The application of deep learning-based artificial intelligence algorithms combined with low-dose scanning protocols...

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Low-dose scanning protocols in computed tomography (CT) can impact image quality.
• Different reconstruction algorithms yield varying results under low-dose conditions.

Purpose of the Study:

• To evaluate the effectiveness of the precise imaging (PI) deep learning artificial intelligence (AI) algorithm combined with low-dose scanning for chest CT.
• To assess the impact on image quality and radiation dose.

Main Methods:

• Retrospective analysis of 100 chest CT scans.
• Comparison of three groups: PI AI with low-dose (Group A), conventional iterative reconstruction with low-dose (Group B), and conventional iterative reconstruction with standard-dose (Group C).
• Objective (SNR, CNR, SD) and subjective image quality assessments, alongside radiation dose metrics (CTDIvol, DLP, ED).

Main Results:

• Group A demonstrated significantly lower noise (SD) and higher SNR/CNR compared to Group B (P<0.01).
• No significant differences in image quality metrics between PI AI low-dose (Group A) and standard-dose (Group C) scans (P>0.05).
• Radiation dose (CTDIvol, DLP, ED) was reduced by 81% in the experimental group (A+B) compared to the control group (C).

Conclusions:

• The PI deep learning AI algorithm with low-dose scanning achieves high image quality comparable to standard-dose protocols.
• This combination effectively reduces radiation exposure in chest CT.
• The PI AI algorithm shows promise for widespread application in chest CT screening.