A better prediction of progression‐free survival in diffuse large B‐cell lymphoma by a prognostic model consisting of baseline TLG and %ΔSUV(max)

In the era of rituximab, the International Prognostic Index (IPI) has been inefficient in initial risk stratification for patients with R‐CHOP‐treated diffuse large B‐cell lymphoma (DLBCL). To estimate the predictive values of PET/CT quantitative parameters and three prognostic models consisting of baseline and interim parameters for three‐year progression‐free survival (PFS), we conducted an analysis of 85 patients in China with DLBCL underwent baseline and interim PET/CT scans and treated at the Department of Hematology of Peking University Third Hospital from November 2012 to November 2017. The PET/CT parameters, viz. the baseline and interim values of standardized uptake value (SUV(max)), total metabolic tumor volume (TMTV), and total lesion glycolysis (TLG), and their rates of change, were analyzed by a receiver operating characteristics curve, Kaplan‐Meier analysis, and log‐rank test. Besides, the National Comprehensive Cancer Network International Prognostic Index (NCCN‐IPI) was also included in the multivariate Cox hazards model. Owing to the strong correlation between TMTV and TLG at baseline and interim (Pearson's correlation coefficient, r = 0.823, P‐value = 0.000, and 0.988, P‐value = 0.000, respectively), only TLG was included in the multivariate Cox hazards model, where TLG(0) > 1036.61 g and %ΔSUV(max) < 86.02% showed predictive value independently (HR = 10.42, 95% CI 2.35‐46.30, P = 0.002, and HR = 4.86, 95% CI 1.27‐18.54, P = 0.021, respectively). Replacing TLG in the equation, TMTV(0) and TMTV(1) both showed significantly predictive abilities like TLG (HR = 8.22, 95% CI 1.86‐32.24, P = 0.005, and HR = 2.96, 95% CI 1.16‐7.54, P = 0.023, respectively). After dichotomy, NCCN‐IPI also gave a significant performance (P = 0.035 and P = 0.010, respectively, in TLG and TMTV models). The baseline variables, that is, TMTV(0), TLG(0) and dichotomized NCCN‐IPI, and the interim variables TMTV(1) and %ΔSUV(max), presented independent prognostic value for PFS. In prognostic model 2 (TLG(0) + %ΔSUV(max)), the group with TLG(0) > 1036.61 g and %ΔSUV(max) < 86.02% recognized 19 (82.6%) of the relapse or progression events, which showed the best screening ability among three models consisting of baseline and interim PET/CT parameters.