Artificial Intelligence for Clinical Decision Support in Acute Ischemic Stroke: A Systematic Review

Area of Science:

• Neurology
• Medical Informatics
• Artificial Intelligence

Background:

• Established treatment groups for acute ischemic stroke patients.
• Artificial intelligence (AI) methods are being explored to correlate patient characteristics with outcomes for clinical decision support.
• This review focuses on AI-based clinical decision support systems in development for acute ischemic stroke.

Approach:

• Systematic review of English language publications on AI decision support systems for adult acute ischemic stroke.
• Analysis of data sources, outcome measures, and AI methodologies used in these systems.
• Evaluation of reporting standards and potential benefits compared to traditional stroke care.

Key Points:

• A high degree of heterogeneity was observed in data sources, methods, and reporting practices across 121 reviewed studies.
• Sixty-five studies were selected for full data extraction, revealing significant validity threats.
• Dissonance in reporting practices and challenges to clinical translation were identified.

Conclusions:

• Significant validity threats and reporting inconsistencies hinder the clinical implementation of AI in acute ischemic stroke.
• Recommendations are provided for enhancing the successful translation of AI research into clinical practice for stroke diagnosis and treatment.