Noninvasive Measurement of Cerebrospinal Fluid Flow in Shunted Hydrocephalus: Protocol for Scanner Calibration and Multisite Data Collection.

Background: Ventricular shunts divert cerebrospinal fluid (CSF) in patients with hydrocephalus, which can be lifesaving. Untreated shunt failure may lead to increased intracranial pressure and neurological injury. The process of diagnosing shunt malfunction can be complex, and historically, there has not been a simple method for noninvasive and quantitative measurement of CSF flow through shunts. The demonstration of successful clinical application of phase-contrast magnetic resonance imaging (PC-MRI) to noninvasively quantify shunt flow is relatively new and will benefit from standardization across varying types of magnetic resonance (MR) hardware to facilitate implementation at multiple medical sites. PC-MRI CSF flow measurement through ventricular shunts has not yet been compared across different types of MR hardware (ie, differing field strengths, radiofrequency coils, slew rates, gradient strengths, models, and manufacturers). Objective: This study describes a protocol aimed to standardize and optimize PC-MRI shunt flow measurements for widespread use on any MR scanner. Methods: To study shunt flow in a manner that would translate effectively to the patient care setting, a phantom model incorporating a human shunt catheter can be constructed to obtain data over a typical physiological range of CSF flow rates. Calibration curves are used to model data comparing known flow rates to flow measured via PC-MRI. The accuracy of each MR scanner is assessed using linear regression. This protocol will be repeated on 8 MR scanners before multisite data collection. Results: The shunt flow phantom was constructed in November 2024. Single-site phantom calibration began in April 2025 and concluded in December 2025. Registration of clinical data collection sites will take place between November 2025 and July 2026. Approval from the Children's Hospital of Los Angeles Institutional Review Board was obtained in January 2023 (before scanner calibration), and single-site data collection is ongoing. Multi-institutional clinical data collection will begin in July 2026 and continue until January 2027. Results and statistical analyses are expected by April 2027. Conclusions: This study protocol provides a methodology to test and implement PC-MRI on any MR scanner using a phantom model that (1) represents real flow and catheter conditions, (2) provides a repeatable means for data collection, (3) is easy to assemble, (4) minimizes artifacts, and (5) is transportable. We also describe analytic methods for interscanner calibration across a range of hardware parameters and provide a framework for multisite data collection.