Patient-specific CFD modeling of cerebrospinal fluid dynamics in lumbar spinal stenosis using weight-bearing MRI: influence of stenosis severity and postural changes.

Abstract

Lumbar spinal stenosis (LSS) is a degenerative condition and a leading cause of pain and disability in adults. Diagnosis relies on clinical and MRI assessment, yet conventional morphological grading often correlates poorly with symptoms.Despite cauda equina nerve root dispersion may better explain symptom variability,cerebrospinal fluid (CSF)flow alterations in LSS remain poorly characterized.To fill this gap,this work introduces a novel computational workflow to quantify how stenosis severity affects CSF pressure, velocity, and shear-related metrics. We retrospectively developed patient-specific computational fluid dynamics (CFD) models of lumbar CSF flow using weight-bearing MRI. Patients with T2-weighted lumbosacral MRI in both supine and upright positions were included and stratified as moderate or severe LSS. Simulations were performed under physiological boundary conditions to evaluate pressure, velocity, and shear metrics. A total of twenty patients yieldedfortysimulationsobtainedfrom the uprightandsupine MRI. Severe stenosis showed higher mean pressure (p = 0.0323), greater pressure range (p < 0.001), andincreased peak velocities (p < 0.001). Maximum pressure negatively correlated withstenosisarea (r =  - 0.548, p = 0.00124).ElevatedTime-Averaged WallShearStress (TAWSS)and less uniformOscillatory Shear Index (OSI)distributionwereobservedwithin the stenotic regions, especially in severe cases, with differences also between supine and upright positions. CSF flow alterations are modulated by canal morphology and posture, contributing to mechanical stress and nerve root redistribution. CFD modelling provides a non-invasive, quantitative approach to characterizing CSF biomechanics in LSS and supports the development of patient-specific diagnostic strategies.