Establishment of an Assessment System for the Prediction of Severe Post-operative Complications after Hepatectomy Based on Preoperative Parameters.

Objective: At present, there is no appropriate system to evaluate the severe complications of liver surgery through the preoperative factors. This study aimed to design and verify a risk assessment system for the prediction of severe post-operative complications after a hepatectomy based on the preoperative parameters.

Methods: A retrospective analysis was performed on 1732 patients who had undergone liver surgery. The severity of the complications was graded by Accordion Severity Grading of post-operative complications. The variables were screened by multivariate analysis, and graded scores were assigned to the selected variables. A logistic regression equation was used to form the liver operation risk formula (LORF) for the prediction of severe post-operative complications. The LORF was verified by the receiver operating characteristic (ROC) curve.

Results: The multivariate correlation analysis revealed the independent influencing factors of the severe post-operative complications of liver surgery were Child-Pugh grade (OR=4.127; P<0.001), medical diseases requiring drug treatment (OR=3.092; P<0.001), the number of liver segments to be removed (OR=2.209; P=0.006), organ invasion (OR=4.538; P=0.024), and pathological type (OR=4.023; P=0.002). The binomial logistic regression model was established to obtain the calculation formula (LORF) of the severe complication risk. The area under the ROC curve (AUC) of the LORF was 0.815. The cut-off value of the expected probability of severe complications was 0.3225 (32.25%). Furthermore, in the validation data set, the corresponding AUC of the LORF was 0.829.

Conclusion: As a novel and simplified assessment system, the LORF could effectively predict the severe post-operative complications of liver surgery through the preoperative factors, and therefore it could be used to evaluate the risk of severe liver surgical complications before surgery.