Risk prediction of second primary malignancies in primary colorectal neuroendocrine neoplasms patients: a population-based study.

Abstract

In this study, we aimed to identify risk factors for developing second primary malignancies (SPMs) in colorectal neuroendocrine neoplasms (NENs) patients and develop a competing-risk nomogram to predict SPMs' probabilities quantitatively. Patients with colorectal NENs were retrospectively collected from the Surveillance, Epidemiology, and End Results (SEER) database during 2000-2013. Potential risk factors for SPMs' occurrence in colorectal NENs' patients were identified by the Fine and Gray's proportional sub-distribution hazards model. Then, a competing-risk nomogram was constructed to quantify SPMs' probabilities. The discriminative abilities and calibrations of this competing-risk nomogram were assessed by the area under the receiver-operating characteristic (ROC) curves (AUC) and calibration curves. We identified 11,017 colorectal NENs' patients, and randomly divided them into training (n = 7711 patients) and validation (n = 3306 patients) cohorts. In the whole cohort, 12.4% patients (n = 1369) had developed SPMs during the maximum follow-up of approximately 19years (median 8.9years). Sex, age, race, primary tumor location, and chemotherapy were identified as risk factors for SPMs' occurrence in colorectal NENs' patients. Such factors were selected to develop a competing-risk nomogram and showed excellent predictive ability for SPMs' occurrence (the 3-, 5-, and 10-year AUC values were 0.631, 0.632, and 0.629 in the training cohort and 0.665, 0.639, 0.624 in the validation cohort, respectively). This research identified risk factors for SPMs' occurrence in colorectal NENs' patients. Competing-risk nomogram was constructed and proved to have good performance.