Prediction of prognosis and pathologic grade in follicular lymphoma using (18)F-FDG PET/CT

PURPOSE: We investigated the utility of a new baseline PET parameter expressing lesion dissemination and metabolic parameters for predicting progression-free survival (PFS) and pathologic grade in follicular lymphoma (FL). METHODS: The baseline (18)F-FDG PET/CT images of 126 patients with grade 1–3A FL were retrospectively analyzed. A novel PET/CT parameter characterizing lesion dissemination, the distance between two lesions that were furthest apart (D (max)), was calculated. The total metabolic tumor volume and total lesion glycolysis (TLG) were computed by using 41% of the maximum standardized uptake value (SUV(max)) thresholding method. RESULTS: The 5-year PFS rate was 51.9% for all patients. In the multivariate analysis, high D (max) [P = 0.046; hazard ratio (HR) = 2.877], high TLG (P = 0.004; HR = 3.612), and elevated serum lactate dehydrogenase (P = 0.041; HR = 2.287) were independent predictors of PFS. A scoring system for prognostic stratification was established based on these three adverse factors, and the patients were classified into three risk categories: low risk (zero to one factor, n = 75), intermediate risk (two adverse factors, n = 29), and high risk (three adverse factors, n = 22). Patients in the high-risk group had a shorter 3-year PFS (21.7%) than those in the low- and intermediate-risk groups (90.6 and 44.6%, respectively) (P < 0.001). The C-index of our scoring system for PFS (0.785) was superior to the predictive capability of the Follicular Lymphoma International Prognostic Index (FLIPI), FLIPI2, and PRIMA-Prognostic Index (C-index: 0.628–0.701). The receiver operating characteristic curves and decision curve analysis demonstrated that the scoring system had better differentiation and clinical utility than these existing indices. In addition, the median SUV(max) was significantly higher in grade 3A (36 cases) than in grades 1 and 2 FL (90 cases) (median: 13.63 vs. 11.45, P = 0.013), but a substantial overlap existed (range: 2.25–39.62 vs. 3.17–39.80). CONCLUSION: TLG and D (max) represent two complementary aspects of the disease, capturing the tumor burden and lesion dissemination. TLG and D (max) are promising metrics for identifying patients at a high risk of progression or relapse. Additionally, SUV(max) seems to have some value for distinguishing grade 3A from low-grade FL but cannot substitute for biopsy.