuppression and acute or fulminant multiorgan dysfunction. There have been many clinical scores proposed for risk of mortality prediction (PELOD‐2, pSOFA, and Phoenix), but it remains unknown how well these scores predict risk in pediatric oncology patients with sepsis. Objective To systematically review and meta‐analysis the predictive performance of clinical scores and evaluate the incremental benefit of biomarkers in pediatric oncology patients. Methods A systematic search was conducted in PubMed, Web of Science, Scopus, and Embase through 2025. Eligible studies were those that assessed PELOD‐2, pSOFA, Phoenix, or prognostic scores in pediatric oncology patients with sepsis. Risk of bias assessment was completed using QUADAS‐2. Random‐effects meta‐analysis was used to pool sensitivity, specificity, and area under the curve (AUC) estimates across studies. Results A total of 32 articles were summarized with cohorts of 50–1200 patients per study. PELOD‐2 and pSOFA demonstrated consistent and strong discrimination for mortality (AUC 0.78–0.88) while the Phoenix score demonstrated moderate discriminatory ability (AUC 0.72–0.83) and little validation. Procalcitonin, C‐reactive protein, lactate, and other biomarkers improved predictive accuracy when combined with clinical scores. In summary, the overall risk of bias was rated to be moderate, largely due to predominately retrospective designs. Conclusions PELOD‐2 and pSOFA are the most validated prognostic tools for pediatric oncology patients with sepsis, while the Phoenix score may be useful in selected settings. Integration of biomarkers improves risk stratification. Prospective multicenter studies are needed to refine prognostic models and guide early interventions in this high‐risk population.
Domínguez-Rojas et al. (2026) studied this question.