Clinical Characteristics Of 1180 Cases Of Primary Hepatic Carcinoma And Analysis Of Related Factors Of Survival Time Based On Bayesian Network Model

Objective:1 Through retrospective analysis,to investigate the Clinical Characteristics of a cohort of Chinese patients with Primary hepatic carcinoma(PHC),and to determine the clinical laboratory indicators and liver stiffness characteristics of PHC in different stages of Barcelona clinic liver cancer.2 To use non-conditional logistics regression model to analyze the influencing factors of the survival time of PHC,and to use Bayesian network model to analyze the complex network relationship among influencing factors of the survival time of PHC.Methods:Details of PHC referred from October 2014 to November 2017 were collected retrospectively.Demographic data,etiology,clinical and biochemical details were retrospectively analyzed,the changes of liver stiffness in different etiologies and different stages of Barcelona clinic liver cancer were especially analyzed.The main factors affecting the survival time of PHC were screened out based on the logistic regression analysis,the Bayesian network topology was constructed to directly reflect the potential relationship among multiple factors,and probabilistic reasoning was used to reflect one or more factors.Results:The onset age of PHC was 60.33±9.11(range 24-84)years,9 cases were ≤ 40 years;572 cases were 41-60 years,599 cases were ≥ 61 years.The proportion of males was higher than that of females,and the proportion of female PHC patients ≥ 61 years old is higher than that of 41-60 years old.Majority of PHC(n=787;66.69%)had HBV infection,second commonest cause was HCV infection(n=217;18.39%).More patients with HBV infection were 41-60 years(50.19%),and more patients with HCV infection were ≥ 61 years(63.32%).There was no statistical difference in etiology,age,gender and distribution of diabetes mellitus among different Barcelona clinicliver cancer stages(P>0.05).The overall PHC was found to be positively correlated with alkaline phosphatase,γ-glutamyltransferase,alpha-fetoprotein and liver stiffness measurement values from stage A to stage D(P<0.05).ANOVA analysis showed that the overall liver stiffness measurement among the four BCLC stages was found to be statistically significant different in HBV-associated and HCV-associated PHC patients.Logistic regression screened factors that may be related to the survival time of patients with PHC.Bayesian network structure showed the age,platelet count,tumor size,tumor number,portal vein invasion,extrahepatic metastasis,andγ-glutamyltransferase of PHC patients are all related to the survival time of PHC patients.Conclusion:Majority(99.24%)were patients aged > 40 years old.Male was a high incidence population.In etiological analysis,HBV dominated PHC occurrence,HBV-,HCVand alcohol-associated PHC had distinct clinical and biochemical characteristics,necessitating different screening policies to optimize PHC surveillance and management.We could conclude from the Bayesian network structure that patients with tumor size > 3cm and γ-GT> 40U/L accompanied by extrahepatic metastasis have the highest probability of survival time <30 months.This conclusion has important guiding significance for the prediction and evaluation of PHC clinical survival time.