Role of myocardial work in identifying anthracycline-induced cardiotoxicity and its correlation with metabolic fingerprinting in patients with breast cancer

Background Global longitudinal strain (GLS) is the standard of care in cardio-oncology to detect subclinical cardiotoxicity (CTX). However, GLS is load-dependent, and differences in systolic blood pressure represent a limitation for serial comparisons. Myocardial Work (MW) permits GLS adjustments for systolic blood pressure, allowing a left ventricular (LV) performance analysis despite afterload variations. Purpose We aimed to investigate the role of MW in the early detection of CTX and the potential role of metabolomics analysis in identifying a metabolic pattern (fingerprint) capable of detecting early adaptive changes suggestive of CTX. Methods Patients with breast cancer treated with cardiotoxic therapies were enrolled.The patients underwent baseline assessment and, upon reaching cumulative doses of anthracycline, were evaluated with echocardiographic examination with image acquisition for GLS and MW analysis. A peripheral blood sampling for metabolomic analysis performed by gas chromatography-mass spectrometry (GC/MS) was collected. Results We enrolled 47 patients; the mean age was 53±11. Baseline LV ejection fraction (EF) was >55% for the whole population, and GLS values were -21± 3%. During treatment, LVEF showed a progressive slight reduction without any individual values falling into the pathological range (67 ± 6% to 65 ± 6%, p=0.90). Similarly, a progressive decline in GLS values was observed starting at a dose of 180 mg/m², becoming more pronounced at a dose of 360 mg/m². Twelve patients exhibited a significant GLS decline. In these patients, heart failure therapy led to a partial recovery of GLS at the dose of 360 mg/m². A similar trend was observed for MW, with significant differences at both 180 mg/m² (p=<0.02 vs basal), and 360 mg/m² (p<0.02 vs basal). Subsequently, we performed two OPLS models using MW values as the Y variable and the data matrix from GC-MS analysis as the X variables, one for each of the two groups analyzed (patients without and with CTX). In both cases, the analysis produced significant results (p< 0.05), confirming MW as a landmark for cardiac muscle workload. Subsequently, we compared the two patient groups performing an OPLS-DA (Figure 1) on the same data matrix. The analysis showed group clustering with good statistical significance, suggesting metabolic remodeling linked to LV performance changes post-chemotherapy. To identify the involved metabolites, VIP analysis revealed a differential distribution of fatty acids, substrates such as ketone bodies, intermediates in energy production such as citrate, and various amino acids, including branched-chain amino acids. Conclusions Our data confirm that CTX can be detected effectively using both GLS and MW. The reduction of MW seems to be associated with a specific metabolic fingerprint. If confirmed, these results could enhance understanding of chemotherapy-induced damage and improve early, sensitive screening in patient follow-up.