Whether regional lymph node involvement exerts significant effect on the prognosis still remains obscure for pancreatic neuroendocrine tumors.To clarify this association and identify predictors for lymph node involvem...Whether regional lymph node involvement exerts significant effect on the prognosis still remains obscure for pancreatic neuroendocrine tumors.To clarify this association and identify predictors for lymph node involvement,we studied the data of patients aged>18 years with regional lymph node involvement histologically confirmed pancreatic neuroendocrine tumors from 2004 to 2014 in the Surveillance,Epidemiology,and End Results database(http://seer.cancer.gov/about).We evaluated Lymph node involvement as a prognostic factor by Cox regression.We reduced 9 variables of demographic and tumor characteristics to 5 potential predictors using least absolute shrinkage and selection operator(LASSO)regression model.We further constructed a lymph node involvement model by logistic regression.The model was verified by the verification set,and the visual expression of the model was realized by a nomogram.A total of 1545 cases of pancreatic neuroendocrine tumors were included in our study.Lymph node positivity was significantly associated with disease-specific survival(P<0.001).Younger patients(P<0.05),patients with tumors in the pancreatic head(P<0.05),patients at high American Joint Committee on Cancer T stage(P<0.001),and patients of an undifferentiated status(P<0.05)showed a significantly higher possibility of developing lymph node involvement.The reliability of this model was verified by cross-validation between the training and testing set,and we obtained good discrimination and calibration power.This model also showed great performance in C-index and area under receiver operating characteristic curve.Lymph node positivity was an important negative prognostic predictor for pancreatic neuroendocrine tumor.We developed a lymph node involvement model based on the predictors including age,marital status,primary site,T status,and tumor grade.展开更多
基金This work was supported by the Zhejiang Natural Sciences Foundation of China(No.LY18H160007,LY17H160029,Q17H160042,LQ16H160003)the Zhejiang Medical Innovative Discipline Construction Project-2016.
文摘Whether regional lymph node involvement exerts significant effect on the prognosis still remains obscure for pancreatic neuroendocrine tumors.To clarify this association and identify predictors for lymph node involvement,we studied the data of patients aged>18 years with regional lymph node involvement histologically confirmed pancreatic neuroendocrine tumors from 2004 to 2014 in the Surveillance,Epidemiology,and End Results database(http://seer.cancer.gov/about).We evaluated Lymph node involvement as a prognostic factor by Cox regression.We reduced 9 variables of demographic and tumor characteristics to 5 potential predictors using least absolute shrinkage and selection operator(LASSO)regression model.We further constructed a lymph node involvement model by logistic regression.The model was verified by the verification set,and the visual expression of the model was realized by a nomogram.A total of 1545 cases of pancreatic neuroendocrine tumors were included in our study.Lymph node positivity was significantly associated with disease-specific survival(P<0.001).Younger patients(P<0.05),patients with tumors in the pancreatic head(P<0.05),patients at high American Joint Committee on Cancer T stage(P<0.001),and patients of an undifferentiated status(P<0.05)showed a significantly higher possibility of developing lymph node involvement.The reliability of this model was verified by cross-validation between the training and testing set,and we obtained good discrimination and calibration power.This model also showed great performance in C-index and area under receiver operating characteristic curve.Lymph node positivity was an important negative prognostic predictor for pancreatic neuroendocrine tumor.We developed a lymph node involvement model based on the predictors including age,marital status,primary site,T status,and tumor grade.
