Related Experiment Video
Updated: Sep 29, 2026
Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
Identification of Discriminatory Factors and Construction of a Nomogram for Differentiating IgG4-RD and DLBCL
Yongkang Chen1, Weimiao Li1, Zhou Zhou1
1The First Clinical College, Xuzhou Medical University, Xuzhou, China.
Objective:
To determine clinically useful indicators that distinguish IgG4-related disease (IgG4-RD) from diffuse large B-cell lymphoma (DLBCL) and to construct a predictive model that may assist early identification.
Methods:
We retrospectively analyzed 19 patients with IgG4-RD and 37 with DLBCL who presented with lymphadenopathy and received treatment at the Affiliated Hospital of Xuzhou Medical University between January 2012 and October 2025. The distribution of continuous variables was checked before group comparisons. Data with an approximately normal distribution are expressed as mean ± standard deviation (mean ± SD) and were compared by independent-samples t-tests; skewed data are reported as median (interquartile range [IQR]) and were compared by the Mann-Whitney U test. Categorical variables are presented as counts and percentages (%) and were compared using the χ2 test or Fisher's exact test, as appropriate. Variables significant in univariate analyses were taken forward into multivariable logistic regression. Predictive performance was evaluated by receiver operating characteristic (ROC) analysis, calibration analysis, and decision curve analysis (DCA).
Results:
Multivariable logistic regression retained lymphocyte count, serum total protein, and prothrombin time (PT) as independent predictive factors. The model yielded an area under the receiver operating characteristic curve (AUC) of 0.91 (95% confidence interval [CI]: 0.82-0.99). Its calibration curve had a mean absolute error of 0.032. Across threshold probabilities from 0.2 to 1.0, decision curve analysis indicated greater net benefit than the "treat-all" and "treat-none" approaches.
Conclusion:
In this cohort, the predictive model showed potential to assist differentiation between IgG4-RD and DLBCL. Because the study was retrospective, single-center, and based on a limited sample, these findings should be regarded as preliminary and require confirmation in larger independent cohorts before clinical use.
