AUTHOR=Gu Yanmei , Deng Haixiao , Wang Daijun , Li Yumin 

TITLE=Metastasis Pattern and Survival Analysis in Primary Small Bowel Adenocarcinoma: A SEER-Based Study

JOURNAL=Frontiers in Surgery

VOLUME=Volume 8 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/surgery/articles/10.3389/fsurg.2021.759162

DOI=10.3389/fsurg.2021.759162

ISSN=2296-875X

ABSTRACT=Background: Small bowel adenocarcinoma (SBA) is a rare gastrointestinal tumor with high malignancy. The aim of this study was to comprehensively assess the distant metastasis pattern of SBA and establish nomograms predicting survival.
Methods: From 2010 to 2015, patients diagnosed with SBA were identified based on the Surveillance, Epidemiology and End Results (SEER) database. To compare survival differences between metastasis pattern, Kaplan Meier survival analysis was conducted. Then, univariate and multivariate cox analysis was applied to identify independent prognostic factors of cancer-specific survival (CSS) and overall survival (OS), and identify risk factors for metastasis of SBA. To assess the discrimination and calibration of nomograms, the concordance index (C-index), calibration curves, receiver operating characteristic curve (ROC) and decision curve analysis (DCA) were calculated. 
Results: Kaplan Meier curves revealed that metastasis pattern were significantly correlated with CSS (p<0.001) and OS (p<0.001). Cox regression analysis indicated that metastasis pattern was an independent prognostic factor of OS and CSS in SBA patients, as well as age, grade, T stage, N stage, surgery, retrieval of regional lymph nodes and chemotherapy. Combining these factors, we constructed prognostic nomograms, which suggested that the metastasis pattern contributed the most to the survival of SBA patients. Nomograms for OS and CSS had C-index of 0.787 and 0.793, respectively. Internal and external calibration curves showed excellent agreement between probability and actual observation. DCA also indicated its clinical value with an improved net benefit. In addition, the models we constructed had better prognostic accuracy and clinical utility than traditional TNM staging based on C-index and ROC. Further Cox regression analysis showed that old age, poor differentiation, N2 and not receiving chemotherapy were risk factors for prognosis in patients with metastatic SBA。
Conclusion: Metastasis pattern was an independent prognostic factor, which exhibited the greatest predictive effect of OS and CSS for SBA patients. Adjuvant chemotherapy had a positive effect on the survival of SBA patients. Nomograms to predict 3- and 5- year OS and CSS of SBA patients were constructed, which could identify patients with higher risk and might be superior predict survival of patients with SBA than TNM staging.