Prognostic Value of Baseline Tumor Burden and Tumor Dissemination Extracted From 18F-FDG PET/CT in a Cohort of Adult Patients With Early or Advanced Hodgkin Lymphoma

Authors:
Publisher website
Google Scholar
Abstract
Purpose: We aimed to assess the prognostic value of baseline tumor burden and dissemination parameters extracted from 18F-FDG PET/CT in patients with early or advanced Hodgkin lymphoma (HL) treated with ABVD (doxorubicin, bleomycin, vinblastine, and dacarbazine) or escalated BEACOPP (increased bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone). Patients and Methods: Patients aged ≥18 years with classical Hodgkin lymphoma were retrospectively included. Progression-free survival (PFS) analysis of dichotomized clinicobiological and PET/CT parameters (SUVmax, TMTV, TLG, Dmax, and Dbulk) was performed. Optimal cutoff values for quantitative metrics were defined as the values maximizing the Youden index from receiver operating characteristic analysis. PFS rates were estimated with Kaplan-Meier curves, and the log-rank test was used to assess statistical significance. Hazard ratios were calculated using Cox proportional hazards models. Results: With a median age of 32 years, 166 patients were enrolled. A total of 111 patients had ABVD or ABVD-like treatment with or without radiotherapy and 55 patients with escalated BEACOPP treatment. The median follow-up was 55 months. Only International Prognostic Score (IPS >1), TMTV >107 cm3, and TLG >1628 were found to be significant prognostic factors for PFS on univariate analysis. Multivariate analysis revealed that IPS and TLG were independently prognostic and, combined, identified 4 risk groups (P < 0.001): low (low TLG and low IPS; 4-year PFS, 95%), intermediate-low (high IPS and low TLG; 4-year PFS, 79%), intermediate-high (low IPS and high TLG; 4-year PFS, 78%), and high (high TLG and high IPS; 4-year PFS, 71%). Conclusions: Combining baseline TLG with IPS could improve PFS prediction.

This publication has 24 references indexed in Scilit: