Objective:

To map the potential for AI to automate, improve, or provide new insights in diagnostic imaging and interventional workflows, ensuring safe and equitable implementation by 2030.

Key Findings:

• Over 75% of FDA-approved AI devices are in radiology, with 1,247 AI-enabled devices authorized by August 2025.
• AI applications in MRI and CT have shown improved diagnostic accuracy and workflow efficiency.
• 70% of workflow steps in MRI-based cancer screening currently have available AI solutions.

Interpretation:

AI technologies are evolving from experimental tools to clinically validated solutions, enhancing diagnostic accuracy and streamlining workflows, while necessitating ongoing human oversight and validation.

Limitations:

• The real-world evidence base for AI remains heterogeneous and context-dependent, impacting generalizability.
• Reported error profiles differ between AI and clinicians, necessitating a collaborative approach to mitigate risks.

Conclusion:

AI has the potential to significantly enhance diagnostic imaging and interventional procedures, but its integration must be carefully managed to ensure safety and effectiveness.