Novel Magnetic Resonance Imaging Methodology for Dynamic Visualization of Respiratory Thoracic Motion: A Pilot Feasibility Study

Masaaki Kobayashi^1,2Hideyuki Fukubayashi¹Katsuma Iwai¹Kazumo Miura²Akio Yamamoto²Kumiko Ono²Akinori Miki²Takashi Ueguchi^2,3,4,5Akira Ishikawa^2*

• ¹Kobe Co-Medical College, kobe, Japan
• ²Graduate School of Health Sciences, Kobe University, Kobe, Hyōgo, Japan
• ³Center for Information and Neural Networks, Advanced ICT Research Institute, National Institute of Information and Communications Technology, Suita, Miyagi, Japan
• ⁴Department of Radiation Oncology, Osaka University Graduate School of Medicine, Suita, Miyagi, Japan
• ⁵Department of Medical Imaging Science, Osaka University Graduate School of Medicine, Suita, Miyagi, Japan

Introduction: Positional management is important in respiratory rehabilitation. Current magnetic resonance (MR) imaging techniques for visualizing respiratory mechanics are limited by external pressure from receiver coils or spatial restrictions within the bore, and there is no established method for visualizing respiratory movements in the semi-prone position. Therefore, we aimed to develop a novel MR imaging and analysis method for visualizing thoracic movements during free breathing, enabling assessment of positional effects.Methods: Five healthy male participants were enrolled. MR images were obtained in the supine and semi-prone positions using a fast imaging sequence, allowing for continuous dynamic imaging during deep breathing. Subsequently, an image processing pipeline was applied to enhance visibility. The thoracic expansion was measured and compared between the two positions. Intra-and interobserver reproducibility and test-retest reproducibility were assessed using intraclass correlation coefficients (ICCs).Results: The proposed method enabled successful dynamic visualization of thoracic movements without using a receiver coil. A significant difference in thoracic expansion between the supine and semi-prone positions was observed in the head-foot and right-left directions. Head-foot expansion was greater in the supine position, while right-left expansion was greater in the semi-prone position. No significant differences were found in the anterior-posterior direction. Both intra-and interobserver reproducibility were high, with ICCs exceeding 0.9 for most thoracic expansion measurements. Test-retest reproducibility also demonstrated high agreement for most measurements, with ICCs ranging from 0.74 to 0.97 across different directions and positions.The developed MR imaging method allows for noninvasive visualization of thoracic movements during natural breathing with robust reproducibility. This method could provide valuable insights into respiratory mechanics, supporting its clinical application in respiratory rehabilitation.