Chemotherapy-induced cardiotoxicity (CIC) is a common adverse effect of antineoplastic drugs, manifesting in adult and paediatric populations. The biochemical background of CIC risk remains unclear. However, recent genomics studies linked the condition to specific gene polymorphisms. The application of metabolomics has the potential to improve our understanding of CIC risk, bridging the gap between the genetic predisposition, the chemotherapy and the onset of CIC. Accordingly, an untargeted metabolomics approach was implemented to determine correlations between patients' metabolomics profiles and the risk of CIC during or shortly following their chemotherapy. A UPLC-ESI-QTOF protocol was developed to analyse plasma samples of 89 paediatricpatients collected prior to the initiation of chemotherapy. The metabolomics data were integrated with the a posteriori knowledge of CIC expression in a post-hoc analysis. KODAMA, OPLS-DA, BORUTA, and FDR t-test were used as complementary classification and variables selection methodologies to point out metabolites with an increased probability of being CIC risk-related. An empirical scale of confidence for DIA identification was developed to assure results' reliability. All classification models succeeded in separating the CIC-risk patients. Overall, CIC-risk patients showed early alterations in pathways related to CVDs, inflammation, oxidation, folate deficiency, and inhibition of GSH production. The 3-hydroxy-9-hexadecenoylcarnitine was determined as a potential predictive biomarker.