Validating the accuracy of real-time phase-contrast MRI and quantifying the effects of free breathing on cerebrospinal fluid dynamics

Understanding of the cerebrospinal fluid (CSF) circulation is essential for physiological studies and clinical diagnosis. Real-time phase contrast sequences (RT-PC) can quantify beat-to-beat CSF flow signals. However, the detailed effects of free-breathing on CSF parameters are not fully understood. This study aims to validate RT-PC's accuracy by comparing it with the conventional phase-contrast sequence (CINE-PC) and quantify the effect of free-breathing on CSF parameters at the intracranial and extracranial levels using a time-domain multiparametric analysis method. Thirty-six healthy participants underwent MRI in a 3T scanner for CSF oscillations quantification at the cervical spine (C2-C3) and Sylvian aqueduct, using CINE-PC and RT-PC. CINE-PC uses 32 velocity maps to represent dynamic CSF flow over an average cardiac cycle, while RT-PC continuously quantifies CSF flow over 45-seconds. Free-breathing signals were recorded from 25 participants. RT-PC signal was segmented into independent cardiac cycle flow curves (Q ) and reconstructed into an averaged Q . To assess RT-PC's accuracy, parameters such as segmented area, flow amplitude, and stroke volume (SV) of the reconstructed Q from RT-PC were compared with those derived from the averaged Q generated by CINE-PC. The breathing signal was used to categorize the Q into expiratory or inspiratory phases, enabling the reconstruction of two Q for inspiration and expiration. The breathing effects on various CSF parameters can be quantified by comparing these two reconstructed Qt. RT-PC overestimated CSF area (82.7% at aqueduct, 11.5% at C2-C3) compared to CINE-PC. Stroke volumes for CINE-PC were 615 mm³ (aqueduct) and 43 mm³ (spinal), and 581 mm³ (aqueduct) and 46 mm³ (spinal) for RT-PC. During thoracic pressure increase, spinal CSF net flow, flow amplitude, SV, and cardiac period increased by 6.3%, 6.8%, 14%, and 6%, respectively. Breathing effects on net flow showed a significant phase difference compared to the other parameters. Aqueduct-CSF flows were more affected by breathing than spinal-CSF. RT-PC accurately quantifies CSF oscillations in real-time and eliminates the need for cardiac synchronization, enabling the quantification of the cardiac and breathing components of CSF flow. This study quantifies the impact of free-breathing on CSF parameters, offering valuable physiological references for understanding the effects of breathing on CSF dynamics.