Abstract
PURPOSE: To verify the capability of (18)F-fluorodeoxy-glucose positron emission tomography/computed tomography (FDG-PET/CT) to identify patients at higher risk of developing doxorubicin (DXR)-induced cardiotoxicity, using a score-based image approach. METHODS: 36 patients underwent FDG-PET/CT. These patients had shown full remission after DXR-based chemotherapy for Hodgkin's disease (DXR dose: 40-50 mg/m² per cycle), and were retrospectively enrolled. Inclusion criteria implied the presence of both pre- and post-chemotherapy clinical evaluation encompassing electrocardiogram (ECG) and echocardiography. Myocardial metabolism at pre-therapy PET was evaluated according to both standardized uptake value (SUV)- and score-based approaches. The capability of the score-based image assessment to predict the occurrence of cardiac toxicity with respect to SUV measurement was then evaluated. RESULTS: In contrast to the SUV-based approach, the five-point scale method does not linearly stratify the risk of the subsequent development of cardiotoxicity. However, converting the five-points scale to a dichotomic evaluation (low vs. high myocardial metabolism), FDG-PET/CT showed high diagnostic accuracy in the prediction of cardiac toxicity (specificity = 100% and sensitivity = 83.3%). In patients showing high myocardial uptake at baseline, in which the score-based method is not able to definitively exclude the occurrence of cardiac toxicity, myocardial SUV mean quantification is able to further stratify the risk between low and intermediate risk classes. CONCLUSIONS: the score-based approach to FDG-PET/CT images is a feasible method for predicting DXR-induced cardiotoxicity. This method might improve the inter-reader and inter-scanner variability, thus allowing the evaluation of FDG-PET/CT images in a multicentral setting.