Nomogram development and validation for predicting early death in primary hepatopancreatobiliary non-hodgkin's lymphoma: A population-based retrospective cohort study.

The factors associated with early death in primary hepatopancreatobiliary non-Hodgkin's lymphoma (PHPB-NHL) patients have not been previously investigated. This study aimed to develop and validate a population-based nomogram for early death prediction in PHPB-NHL patients. Data on PHPB-NHL patients diagnosed from 2000 to 2021 were collected from the Surveillance, Epidemiology and End Results database. Patients diagnosed between 2000 and 2015 formed the training cohort, while those from 2016 to 2021 represented the validation cohort. Univariate and multivariate logistic regression analyses identified factors affecting early death, which were used to create a nomogram. The discriminatory ability, calibration, and clinical utility of the nomogram were evaluated using the area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis, respectively. The Surveillance, Epidemiology and End Results database identified 2802 PHPB-NHL patients from 2000 to 2021, with eligible patients divided into a training cohort (1240) and a validation cohort (509). In the training cohort, 29.4% (365/1240) experienced early death. Factors for early death included age, race, primary site, histological type, surgery, radiotherapy, chemotherapy, median household income, and time from diagnosis to treatment. A nomogram based on these factors showed good discrimination (AUC 0.815, 95% CI: 0.789-0.842) and was validated internally (optimism-corrected AUC 0.805, 95% CI: 0.780-0.831). Calibration curves showed strong agreement between predictions and actual outcomes. The validation cohort also showed good discrimination (AUC 0.827, 95% CI: 0.787-0.867) and calibration. Decision curve analysis indicated a positive net benefit. We developed and validated a nomogram to predict early death in PHPB-NHL patients, demonstrating strong predictive accuracy and clinical utility.