Development and validation of a prognostic nomogram for gallbladder papillary adenocarcinoma.

Abstract

Gallbladder papillary adenocarcinoma (GBPA) is an uncharacteristically gallbladder cancer subtype. Although some studies have shown that the prognosis of GBPA patients is significantly better than that of gallbladder adenocarcinoma (GBA) and gallbladder mucinous adenocarcinoma (GBMA) due to its rarity, there is a lack of large sample studies necessary to confirm the clinical characteristics and survival rate of GBPA. Therefore, this study aimed to describe the clinicopathological characteristics affecting survival in GBPA. This data was then used to establish a prognostic nomogram for GBPA. The data of patients diagnosed with gallbladder cancer between 2004 to 2015 were extracted from the Surveillance, Epidemiology, and End Results (SEER) database. The clinical features and survival of patients with GBPA were compared with those of GBA and GBMA after balancing the baseline characteristics using propensity score matching (PSM). Univariate and multivariate Cox analyses were used to identify the prognostic factors for GBPA. Subsequently, the overall survival (OS) and cancer-specific survival (CSS) nomograms were established to predict GBPA prognosis. The performance and discrimination of the nomogram were measured using concordance index (C-index), calibration curves, receptor operating characteristic curves(ROC), and decision curve analysis (DCA) was applied to examine the net benefit of tients with GBPA, 5798 patients with GBA, and 223 patients with GBMA. The mean 1-, 3- and 5-year OS rates for GBPA were 81.3%, 58.8%, and 49.1%, respectively, while the mean 1-, 3- and 5-year CSS rates were 85.0%, 68.1%, and 61.0%, respectively. The median OS rates was 58 months (95% CI: 43-88), while the median CSS was not reached. The PSM analysis showed a differ statistically significantly in the OS between GBPA and GBA. However, there has no statistically difference in CSS. Conversely, the OS and CSS between GBPA and GBMA have statistically significant differences. Age, marital, T stage, and M stage were strongly linked to the prognosis for OS, while T-stage, M-stage, and surgery were significantly associated with the prognosis for CSS in GBPA patients. The AUC for the 1-, 3-, and 5-year OS were 0.722 (95%CI: 0.630-0.813), 0.728 (95%CI: 0.665-0.790), and 0.706 (95%CI: 0.641-0.771), respectively. The AUC for the 1-, 3-, and 5-year CSS were 0.749 (95%CI: 0.659-0.840), 0.698 (95%CI: 0.627-0.770), and 0.665 (95%CI: 0.594-0.735), respectively. The C-indices for the OS and CSS nomograms were 0.701 (95% CI: 0.634-0.744) and 0.651 (95% CI: 0.598-0.703), respectively. The calibration curves showed that the nomograms were well consistency. The DCA showed that compared with the TNM system, the nomogram models had a significant positive net benefit in survival prediction. GBPA has distinct clinicopathological characteristics and survival compared to other gallbladder carcinomas. The established nomogram provided a better prediction of survival for GBPA patients than the traditional TNM models.