A comparison between manual and artificial intelligence-based automatic positioning in CT imaging for COVID-19...

Area of Science:

• Radiology and Medical Imaging
• Artificial Intelligence in Healthcare
• Thoracic Imaging

Background:

• Conventional manual positioning (MP) in CT scans can lead to variability in radiation dose and image quality.
• Optimizing imaging workflow and reducing radiation exposure are critical in managing COVID-19 patients.

Purpose of the Study:

• To compare the imaging workflow, radiation dose, and image quality between manual positioning (MP) and AI-based automatic positioning (AP) for COVID-19 patients undergoing chest CT.
• To evaluate the efficacy of AI in improving CT scan procedures.

Main Methods:

• A study involving 127 adult COVID-19 patients comparing MP and AP methods for chest CT scans.
• Recorded and compared radiation dose, patient positioning time, off-center distance, image noise, and signal-to-noise ratio (SNR).

Main Results:

• AI-based automatic positioning (AP) was successful in all patients, reducing total positioning time by 28% compared to MP.
• AP resulted in significantly less off-center distance (1.56 cm vs. 4.05 cm), higher positioning accuracy (99% vs. 92%), a 16% reduction in radiation dose, and 9% less image noise in the erector spinae.
• AP also improved image quality with lower noise and higher SNR in pulmonary peripheral lesions.

Conclusions:

• AI-based automatic positioning is a promising technique for enhancing CT imaging of the chest.
• AP optimizes imaging workflow, reduces radiation dose, and improves image quality for COVID-19 patients.
• The AI method offers significant advantages over traditional manual positioning in clinical practice.