Speed and efficiency: evaluating pulmonary nodule detection with AI-enhanced 3D gradient echo imaging.

Area of Science:

• Radiology and Medical Imaging
• Artificial Intelligence in Medicine
• Pulmonary Medicine

Background:

• Pulmonary MRI traditionally has long scan times, limiting its use in lung cancer screening.
• Accelerated imaging techniques are crucial for improving the efficiency and feasibility of MRI-based lung nodule detection.

Purpose of the Study:

• To evaluate the diagnostic feasibility of accelerated pulmonary MRI for detecting and characterizing pulmonary nodules.
• To assess the impact of artificial intelligence-aided compressed sensing (AI-CS) on MRI scan time and image quality.

Main Methods:

• Prospective trial comparing chest CT and pulmonary MRI in 37 patients with lung nodules.
• Pulmonary MRI utilized a 3D gradient echo sequence accelerated with parallel imaging, compressed sensing, and deep learning reconstruction (CS-AI factors 7, 10, 15).
• Two blinded readers assessed image quality, nodule detection, size, and morphology against CT reference standard.

Main Results:

• Scan times were reduced to as low as 1:50 min (CS-AI-15) with diagnostic image quality maintained even at higher acceleration.
• Pulmonary nodule detection rates were high across all factors (96.8%–100%).
• Nodule morphology characterization was comparable to CT, with minimal deviation in size (<1 mm).

Conclusions:

• AI-aided compressed sensing significantly reduces pulmonary MRI scan times while preserving diagnostic accuracy for nodule detection and characterization.
• This accelerated MRI technique shows promise for efficient and effective lung cancer screening.
• Integrating CS and AI into pulmonary MRI enhances efficiency without compromising diagnostic quality.