Performance of the Phoenix Sepsis Score for in-hospital mortality in critically ill pediatric oncology patients with suspected infection

BACKGROUND: Pediatric oncology patients with infection face a high risk of death. Although the Phoenix Sepsis Score (PSS) has recently been introduced as an organ dysfunction score for pediatric sepsis, its performance in pediatric oncology populations remains insufficiently characterized. METHODS: This retrospective study enrolled pediatric oncology patients admitted to the pediatric intensive care unit (PICU) with suspected infection. The performance of the PSS was evaluated and compared with other pediatric critical care scores for discrimination of in-hospital mortality. RESULTS: The study cohort comprised 224 patients, accounting for 254 PICU admissions. The overall hospital mortality rate was 17.7%. All scoring systems except the Pediatric Risk of Mortality III (PRISM III) score demonstrated good discrimination. Phoenix-8 showed the highest area under the curve (AUC) (0.855, 95% CI: 0.791-0.919), followed by PSS (0.837) and Pediatric Sequential Organ Failure Assessment (pSOFA) score (0.836). Phoenix-8 had significantly higher AUCs than PHO-Phoenix and PRISM III; differences among the other scoring systems were not statistically significant. Sensitivity analyses were broadly consistent with the primary analysis but showed lower AUCs after inclusion of deaths following discontinuation of life-sustaining therapy. CONCLUSIONS: The PSS showed good discrimination for in-hospital mortality in PICU-admitted pediatric oncology patients with infection, with an AUC of similar magnitude to those of pSOFA and PELOD-2. Its lower data requirements may support its use as a clinically feasible tool for mortality risk assessment in this high-risk population.