RESUMO
PURPOSE: We investigated the ability of baseline 2-deoxy-2-[18F]fluoro-D-glucose PET/CT parameters, acquired before the start of immunotherapy, to predict development of hyperprogressive disease (HPD) in melanoma patients. We also evaluated the diagnostic performances of ratios of baseline and first restaging PET/CT parameters to diagnose HPD without information of the tumor growth kinetic ratio (TGKR) that requires pre-baseline imaging before baseline imaging (3 timepoint imaging). PROCEDURES: Seventy-six patients who underwent PET/CT before and approximately 3 months following initiation of immunotherapy were included. PET/CT parameters, including metabolic tumor volume (MTV) for all melanoma lesions and total measured tumor burden (TMTB) based on irRECIST, were measured from baseline PET/CT (MTVbase and TMTBbase) and first restaging PET/CT (MTVpost and TMTBpost). The ratios of MTV (MTVpost/MTVbase, MTVr) and TMTB (TMTBpost/TMTBbase, TMTBr) were calculated. RESULTS: MTVbase of HPD patients (n = 9, TGKR ≥ 2) was larger than that of non-HPD (n = 67, TGKR < 2) patients (P < 0.05), and HPD patients demonstrated shorter median overall survival (7 vs. more than 60 months, P < 0.05). The area under the curve (AUC) of MTVbase (≥ 155.5 ml) to predict the risk of HPD was 0.703, with a sensitivity of 66.7 % and specificity of 81.2 %. The AUCs of MTVr (≥ 1.25) and TMTBr (≥ 1.27) to diagnose HPD without information of TGKR were 0.875 and 0.977 with both sensitivities of 100 %, and specificities of 79 % and 83.9 %, respectively. CONCLUSIONS: Patients at high risk of developing HPD could not be accurately identified based on baseline PET/CT parameters. The ratios of baseline and first restaging PET/CT parameters may be helpful to diagnose HPD, when patients do not undergo pre-baseline imaging.