Development and validation of a prognostic nomogram for diffuse large B-cell lymphoma: a multicenter cohort study from Northern China.

OBJECTIVE

This study aimed to refine the prognostic stratification provided by the International Prognostic Index (IPI) for patients diagnosed with diffuse large B-cell lymphoma (DLBCL) in a northern Chinese cohort and to construct a more precise prognostic nomogram to facilitate individualized treatment strategies and improve survival outcomes.

METHODS

Clinical data from 1,364 patients diagnosed with DLBCL between January 2000 and December 2019 were retrospectively analyzed across multiple centers in Beijing and Shanxi, China. A prognostic nomogram model for overall survival (OS) was developed incorporating the following variables: age, disease stage, Eastern Cooperative Oncology Group (ECOG) performance status, lactate dehydrogenase (LDH) levels, involvement of various extranodal sites, and disease progression within 6 months (POD6). Model performance was evaluated through receiver-operating characteristic (ROC) curve analysis, calibration curves, concordance index (C-index), and area under the curve (AUC) metrics.

RESULTS

The developed nomogram demonstrated consistent predictive capacity, with AUC values exceeding 0.7 at 1-, 3-, and 5-year time points. Calibration plots indicated close alignment with the ideal reference line, with slopes approximating 1, supporting the model's predictive accuracy and clinical relevance. POD6 was identified as the most significant high-risk factor associated with decreased OS (hazard ratio [HR] = 5.13, 95% confidence interval [CI]: 1.03-2.51, p < 0.0001). Among the IPI components, all except ECOG performance status and extranodal involvement remained significant. Notably, central nervous system involvement exhibited the strongest adverse prognostic effect among extranodal sites (HR = 1.61, 95% CI: 1.03-2.51, p = 0.035).

CONCLUSION

A novel prognostic nomogram was established for patients with DLBCL, offering improved prognostic accuracy compared to the traditional IPI. This model presents clinical utility in supporting personalized management and treatment planning.

CLINICAL REGISTRATION NUMBER

Not applicable.