Prediction nomogram based on 18F-FDG PET/CT and clinical parameters for patients with diffuse large B-cell lymphoma.

Abstract

The objective of this study was to develop a nomogram including parameters assessed by 18F-FDG PET/CT and clinical parameters for patients with diffuse large B-cell lymphoma (DLBCL) to predict progression-free survival (PFS). A total of 181 patients with pathologically diagnosed DLBCL at Sichuan Cancer Hospital and Institute from March 2015 to December 2020 were enrolled in this retrospective study. The area under the receiver operating characteristic (ROC) curve (AUC) was used to calculate the optimal cutoff values of the semiquantitative parameters (SUVmax, TLG, MTV, and Dmax) for PFS. A nomogram was constructed according to multivariate Cox proportional hazards regression. The predictive and discriminatory capacities of the nomogram were then measured using the concordance index (C-index), calibration plots, and Kaplan-Meier curves. The predictive and discriminatory capacities of the nomogram and the International Prognostic Index of the National Comprehensive Cancer Network (NCCN-IPI) were compared via the C-index and AUC. Multivariate analysis demonstrated that male gender and pretreatment Ann Arbor stage III-IV, non-GCB, elevated lactate dehydrogenase (LDH), number of extranodal organ involvement (Neo)>1, MTV≥152.8 cm3, and Dmax ≥53.9 cm were associated with unfavorable PFS (all p<0.05). The nomogram, including gender, Ann Arbor stage, pathology type, Neo, LDH levels, MTV, and Dmax, showed good prediction accuracy, with a C-index of 0.760 (95% CI: 0.727-0.793), which was higher than that of NCCN-IPI (0.710; 95% CI: 0.669-751). The calibration plots for 2-year demonstrated good consistency between the predicted and observed probabilities for survival time. We established a nomogram including MTV, Dmax, and several clinical parameters to predict the PFS of patients with DLBCL, and the nomogram showed better predictability and higher accuracy than NCCN-IPI.