Purpose: We aimed to create a novel prognostic score, the inflammation-based prognosis score (IBPS). In addition, we attempted to establish and validate a nomogram to predict the overall survival (OS) of patients with DLBCL.
Patients and Methods: We retrospectively investigated the data of 213 patients with DLBCL diagnosed and treated in the Affiliated Hospital of Jiangnan University and used these data to develop nomograms. At the same time, 89 patients diagnosed and treated in Wuxi People’s Hospital Affiliated with Nanjing Medical University from January 2015 to June 2021 were collected as an external validation cohort. We developed IBPS through the least absolute shrinkage and selection operator (LASSO) Cox regression. The univariate and multivariate Cox regression method was used to develop the nomogram. We used the concordance index (C-index), calibration chart, time-dependent receiver operating characteristic (ROC) analysis, decision curve analysis (DCA), and the Kaplan–Meier curve were used to assess the nomogram.
Results: The systemic immune inflammation index (SII), prognostic nutrition index (PNI), and modified Glasgow prognostic score (mGPS) were used to construct IBPS. The Eastern Cooperative Oncology Group performance status (ECOG PS), IBPS, response to treatment, and whether accept surgery were used to construct the nomogram to predict the OS of DLBCL patients. The C-index in the training and validation cohorts were 0.844 and 0.828, respectively. According to the time-dependent ROC curve and DCA, the nomogram has good predictive accuracy and clinical net benefit. The Kaplan–Meier curve showed that according to the nomogram score, patients in the training and validation cohorts could be classified into three risk groups.
Conclusion: In patients with DLBCL, baseline IBPS was a reliable predictor of OS. The survival probability of DLBCL patients can be precisely predicted using the prognosis nomogram based on IBPS.
Keywords: diffuse large B-cell lymphoma, inflammatory indicators, overall survival, prediction model