Factors affecting refractoriness or recurrence in diffuse large B-cell lymphoma: development and validation of a novel predictive nomogram.

Abstract

Background: Relapsed/Refractory (R/R) diffuse large B-cell lymphoma (DLBCL) represents a subgroup with a high incidence and dismal prognosis. Currently, there is a lack of robust models for predicting R/R DLBCL. Therefore, we conducted a retrospective study to identify key determinants to be incorporated into a novel nomogram to enhance the identification of DLBCL patients at elevated risk of refractoriness/recurrence.

Methods: We included 293 newly-diagnosed DLBCL patients from Harbin Medical University Cancer Hospital, collected from 2008-2017. Patients were randomly divided into a training cohort (n = 206) and a validation cohort (n = 87) at a 7:3 ratio. The training cohort underwent univariable analysis to select variables for a binary logistic regression model. These variables were also prioritized using a random forest algorithm. The developed nomogram was evaluated with the receiver-operator characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA) for its clinical utility.

Results: Univariable analysis pinpointed several factors significantly associated with refractoriness/recurrence, including pathological subtype, lactate dehydrogenase (LDH), International Prognostic Index (IPI), treatment, absolute lymphocyte count (ALC), lymphocyte/monocyte ratio (LMR), and prognostic nutritional index (PNI). Binary logistic regression highlighted pathological subtype, LDH, treatment, and ALC as key predictors, which were incorporated into the nomogram. The nomogram showed excellent calibration and accuracy in both cohorts, and comparative DCA and ROC analysis demonstrated its superior net benefit and area under the curve (AUC) compared to traditional indexes like IPI, R-IPI, and NCCN-IPI.

Conclusion: This nomogram serves as a valuable tool for predicting the likelihood of refractoriness or recurrence in DLBCL patients.