Comprehensible knowledge model creation for cancer treatment decision making

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Clinical Decision Support

Background:

• Electronic health records (EHRs) contain vast clinical data valuable for developing knowledge-based systems.
• Automated knowledge acquisition from EHRs is crucial for enhancing clinical decision-making and education.
• Current limitations in automated knowledge acquisition hinder the development of effective clinical support systems.

Purpose of the Study:

• To propose an automated knowledge acquisition methodology with a comprehensible knowledge model for cancer treatment (CKM-CT).
• To enable automated data acquisition from clinical documents and ensure data quality through error correction and standardization.
• To develop predictive models for clinical decision support and education.

Main Methods:

• The CKM-CT methodology was employed for automated knowledge acquisition from clinical documents.
• Data preprocessing included dimensionality reduction and missing value imputation.
• The Classification and Regression Trees (CRT) algorithm was selected for predictive modeling.

Main Results:

• Data from 13,788 head and neck cancer (HNC) documents were analyzed.
• The CRT algorithm achieved 69.0% accuracy in predicting HNC treatment plans.
• 11 decision paths and 11 decision rules were generated for treatment prediction.

Conclusions:

• The CKM-CT methodology effectively extracts hidden knowledge from clinical documents.
• The developed prediction models assist and educate clinicians in informed decision-making.
• The methodology is generalizable to other cancer domains, such as breast cancer.