Development and validation of a nomogram to predict overall survival in elderly patients with pancreatic cancer without surgical treatment

Elderly patients with pancreatic cancer (PC) not treated by surgery have extremely low overall survival (OS), and the aim of this study is to investigate the prognostic factors of elderly patients with PC not treated by surgery and to establish and validate a nomogram to predict their OS. Data on PC patients over 65 years of age who were not treated surgically were downloaded from the Surveillance, Epidemiology, and End Results database from 2010 to 2015. Patients from 2010 to 2014 were randomized into a training group (n = 2044) and a validation group (n = 877). Patients from 2015 (n = 662) were used for external validation. Cox analysis was performed to determine prognostic factors, and a nomogram was constructed to predict 1-, 2-, and 3-year OS in elderly PC patients without surgical treatment. The nomogram was validated using the area under the receiver operating characteristic curve (AUC) and the calibration curve. Patients were categorized into low- and high-risk groups based on the median risk score for Kaplan-Meier survival analysis. Age, tumor size, grade, radiotherapy, and chemotherapy were independent prognostic factors affecting OS, and a nomogram was constructed. In the training cohort, the AUCs of the predicted nomogram for 1, 2, and 3 years were 0.725, 0.738, and 0.741, and in the validation cohort, the AUCs of the predicted nomogram for 1, 2, and 3 years were 0.716, 0.733, and 0.731, while in the test cohort, the AUCs of the predicted nomogram for 1, 2, and 3 years were in the test cohort were 0.777, 0.794, and 0.813, respectively. The calibration curves showed that the actual OS at 1, 2, and 3 years was in good agreement with the predicted OS. In addition, the OS of patients in the low-risk group was significantly better than that of high-risk patients. We developed a new nomogram to assess the prognosis of elderly patients with non-operatively treated PC and to develop clinical treatment strategies.