Can AI Make Paediatric PET/CT Lower Dose? Opportunities, Evidence and Cautions

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Paediatric Radiology

Background:

• Paediatric PET/CT requires dose optimization to minimize radiation exposure.
• Artificial intelligence (AI) presents potential solutions for reducing radiation doses in paediatric imaging.
• Ensuring diagnostic accuracy and clinical confidence is paramount when implementing dose reduction strategies.

Purpose of the Study:

• To review AI-enabled strategies for optimizing radiation dose in paediatric PET/CT.
• To assess the potential of AI in achieving dose reduction without compromising image quality and diagnostic performance.
• To highlight the importance of validation and ethical considerations for AI in paediatric imaging.

Main Methods:

• Review of AI applications in PET/CT dose reduction, including denoising, image enhancement, synthetic CT, and CT-free correction.
• Examination of AI's role in improving image reconstruction and motion correction.
• Analysis of indirect AI approaches like automated quality control and protocol optimization.

Main Results:

• AI offers direct opportunities in low-count PET enhancement and AI-supported CT dose reduction.
• Synthetic CT and CT-free attenuation correction can potentially replace or reduce the need for CT.
• Indirect AI methods can improve imaging efficiency and reduce repeat scans, thereby lowering overall dose.

Conclusions:

• AI can facilitate lower-dose paediatric PET/CT by enhancing image quality and optimizing protocols.
• Rigorous validation of AI tools for lesion detection, quantitative accuracy, and clinical safety is essential.
• Ethical and social considerations must be addressed to prevent widening health disparities; AI must preserve clinical value within a paediatric-specific framework.