Development of a LASSO machine learning algorithm-based model for postoperative delirium prediction in hepatectomy...

Area of Science:

• Hepatobiliary Surgery
• Critical Care Medicine
• Medical Informatics

Background:

• Postoperative delirium (POD) is a significant complication following hepatectomy.
• Accurate prediction of POD risk is crucial for patient management and resource allocation.

Purpose of the Study:

• To develop and validate a clinically applicable nomogram for predicting the risk of delirium after hepatectomy.
• To identify independent risk factors associated with POD in hepatectomy patients.

Main Methods:

• Least Absolute Shrinkage and Selection Operator (LASSO) regression was used to identify risk factors.
• A nomogram model was developed and validated using R software.
• Receiver Operating Characteristic (ROC) curve analysis and Decision Curve Analysis (DCA) were employed for performance evaluation.

Main Results:

• Key predictors identified include Ramelteon, Age, Sex, Alcohol use, Viral status, Cardiovascular disease, ASA class, Total bilirubin, Prothrombin time, Laparoscopic approach, and Blood transfusion.
• The nomogram achieved an Area Under the Curve (AUC) of 0.854 for prediction accuracy.
• The model demonstrated high sensitivity (91.9%) and good specificity (68.8%) at the optimal cutoff.

Conclusions:

• LASSO-based regression effectively constructed a nomogram for predicting post-hepatectomy delirium.
• The developed nomogram is clinically valid and useful for predicting POD risk in hepatectomy patients.