Early 18F-FDG PET for Prediction of Prognosis in Patients with Diffuse Large B-Cell Lymphoma: SUV-Based Assessment Versus Visual Analysis

TLDR

The study evaluates whether early (18)F‑FDG PET using SUV metrics predicts prognosis better than visual analysis in newly diagnosed diffuse large B‑cell lymphoma patients. The authors performed early (18)F‑FDG PET on 92 DLBCL patients before and after two chemotherapy cycles, calculated SUVmax, SUVmean, and tumor‑to‑normal ratios, and used ROC and Kaplan‑Meier analyses to compare these quantitative metrics with visual assessment for predicting event‑free and overall survival. SUV‑based assessment improved prognostic accuracy, achieving 76.1% for event‑free survival versus 65.2% for visual analysis, and identified a 65.7% SUVmax reduction cutoff that separated 21% from 79% 2‑year EFS, demonstrating better outcome prediction in DLBCL.

Abstract

The purpose of this study was to assess the prognostic value of early (18)F-FDG PET using standardized uptake value (SUV) compared with visual analysis in patients with diffuse large B-cell lymphoma (DLBCL).Ninety-two patients with newly diagnosed DLBCL underwent (18)F-FDG PET prospectively before and after 2 cycles of chemotherapy (at midtherapy). Maximum SUV (SUVmax) and mean SUV (SUVmean) normalized to body weight and body surface area, as well as tumor-to-normal ratios, were computed on the most intense uptake areas. The SUVs, tumor-to-normal ratios, and their changes over time were compared with visual analysis for predicting event-free survival (EFS) and overall survival, using receiver-operating-characteristic (ROC) analysis. Survival curves were estimated with Kaplan-Meier analysis and compared using the log-rank test.With visual analysis, the accuracy of early PET to predict EFS was 65.2%. The 2-y estimate for EFS was 51% (95% confidence interval [CI], 34%-68%) in the PET-positive group compared with 79% (95% CI, 68%-90%) in the PET-negative group (P = 0.009). An optimal cutoff value of 65.7% SUVmax reduction from baseline to midtherapy obtained from ROC analysis yielded an accuracy of 76.1% to predict EFS. The 2-y estimate for EFS was 21% (95% CI, 0%-42%) in patients with SUVmax reduction <or= 65.7% compared with 79% (95% CI, 69%-88%) in those with reduction > 65.7% (P < 0.0001). Fourteen patients considered as positive on visual analysis could have been reclassified as good responders.SUV-based assessment of therapeutic response during first-line chemotherapy improves the prognostic value of early (18)F-FDG PET compared with visual analysis in DLBCL.