Establishment of prognostic nomogram in diffuse large B-cell lymphoma patients with bone marrow infiltration treated by rituximab and its predictive effect / 肿瘤研究与临床

ABSTRACT

Objective:To investigate the predictive efficacy of the established prognostic nomogram of rituximab in treatment of diffuse large B-cell lymphoma (DLBCL) patients with bone marrow infiltration.Methods:The clinicopathological data of 71 DLBCL patients with bone marrow infiltration who received first-line treatment with rituximab between January 2014 and June 2016 in Shanxi Province Cancer Hospital were retrospectively analyzed. Progression-free survival (PFS) analysis was performed by using Kaplan-Meier method, and influencing factors of PFS were analyzed by using univariate and multivariate Cox proportional hazards models. The nomogram was drawn with R software based on independent influencing factors of PFS from Cox regression analysis. Receiver operating characteristic (ROC) curve was applied to evaluate the effects of nomogram models predicting the PFS of patients; Bootstrap method was used for internal validation of the model. A nomogram calibration curve was plotted to compare the consistency between the nomogram model prediction and the actual PFS.Results:The median follow-up time of all patients was 48 months (12-84 months), and the 3-year and 5-year PFS rates were 39.44% and 26.76%, respectively. Age > 60 years ( HR = 1.593, 95% CI 1.379-1.840, P < 0.001), Ann-Arbor staging Ⅲ-Ⅳ ( HR = 1.444, 95% CI 1.092-1.910, P = 0.010), international prognostic index (IPI) score 3-5 ( HR = 1.648, 95% CI 1.249-2.333, P < 0.001), complicated with type 2 diabetes ( HR = 5.880, 95% CI 1.645-21.023, P = 0.006) were independent influencing factors of PFS in DLBCL patients with bone marrow infiltration. The independent influencing factors of PFS were included to establish the prognostic nomogram model. Bootstrap method internal validation showed that the consistency index of the prediction model was 0.71 (95% CI 0.69-0.78), and the ROC curve showed that the area under the curve (AUC) of 3-year PFS predicted by nomogram model was 0.708, 5-year PFS predicted by nomogram model was 0.716, indicating that nomogram model had a good degree of differentiation; and the calibration curve results showed that the 3-year and 5-year PFS rates predicted by nomogram model had a good consistency with the actual 3-year and 5-year PFS rates. Conclusions:The nomogram model constructed by age, Ann-Arbor staging, IPI score, complicated with or without type 2 diabetes could be used to predict the prognosis of DLBCL patients with bone marrow infiltration treated with rituximab, which is helpful for clinicians to implement treatment strategies.