Artificial intelligence (AI) and CT in abdominal imaging: image reconstruction and beyond

Area of Science:

• Radiology and Medical Imaging
• Artificial Intelligence in Healthcare
• Medical Image Processing

Background:

• Computed tomography (CT) is essential for abdominal diagnosis, treatment planning, and monitoring.
• Traditional reconstruction methods (FBP, IR) have limitations like noise and artificial texture.
• Artificial intelligence (AI) offers advanced solutions for medical imaging challenges.

Purpose of the Study:

• To review the principles, advancements, and future directions of AI-driven CT image reconstruction in abdominal imaging.
• To highlight the benefits of deep learning-based reconstruction (DLR) over traditional methods.
• To explore AI's expanding role beyond reconstruction in abdominal radiology.

Main Methods:

• Review of recent advances in vendor-specific and vendor-agnostic DLR algorithms (e.g., TrueFidelity, AiCE, Precise Image).
• Analysis of DLR's impact on image quality metrics (contrast-to-noise ratio) and diagnostic performance.
• Exploration of AI applications in lesion detection, quantitative imaging, and workflow optimization.

Main Results:

• DLR significantly improves image quality, reduces radiation dose, and enhances workflow efficiency compared to FBP and IR.
• AI-based DLR algorithms demonstrate improved contrast-to-noise ratio and lesion detection in abdominal organs.
• AI applications enhance radiologists' efficiency and diagnostic accuracy in abdominal imaging.

Conclusions:

• AI-driven DLR represents a significant advancement in abdominal CT imaging, overcoming limitations of traditional methods.
• DLR enhances diagnostic confidence and accuracy for various abdominal pathologies.
• Further clinical validation, standardization, and adoption are crucial for realizing AI's full potential in abdominal radiology.