Thyroid disorders: automatic diagnosis in CONSULT I

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Diagnostic Systems

Background:

• Thyroid disease diagnosis relies on complex interpretation of clinical signs, symptoms, and laboratory tests.
• Computer-assisted diagnosis aims to improve accuracy and efficiency in medical decision-making.
• Existing diagnostic methods can be time-consuming and prone to human error.

Purpose of the Study:

• To implement and evaluate an automated system for thyroid disease diagnosis.
• To assess the performance of the CONSULT I microcomputer system in classifying thyroid diseases.
• To compare the diagnostic accuracy of the automated system with that of physicians.

Main Methods:

• Development of a thyroid disease diagnostic subsystem within the CONSULT I framework.
• Utilizing the Patrick model for computer-assisted diagnosis.
• Employing 16 features (signs, symptoms, lab tests) to classify 19 thyroid disease classes.
• Analysis of 76 patient records as recognition samples.

Main Results:

• The automated system correctly identified the true thyroid disease in 89% of test cases based on the highest probability.
• The system achieved 100% accuracy in providing a differential diagnosis.
• Performance metrics were established for comparison with human physician diagnoses.

Conclusions:

• The CONSULT I system demonstrates high accuracy in the automatic diagnosis of thyroid diseases.
• Automated diagnostic systems show potential for supporting clinical decision-making in endocrinology.
• Further research can explore probability density estimation for enhanced diagnostic accuracy.