[Evaluation of the Latest Motion Correction Techniques in Periodically Rotated Overlapping ParallEL Lines with Enhanced Reconstruction (PROPELLER) Imaging across Different Vendors]

Abstract

Purpose: To evaluate the robustness of the latest periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) technology from each vendor against head movements and to investigate their characteristics for effective clinical use.

Methods: Using a phantom simulating the T2-weighted image of the human brain, images were acquired with devices from CANON MEDICAL SYSTEMS (Tochigi, Japan; hereinafter "Canon"), GE HealthCare (Chicago, IL, USA; hereinafter "GE"), Philips (Amsterdam, Netherlands), and Siemens Healthineers (Forchheim, Germany; hereinafter "SIEMENS"). The head motion patterns were divided into rotation angle dependency (single rotation and multiple rotations) and rotation frequency dependency and evaluated using structural similarity (SSIM).

Results: For rotation angle dependency, Canon was robust against small rotation angles and fine movements. Despite the rotation angle, GE was robust against movements, with deep learning reconstruction (DLR) improving correction functionality. Philips could be used with compressed sensitivity encoding (CS), and robustness varied with blade width. SIEMENS was robust against large movements. For rotation frequency dependency, results were similar across the 4 vendors.

Conclusion: The rotation angle and rotation frequency dependencies of the PROPELLER technology from the 4 vendors were quantitatively evaluated. Understanding the characteristics of PROPELLER allows for the possibility of providing diagnostic-quality images even for patients who move during head MRI exams by appropriately using PROPELLER.

Keywords: PROPELLER; brain; compressed SENSE; deep learning reconstruction; motion correction.