Development and validation of a population-based prognostic nomogram for primary colorectal lymphoma patients

BACKGROUND: Primary colorectal lymphoma (PCL) is a relatively rare cancer type, constituting 15%–20% of primary gastrointestinal lymphoma and <1% of all colorectal malignancies. Given its low incidence, standard guidelines for case management are not available. This large population-based study aims to construct a nomogram to predict survival outcomes and to help tailor individualised treatment decisions in patients with PCL. METHODS: A retrospective cohort study of patients with PCL was developed using data registered in the Surveillance, Epidemiology, and End Results (SEER) database between 1990 and 2015. The prognostic nomogram was constructed using R software after univariate and multivariate Cox regression analyses. Cox regression models were assessed using the proportional hazards (PH) assumption. Kaplan−Meier survival analysis was used to analyze survival outcomes. The 1-, 3-, 5-, and 10-year area under the curve (AUC) values of ROC (receiver operating characteristic) curves, the concordance index (C-index), and calibration curves were calculated to verify the predictive performance of the nomogram. RESULTS: The final nomogram included age, Ann Arbor stage, histology, location, marital status, and treatment, all of which had an important effect on overall survival (OS). The discrimination of the nomogram revealed good prognostic accuracy and clinical applicability as indicated by C-index values of 0.713 and 0.711 in the training and validation cohorts, respectively. Kaplan−Meier survival curves were significantly different for distinct conditions. CONCLUSION: This study developed and validated a six-factor nomogram for predicting PCL patient prognosis. This nomogram might be useful for risk stratification and making better individualised decisions for PCL patients.