A new clinical-laboratory nomogram has been developed and internally validated to help identify patients with diffuse large B-cell lymphoma (DLBCL) who are at high risk of experiencing primary refractory disease or early relapse following standard first-line R-CHOP immunochemotherapy.
First-line R-CHOP immunochemotherapy controls diffuse large B-cell lymphoma in most patients, but 30% to 40% develop primary refractory disease or early relapse. Refractory disease and early relapse, which occur in a substantial subset of patients, are both associated with poor survival outcomes. The International Prognostic Index captures tumor biology only partially, and no bedside tool integrates routine inflammatory, metabolic, and proliferative markers. The resulting nomogram incorporates six specific routine laboratory markers: lactate dehydrogenase (LDH), β2-microglobulin (β2-MG), the neutrophil-to-lymphocyte ratio (NLR), the platelet-to-lymphocyte ratio (PLR), Ki-67, and the lymphocyte-to-monocyte ratio (LMR).
The study was designed as a retrospective analysis where the total cohort of 342 patients was randomly partitioned into a training cohort consisting of 239 individuals and an internal validation cohort of 103 individuals, maintaining a 7:3 ratio. The research team first conducted a univariate analysis to screen for potential predictors, requiring a P-value of less than 0.05 to retain candidate variables. Univariate analysis was followed by LASSO and multivariable logistic regression.
To rigorously evaluate the predictive power of their chosen logistic regression model, the team compared it against other common machine learning approaches, including random forest (RF) and support vector machines (SVM). Performance was assessed using a variety of metrics, including the area under the receiver operating characteristic curve (ROC-AUC), calibration plots, and decision curve analysis. Furthermore, SHAP analysis was used to quantify predictor contributions on the log-odds scale.
The findings indicated that elevated levels of LDH, β2-MG, NLR, PLR, and Ki-67, coupled with reduced levels of LMR, were independent predictors of early treatment failure. The logistic regression model achieved an AUC of 0.815 in the training cohort, with a 95% confidence interval ranging from 0.747 to 0.883, and maintained a performance of 0.776 in the validation cohort. There were no significant between-model AUC differences, and the logistic model was selected for parsimony and interpretability.
Assessment of the model’s reliability showed that the calibration slopes and intercepts were 0.97 and 0.03 in the training set, and 0.93 and 0.07 in the validation set, indicating that the nomogram’s predictions align closely with the observed outcomes. Decision curve analysis further demonstrated the clinical utility of the model, revealing positive net benefits at risk thresholds of 0.2, 0.3, and 0.4.
Despite these promising results, the study authors emphasize that the current nomogram remains an exploratory tool. A critical limitation of this research is that it was restricted to a single-center cohort, which may limit the generalizability of the findings across more diverse patient populations treated in different clinical environments. Moreover, the model was not formally benchmarked against established clinical prognostic systems such as the standard IPI, NCCN-IPI, or R-IPI.
For this nomogram to transition into actual clinical use, it requires extensive multicenter prospective external validation.
Subject of Research: Clinical research on a predictive nomogram for early treatment failure in DLBCL.
Article Title: A clinical-laboratory nomogram integrating inflammatory, metabolic and proliferative markers to predict refractory disease or early relapse in diffuse large B-cell lymphoma
Article References: Li, S., Yang, Q., Zhang, W., & Chen, W. (2026). A clinical-laboratory nomogram integrating inflammatory, metabolic and proliferative markers to predict refractory disease or early relapse in diffuse large B-cell lymphoma. Frontiers in Medicine, 13, Article 1995591. https://doi.org/10.3389/fmed.2026.1995591
Image Credits: AI Generated
DOI: 10.3389/fmed.2026.1995591
Keywords: Diffuse large B-cell lymphoma, DLBCL, R-CHOP, immunochemotherapy, refractory disease, early relapse, oncology, clinical prognosis, biomarker, nomogram, LDH, beta-2 microglobulin, neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, Ki-67, lymphocyte-to-monocyte ratio, machine learning, clinical prediction model, hematology
News Source: Nathaniel Bowman. (November 17, 2026). Nomogram Predicts Early Treatment Failure in Diffuse Large B-Cell Lymphoma Patients. Scienmag.



