A Nomogram Combining Clinical, Metabolic, and Serum Markers to Predict the Risk of Primary Breast Lymphoma Recurrence

Background Primary breast lymphoma (PBL) is a rare yet aggressive extranodal non-Hodgkin lymphoma characterized by a high recurrence rate. This study aimed to develop and validate a nomogram integrating clinical features, ¹⁸ F-FDG PET/CT metabolic parameters, and serum biomarkers to individually predict postoperative recurrence risk in PBL patients. Methods We retrospectively analyzed 44 patients with pathologically confirmed PBL, categorizing them into recurrence and non-recurrence groups. Clinical factors, PET/CT metabolic parameters (SUVmax, MTV, TLG), and serum biomarkers (β2-microglobulin [β2-MG], LDH, Ki-67) were compared. Variables with P  < 0.05 in univariate analysis were incorporated into multivariate logistic regression to identify independent predictors. A predictive nomogram was constructed and evaluated using ROC curves, calibration curves, and decision curve analysis (DCA). Results Significant differences were observed in tumor multiplicity, TLG, and β2-MG levels ( P  < 0.05). Multivariate analysis identified tumor multiplicity (OR = 11.732, P = 0.047), TLG (per unit increase, OR = 1.029, P = 0.036), and β2-MG level (OR = 1.002, P  = 0.042) as independent predictors. The comprehensive model integrating all three predictors demonstrated superior performance (AUC = 0.921, sensitivity 88.9%, specificity 84.6%). The nomogram showed high calibration accuracy and substantial net clinical benefit on DCA. Conclusion Tumor multiplicity, elevated TLG, and elevated β2-MG were independent predictors of recurrence in PBL. The constructed nomogram effectively integrates these factors, providing a practical tool for individualized recurrence risk assessment.