Correlation of 18-fluorodeoxyglucose PET/computed tomography parameters and clinical features to predict outcome for diffuse large B-cell lymphoma.

PURPOSE To determine if the correlation between different metabolic parameters along with clinical features can create an improved model of prognostication for diffuse large B-cell lymphoma (DLBCL) patients. METHODS We retrospectively evaluated 89 patients with DLBCL. All patients had a baseline and an interim 18F-FDG PET/CT. Seventy-nine also had an end-of-treatment PET/CT (EOT-PET). For each scan, we collected standardized uptake value (SUVmax, SUVmean, SUVpeak), metabolic tumor volume (MTV), total lesion glycolysis (TLG), SUVmaxsum, SUVmeansum, MTVsum, and TLGsum. These metabolic parameters were combined with clinical features in order to identify a new prognostic model. The predictive value of interim PET and EOT-PET using Deauville score was also determined. RESULTS Baseline SUVmaxsum and SUVmeansum were significantly correlated to overall survival (OS) (P value = 0.012 and 0.011, respectively). The percentage change of MTV and TLG sum from baseline to EOT was predictive of progression-free survival (PFS) (P value = 0.003 and 0.022, respectively). The combination of either Deauville score at the EOT and SUVmaxsum at baseline significantly predicted OS (P value <0.001); Eastern Cooperative Oncology Group performance status, presence of extranodal disease and percentage change of MTVsum from baseline to EOT were significant predictors of PFS (P value = 0.001). CONCLUSIONS SUVmaxsum and SUVmeansum at baseline and percentage change in MTV and TLG sum from baseline to EOT are predictors of outcome in DLBCL patients. These metabolic parameters combined to Deauville score and some clinical features could be used together to stratify patients.