Prediction of response to neoadjuvant chemotherapy in osteosarcoma using dual-phase (18)F-FDG PET/CT

We evaluated the ability of dual-phase (18)F-FDG PET/CT to predict the histological response after neoadjuvant chemotherapy (NAC) in osteosarcoma. Thirty-one patients with osteosarcoma treated with NAC and surgery were prospectively enrolled. After injection of (18)F-FDG, both early (~60 min) and delayed (~150 min) PET were acquired before and after the completion of NAC. SUVmax, early/delayed SUVmax change (RImax), and early/delayed SUVmean change (RImean) of tumour were measured before (SUV1, RImax1, and RImean1) and after NAC (SUV2, RImax2, and RImean2). Then, we calculated the percentage changes between SUV1 and SUV2 (%SUV). Twelve patients (39%) exhibited good histological response after NAC. SUVmax, RImax, and RImean significantly decreased after NAC. Before NAC, only RImean1 predicted good histological response with the optimal criterion of < 10%, sensitivity of 92%, specificity of 57%, and accuracy of 71%. After NAC, %SUV, SUV2, and RImax2 predicted histological response. By using combined criterion of %SUV and RImax2 or SUV2 and RImean1 or SUV2 and RImax2, accuracies were 81%, 77%, and 77%, respectively. The histological response after NAC could be predicted by using RImean1 before the initiation of NAC in osteosarcoma. The combined use of SUV and RI values may provide a better prediction. • Pretreatment dual-phase FDG-PET was useful to predict histological response in osteosarcoma. • A combination of early and delayed PET may increase the predictive value. • Early/delayed SUV change of tumours significantly decreased after neoadjuvant chemotherapy.