Metabolomic profiling in neuroblastoma.

Background And Objectives: Previous studies on several cancer types show that metabolomics provides a potentially useful noninvasive screening approach for outcome prediction and accurate response to treatment assessment. Neuroblastoma (NB) accounts for at least 15% of cancer-related deaths in children. Although current risk-based treatment approaches in NB have resulted in improved outcome, survival for high-risk patients remains poor. This study aims to evaluate the use of metabolomics for improving patients' risk-group stratification and outcome prediction in NB.

Design And Methods: Plasma samples from 110 patients with NB were collected at diagnosis prior to starting therapy and at the end of treatment if available. Metabolomic analysis of samples was carried out by ultra-performance liquid chromatography-time of flight mass spectrometry (UPLC-MS).

Results: The metabolomic analysis was able to identify different plasma metabolic profiles in high-risk and low-risk NB patients at diagnosis. The metabolic model correctly classified 16 high-risk and 15 low-risk samples in an external validation set providing 84.2% sensitivity (60.4-96.6, 95% CI) and 93.7% specificity (69.8-99.8, 95% CI). Metabolomic profiling could also discriminate high-risk patients with active disease from those in remission. Notably, a plasma metabolomic signature at diagnosis identified a subset of high-risk NB patients who progressed during treatment.

Conclusions: To the best of our knowledge, this is the largest NB study investigating the prognostic power of plasma metabolomics. Our results support the potential of metabolomic profiling for improving NB risk-group stratification and outcome prediction. Additional validating studies with a large cohort are needed.