An intelligent model for liver disease diagnosis

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Hepatology

Background:

• Liver disease is a leading cause of mortality, often diagnosed late.
• Accurate early diagnosis is crucial for effective liver disease treatment.
• Developing intelligent diagnosis models is essential for improving patient outcomes.

Purpose of the Study:

• To develop an intelligent diagnosis model for liver disease using Classification and Regression Trees (CART) and Case-Based Reasoning (CBR).
• To enhance the accuracy of liver disease diagnosis and classification.
• To provide a comprehensive analytic framework for early detection and treatment.

Main Methods:

• Utilized data from 510 outpatients diagnosed with liver conditions (ICD-9 codes) from 2005-2006.
• Applied CART for initial diagnosis of liver disease presence (340 patients for model development).
• Employed CBR for diagnosing the specific type of liver disease in confirmed cases (170 patients for comparative analysis).

Main Results:

• CART achieved a diagnostic accuracy rate of 92.94% in identifying liver disease.
• CBR demonstrated a diagnostic accuracy rate of 90.00% in classifying liver disease types.
• The integrated CART-CBR model proved effective in examining liver diseases with considerable accuracy.

Conclusions:

• The intelligent diagnosis model effectively integrates CART and CBR for accurate liver disease diagnosis and classification.
• The model serves as a valuable supporting system for physicians in clinical decision-making.
• Extracted CART rules and CBR case retrieval assist in reducing diagnostic errors and improving treatment quality.