Abstract

e19069 Background: The prognostic factors influencing survival and optimal management of primary pulmonary lymphomas (PPLs) have not been clearly defined due to rarity of the disease. This study sought to characterize the significant prognostic factors of PPL and develop a prognostic nomogram for individualized prediction of survival outcomes in patients with PPL. Methods: Patients diagnosed with PPL between 1983 and 2010 were identified using the Surveillance, Epidemiology, and End Results (SEER) Program database. Kaplan-Meier survival analysis and Cox proportional hazards regression model were performed to identify significant independent prognostic factors for overall survival (OS) in patients with PPL. A nomogram was constructed for the prediction of 5-year OS probability based on these variables. Results: The study cohort of 2325 PPL patients has a 5-year OS rate of 52% and a median OS of 67 months. Older age at diagnosis (HR 1.031; 95% CI, 1.026-1.036; p<0.001], males (HR 1.434; 95% CI, 1.267-1.622; p<0.001), Hispanic race (HR 1.370; 95% CI, 1.113-1.687; p=0.003), non-marginal zone B-cell lymphoma of mucosa associated lymphoid tissue (MALT) histology (p<0.001), primary site at main bronchus (HR 1.326; 95% CI, 1.014-1.735; p=0.039), Ann Arbor stage IV (HR 1.542; 95% CI, 1.324-1.795; p<0.001) were significantly associated with worse OS. All treatment modalities, including chemotherapy (HR 0.615; 95% CI, 0.536-0.706; p<0.001), surgery (HR 0.666;95% CI, 0.577-0.769; p<0.001) and radiotherapy (HR 0.829; 95% CI, 0.693-0.992; p=0.041) were independent predictors of survival. The nomogram constructed using these variables has a higher concordance index of 0.716 (95% CI, 0.699-0.734) compared to that of Ann Arbor staging system [0.571 (95% CI,0.552-0.591); p<0.001]. Conclusions: Older age, male sex, Hispanic race, non-MALT histology, primary site at main bronchus, advanced Ann Arbor stage, not receiving treatment were independent prognostic factors that are associated with worse OS in patients with PPL. The nomogram demonstrated good agreement between the predicted probabilities and actual observations on calibration plots.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call