The Road Map for ACR Practice Accreditation for Radiology Artificial Intelligence

Area of Science:

• Radiology
• Artificial Intelligence
• Healthcare Quality Management

Background:

• Artificial intelligence (AI) performance in real-world radiology settings often deviates from premarket testing.
• Robust quality management (QM) programs are essential for safe AI implementation in healthcare.
• Current ACR accreditation processes do not encompass AI in radiology.

Purpose of the Study:

• To outline the plan for establishing an ACR accreditation program for AI in radiology.
• To ensure the safe and effective integration of AI into clinical radiology practices.
• To address the need for standardized quality assessment of AI tools.

Main Methods:

• Leveraging the existing ACR accreditation framework, which includes Practice Parameters and Technical Standards.
• Establishing the ACR Recognized Center for Healthcare-AI (ARCH-AI) as a precursor program.
• Gathering insights from ARCH-AI participants on governance, model selection, testing, and monitoring.

Main Results:

• The ARCH-AI program serves as a foundational step towards formal AI accreditation.
• Insights from ARCH-AI will inform the development of comprehensive accreditation criteria.
• Formal ACR accreditation for radiology AI is anticipated by spring 2027.

Conclusions:

• ACR accreditation for AI in radiology is crucial for maintaining high standards of care.
• The proposed program aims to enhance patient safety and optimize AI's clinical benefits.
• Ongoing dialogue among stakeholders will guide the successful implementation of AI accreditation.