Leveraging artificial intelligence for equitable women's health outcomes through imaging

Area of Science:

• Women's health research
• Medical imaging analysis
• Artificial intelligence applications

Background:

• Women face a disproportionate disease burden, including cardiovascular disease, cancer, and maternal complications.
• Current diagnostic tools and risk models often overlook sex-specific factors, leading to care delays.
• Artificial intelligence (AI) presents a transformative opportunity to enhance women's health outcomes.

Purpose of the Study:

• To review AI applications in radiology for women's health.
• To focus on diseases unique to women and those affecting both sexes, emphasizing women's outcomes.
• To identify and recommend mitigation strategies for AI-related biases in women's health.

Main Methods:

• Literature review of AI applications in radiology for women's health.
• Analysis of AI's impact on screening, early detection, and risk prediction.
• Examination of biases affecting AI in women's health and proposed solutions.

Main Results:

• AI demonstrates significant potential in improving diagnostic accuracy and risk prediction for various conditions in women.
• Identified biases in AI algorithms can exacerbate existing health disparities.
• Sociotechnical approaches are crucial for equitable AI development and deployment in women's health.

Conclusions:

• AI integration in radiology can accelerate care and improve outcomes for women.
• Addressing and mitigating biases is essential for equitable AI implementation.
• A sociotechnical framework is recommended for AI development to overcome systemic disparities in women's health.