'Carrusel': an expert system for vestibular diagnosis

Area of Science:

• Otolaryngology
• Neuroscience
• Medical Informatics

Background:

• Neuro-otology relies heavily on deductive reasoning and patient history for diagnosing vestibular disorders.
• Despite technological advancements, subjective patient reports of vertigo and dizziness remain crucial for accurate diagnosis.
• Artificial intelligence (AI) has been explored as a tool to assist in complex clinical problem-solving.

Purpose of the Study:

• To evaluate the diagnostic performance of an AI-powered expert system in neuro-otology.
• To compare the efficacy of an AI system against human expert diagnosis in vestibular disorders.
• To highlight the potential of AI in enhancing diagnostic accuracy within specialized medical fields.

Main Methods:

• Development of "Carrusel," an expert system utilizing Prolog for diagnosing vestibular disorders.
• Systematic comparison of the diagnostic outcomes generated by "Carrusel" against those of human experts.
• Evaluation of diagnostic success rates based on established clinical criteria.

Main Results:

• The "Carrusel" expert system demonstrated a high diagnostic success rate of 97%.
• This success rate was achieved when benchmarked against the diagnoses provided by the human experts involved in its development.
• The findings indicate a significant level of accuracy for the AI system in the context of neuro-otological diagnosis.

Conclusions:

• The "Carrusel" expert system shows considerable promise as an aid in the diagnosis of vestibular disorders.
• AI-driven diagnostic tools can achieve high accuracy, complementing traditional clinical methods in neuro-otology.
• The study underscores the potential for AI to support clinicians in complex diagnostic scenarios, improving patient care.