Evaluating early interim 18F-FDG PET/CT with Peking criteria for predicting the outcome of diffuse large B cell lymphoma

1603 Objectives The prognostic value of early interim fluorodeoxyglucose (18F-FDG) positron emission tomography/ computed tomography (PET/CT) is debated in patients with diffuse large B cell lymphoma (DLBCL) due to heterogeneous interpretation scales. This study investigated whether a new interpretation scale, the Peking Criteria, could be a superior method for evaluating early interim PET, compared to 5-PS and ΔSUVmax criteria. Methods One hundred and nineteen patients with DLBCL underwent 18F-FDG PET/CT scans at baseline (PET0) and after 2 cycles of chemotherapy (PET2). PET2 scans were evaluated with Peking criteria, comparing to 5-PS criteria, and ΔSUVmax criteria. For each PET interpreting with Peking criteria, the lymphoma residue with the most intense 18F-FDG uptakes was identified among all foci, and measured and calculated via ROI technique. Liver uptake was designed to measure on the maximum cross section. Liver SUVmax (SUVmax-liver) were taken as a reference and The optimal threshold was decided via interobserver agreements and prognostic. Residue SUVmax higher than the optimal threshold or new 18F-FDG avid lesions indicated the positive lesion in the interim PET. Treatment response was evaluated using the revised criteria of Cheson et al. at the end of the first line of treatment or at the time of progression. Prognostic values using three criteria were compared via the survival analysis. Uni- and multivariate analyses of outcomes were performed using clinical variables and PET2 scans. Results Over the median follow-up of 19 months, 23 patients (19.3%) achieved a complete response, 91 patients (76.5%) achieved a partial response, and 5 patients (4.2%) showed a stable disease or progression. Twenty-two patients died during follow-up. Three criteria attributed by three observers are shown in Table 1. The Cohen k for agreement between pairs of observers ranged from 0.83 to 0.94 in the Peking Criteria, revealing almost perfect agreement. The optimal threshold for Peking criteria is 1.6 fold of SUVmax-liver. The Cohen k values for reproducibility of Peking criteria were above 0.90, which was better than those of 5-PS or ΔSUVmax interpretation. Prognostic value of PET2 using Peking, 5-PS, and 66%ΔSUVmax scale were compared in Table 2. Overall, the best accuracy (76.5%) in predicting 3-year PFS was obtained when using Peking Criteria. The predicting accuracy of 3-year OS for Peking Criteria (73.9%) was detected as same as 66%ΔSUVmax scale, but better than the visual analysis. Negative predictive values (NPV) of PFS (74.7% to 77.9%) and OS (83.5% to 85.9%) were high for the interim PET2. The 3-year PFS and OS were 75.1% and 78.6% in patients with positive residue, using the Peking Criteria, versus 15.8% and 36.9% in patients with negative residue (P 60 years, advanced stage, elevated LDH and BMG, high IPI score, positive residue interpreting with Peking Criteria, 5-PS, and ΔSUVmax scale were significant predictors of PFS and OS (P Conclusions Early interim 18F-FDG PET/CT can effectively predict the outcome in patients with DLBCL by using the Peking Criteria. Figure 1 PFS and OS based on evaluating the interim PET2 scans using the Peking Criteria(liver×1.6)(A,B), 5-PS Criteria (score 4) (C,D), or 66%ΔSUVmax criteria (E,F). There are statistically significant differences (P < 0.001) $$graphic_4CAB5D57-109A-4E24-89A5-4C104F816E54$$ $$graphic_B2C0D672-E3DF-43BD-A39D-50E7FF75061B$$ $$graphic_A2B62A80-2096-4213-A9B2-75DC5BCB8BBF$$ $$graphic_DE87A655-9A1F-4DA6-A25E-9C5E6D8C0D60$$