Volume 29, Issue 5, 2020
DOI: 10.24205/03276716.2020.1017
Prognostic Nomograms for Predicting Survival in Gastric Carcinoma Patients with History of a Prior Malignancy
Abstract
Objective: Prior malignancy is a frequent exclusion criterion in gastric carcinoma (GC) trials. However, the impact of prior malignancy on GC prognosis is not known. We aim to investigate the impact of history of a prior malignancy on overall survival (OS) and cancer-specific survival (CSS) in patients with GC, and develop nomograms for prognostic prediction of these patients using SEER database.
Methods: A total of 12676 histologically confirmed GC patients were obtained from the SEER database. Propensity-score matching (PSM) was performed to reduce potential selective bias. All matched patients were randomly divided into development (n=1155) and validation (n=1155) cohorts. We used univariate and multivariate analyses to identify the independent variables for OS and CSS. A nomogram was established based on independent prognostic factors. The prognostic performances of nomograms were validated using the concordance index (C-index) and calibration plots both internally and externally. The clinical usefulness of nomograms was compared using decision curve analysis (DCA) between nomograms and American Joint Committee on Cancer (AJCC) 7th staging system.
Results: Two nomograms were built based on the independent variables. In the development cohort, the C-indexes for the constructed nomogram to predict OS and CSS were 0.699 and 0.744, respectively. The nomograms achieved favorable discriminative ability in the validation cohort to predict OS and CSS, with C-indexes of 0.698 and 0.744, respectively, which were higher than C-indexes of the AJCC 7th staging system. The calibration plots displayed good agreement between nomogram-predicted survival probability and the actual observed outcomes. Furthermore, DCA indicated that the nomograms offered advantage over the AJCC staging system with bringing more clinical net benefit.
Conclusions: The novel proposed nomograms based on history of a prior malignancy can more effectively prediction the individualized probability of OS and CSS in patients with GC than the AJCC staging system, and the predictive power can help clinicians formulate suitable individual treatment and conduct personalized prognostic evaluation.
Keywords
nomogram; prior malignancy; gastric carcinoma; survival