A comparative analysis of compartmental and computational fluid dynamics modeling of cerebrospinal fluid clearance

The global prevalence of Alzheimer's Disease (AD), a neurodegenerative disease linked to aging, is rising. With limited current treatments, early diagnosis and intervention are crucial to reducing mortality rates. Recent studies indicated that slow clearance of toxic brain waste may cause neurodegeneration, sparking interest in cerebrospinal fluid (CSF) dynamics as it removes brain waste. Due to limited noninvasive methods for humans, most data come from animal models. A promising in-vivo approach uses dynamic PET imaging to track radioactive contrast material, requiring complex image processing based on traditional compartmental modeling. An alternative modeling approach is to use computational fluid dynamics (CFD) to analyze the spatiotemporal transport of the PET contrast agent. Both methods estimate the CSF clearance rate as turnover time. Our study compares these models by solving an inverse problem to match model predictions with clinical data. The CFD model uniquely accounts for variations based on simulated brain compartment volume. In a sample of 10 patients, the compartmental model better fits clinical measurements, with a mean absolute percent error of 6.9% compared to 12% for the CFD model. The higher error produced by CFD might stem from simplified assumptions and a limited sample size. Nevertheless, CFD can provide valuable insight and has the potential to aid in the early diagnosis of AD.