Characterization of the discrepancies between four-dimensional phase-contrast magnetic resonance imaging and in-silico simulations of cerebrospinal fluid dynamics

Four-dimensional flow MRI for quantitative assessment of cerebrospinal fluid dynamics: Status and opportunities

Neurological disorders can manifest with altered neurofluid dynamics in different compartments of the central nervous system. These include alterations in cerebral blood flow, cerebrospinal fluid (CSF) flow, and tissue biomechanics. Noninvasive quantitative assessment of neurofluid flow and tissue motion is feasible with phase contrast magnetic resonance imaging (PC MRI). While two-dimensional (2D) PC MRI is routinely utilized in research and clinical settings to assess flow dynamics through a single imaging slice, comprehensive neurofluid dynamic assessment can be limited or impractical. Recently, four-dimensional (4D) flow MRI (or time-resolved three-dimensional PC with three-directional velocity encoding) has emerged as a powerful extension of 2D PC, allowing for large volumetric coverage of fluid velocities at high spatiotemporal resolution within clinically reasonable scan times. Yet, most 4D flow studies have focused on blood flow imaging. Characterizing CSF flow dynamics with 4D flow (i.e., 4D CSF flow) is of high interest to understand normal brain and spine physiology, but also to study neurological disorders such as dysfunctional brain metabolite waste clearance, where CSF dynamics appear to play an important role. However, 4D CSF flow imaging is challenged by the long T1 time of CSF and slower velocities compared with blood flow, which can result in longer scan times from low flip angles and extended motion-sensitive gradients, hindering clinical adoption. In this work, we review the state of 4D CSF flow MRI including challenges, novel solutions from current research and ongoing needs, examples of clinical and research applications, and discuss an outlook on the future of 4D CSF flow.

Intrathecal Drug Delivery: Advances and Applications in the Management of Chronic Pain Patient

Advances in our understanding of the biology of spinal systems in organizing and defining the content of exteroceptive information upon which higher centers define the state of the organism and its role in the regulation of somatic and automatic output, defining the motor response of the organism, along with the unique biology and spatial organization of this space, have resulted in an increased focus on therapeutics targeted at this extracranial neuraxial space. Intrathecal (IT) drug delivery systems (IDDS) are well-established as an effective therapeutic approach to patients with chronic non-malignant or malignant pain and as a tool for management of patients with severe spasticity and to deliver therapeutics that address a myriad of spinal pathologies. The risk to benefit ratio of IDD makes it a useful interventional approach. While not without risks, this approach has a significant therapeutic safety margin when employed using drugs with a validated safety profile and by skilled practioners. The present review addresses current advances in our understanding of the biology and dynamics of the intrathecal space, therapeutic platforms, novel therapeutics, delivery technology, issues of safety and rational implementation of its therapy, with a particular emphasis upon the management of pain.

Deep Learning for Computational Hemodynamics: A Brief Review of Recent Advances

Computational fluid dynamics (CFD) modeling of blood flow plays an important role in better understanding various medical conditions, designing more effective drug delivery systems, and developing novel diagnostic methods and treatments. However, despite significant advances in computational technology and resources, the expensive computational cost of these simulations still hinders their transformation from a research interest to a clinical tool. This bottleneck is even more severe for image-based, patient-specific CFD simulations with realistic boundary conditions and complex computational domains, which make such simulations excessively expensive. To address this issue, deep learning approaches have been recently explored to accelerate computational hemodynamics simulations. In this study, we review recent efforts to integrate deep learning with CFD and discuss the applications of this approach in solving hemodynamics problems, such as blood flow behavior in aorta and cerebral arteries. We also discuss potential future directions in the field. In this review, we suggest that incorporating physiologic understandings and underlying fluid mechanics laws in deep learning models will soon lead to a paradigm shift in the development novel non-invasive computational medical decisions.

Association Between Resistance to Cerebrospinal Fluid Flow Near the Foramen Magnum and Cough-Associated Headache in Adult Chiari Malformation Type I

Cough-associated headaches (CAHs) are thought to be distinctive for Chiari malformation type I (CMI) patients and have been shown to be related to the motion of cerebrospinal fluid (CSF) near the foramen magnum (FM). We used computational fluid dynamics (CFD) to compute patient-specific resistance to CSF motion in the spinal canal for CMI patients to determine its accuracy in predicting CAH. Fifty-one symptomatic CMI patients with cerebellar tonsillar position (CTP) ≥ 5 mm were included in this study. The patients were divided into two groups based on their symptoms (CAH and non-CAH) by review of the neurosurgical records. CFD was utilized to simulate CSF motion, and the integrated longitudinal impedance (ILI) was calculated for all patients. A receiver operating characteristic (ROC) curve was evaluated for its accuracy in predicting CAH. The ILI for CMI patients with CAH (776 dyn/cm5, 288-1444 dyn/cm5; median, interquartile range) was significantly larger compared to non-CAH (285 dyn/cm5, 187-450 dyn/cm5; p = 0.001). The ILI was more accurate in predicting CAH in CMI patients than the CTP when the comparison was made using the area under the ROC curve (AUC) (0.77 and 0.70, for ILI and CTP, respectively). ILI ≥ 750 dyn/cm5 had a sensitivity of 50% and a specificity of 95% in predicting CAH. ILI is a parameter that is used to assess CSF blockage in the spinal canal and can predict patients with and without CAH with greater accuracy than CTP.

In vitro evaluation of cerebrospinal fluid velocity measurement in type I Chiari malformation: repeatability, reproducibility, and agreement using 2D phase contrast and 4D flow MRI

Background

Phase contrast magnetic resonance imaging, PC MRI, is a valuable tool allowing for non-invasive quantification of CSF dynamics, but has lacked adoption in clinical practice for Chiari malformation diagnostics. To improve these diagnostic practices, a better understanding of PC MRI based measurement agreement, repeatability, and reproducibility of CSF dynamics is needed.

Methods

An anatomically realistic in vitro subject specific model of a Chiari malformation patient was scanned three times at five different scanning centers using 2D PC MRI and 4D Flow techniques to quantify intra-scanner repeatability, inter-scanner reproducibility, and agreement between imaging modalities. Peak systolic CSF velocities were measured at nine axial planes using 2D PC MRI, which were then compared to 4D Flow peak systolic velocity measurements extracted at those exact axial positions along the model.

Results

Comparison of measurement results showed good overall agreement of CSF velocity detection between 2D PC MRI and 4D Flow (p = 0.86), fair intra-scanner repeatability (confidence intervals ± 1.5 cm/s), and poor inter-scanner reproducibility. On average, 4D Flow measurements had a larger variability than 2D PC MRI measurements (standard deviations 1.83 and 1.04 cm/s, respectively).

Conclusion

Agreement, repeatability, and reproducibility of 2D PC MRI and 4D Flow detection of peak CSF velocities was quantified using a patient-specific in vitro model of Chiari malformation. In combination, the greatest factor leading to measurement inconsistency was determined to be a lack of reproducibility between different MRI centers. Overall, these findings may help lead to better understanding for application of 2D PC MRI and 4D Flow techniques as diagnostic tools for CSF dynamics quantification in Chiari malformation and related diseases.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12987-021-00246-3.

Compressed-sensing accelerated 4D flow MRI of cerebrospinal fluid dynamics

BACKGROUND

4D flow magnetic resonance imaging (MRI) of CSF can make an important contribution to the understanding of hydrodynamic changes in various neurological diseases but remains limited in clinical application due to long acquisition times. The aim of this study was to evaluate the accuracy of compressed SENSE accelerated MRI measurements of the spinal CSF flow.

METHODS

In 20 healthy subjects 4D flow MRI of the CSF in the cervical spine was acquired using compressed sensitivity encoding [CSE, a combination of compressed sensing and parallel imaging (SENSE) provided by the manufacturer] with acceleration factors between 4 and 10. A conventional scan using SENSE was used as reference. Extracted parameters were peak velocity, absolute net flow, forward flow and backward flow. Bland-Altman analysis was performed to determine the scan-rescan reproducibility and the agreement between SENSE and compressed SENSE. Additionally, a time accumulated flow error was calculated. In one additional subject flow of the spinal canal at the level of the entire spinal cord was assessed.

RESULTS

Averaged acquisition times were 10:21 min (SENSE), 9:31 min (CSE4), 6:25 min (CSE6), 4:53 min (CSE8) and 3:51 min (CSE10). Acquisition of the CSF flow surrounding the entire spinal cord took 14:40 min. Bland-Altman analysis showed good agreement for peak velocity, but slight overestimations for absolute net flow, forward flow and backward flow (< 1 ml/min) in CSE4-8. Results of the accumulated flow error were similar for CSE4 to CSE8.

CONCLUSION

A quantitative analysis of acceleration factors CSE4-10 showed that CSE with an acceleration factor up to 6 is feasible. This allows a scan time reduction of 40% and enables the acquisition and analysis of the CSF flow dynamics surrounding the entire spinal cord within a clinically acceptable scan time.

Anthropomorphic Model of Intrathecal Cerebrospinal Fluid Dynamics Within the Spinal Subarachnoid Space: Spinal Cord Nerve Roots Increase Steady-Streaming

Cerebrospinal fluid (CSF) dynamics are thought to play a vital role in central nervous system (CNS) physiology. The objective of this study was to investigate the impact of spinal cord (SC) nerve roots (NR) on CSF dynamics. A subject-specific computational fluid dynamics (CFD) model of the complete spinal subarachnoid space (SSS) with and without anatomically realistic NR and nonuniform moving dura wall deformation was constructed. This CFD model allowed detailed investigation of the impact of NR on CSF velocities that is not possible in vivo using magnetic resonance imaging (MRI) or other noninvasive imaging methods. Results showed that NR altered CSF dynamics in terms of velocity field, steady-streaming, and vortical structures. Vortices occurred in the cervical spine around NR during CSF flow reversal. The magnitude of steady-streaming CSF flow increased with NR, in particular within the cervical spine. This increase was located axially upstream and downstream of NR due to the interface of adjacent vortices that formed around NR.

Reconciling PC-MRI and CFD: An in-vitro study

Several well-resolved 4D Flow MRI acquisitions of an idealized rigid flow phantom featuring an aneurysm, a curved channel as well as a bifurcation were performed under pulsatile regime. The resulting hemodynamics were processed to remove MRI artifacts. Subsequently, they were compared with CFD predictions computed on the same flow domain, using an in-house high-order low dissipative flow solver. Results show that reaching a good agreement is not straightforward but requires proper treatments of both techniques. Several sources of discrepancies are highlighted and their impact on the final correlation evaluated. While a very poor correlation (r²  = 0.63) is found in the entire domain between raw MRI and CFD data, correlation as high as r²  = 0.97 is found when artifacts are removed by post-processing the MR data and down sampling the CFD results to match the MRI spatial and temporal resolutions. This work demonstrates that, in a well-controlled environment, both PC-MRI and CFD might bring reliable and correlated flow quantities when a proper methodology to reduce the errors is followed.

Development of the MRI Flow Phantom System Focused on Low Speed Flows in Fluid Machinery

At present, magnetic resonance imaging (MRI), which is one of the noninvasive diagnostic imaging techniques for medical use, has been applied to industrial product inspection. In this study, a hydraulic model test system, called the “flow phantom system,” which can generate steady flow and oscillating flow for time-resolved MRI, was developed to aid optimum design of fluid machinery. Simple flow phantoms are tested under laminar flow or oscillating flow to evaluate the validity and effectiveness of the proposed system. In this article, the test results of double cylindrical pipe flow, elbow pipe flow, cylindrical valve flow, and jet pump flow, which are often seen in fluid machines, are tested using the 2D time-spatial labeling inversion pulse (Time-SLIP) method, which can track a labeled water mass and visualize it using two-dimensional images. MRI-detected flow patterns were compared with particle image velocimetry (PIV) or numerical simulation.

Nonuniform Moving Boundary Method for Computational Fluid Dynamics Simulation of Intrathecal Cerebrospinal Flow Distribution in a Cynomolgus Monkey

A detailed quantification and understanding of cerebrospinal fluid (CSF) dynamics may improve detection and treatment of central nervous system (CNS) diseases and help optimize CSF system-based delivery of CNS therapeutics. This study presents a computational fluid dynamics (CFD) model that utilizes a nonuniform moving boundary approach to accurately reproduce the nonuniform distribution of CSF flow along the spinal subarachnoid space (SAS) of a single cynomolgus monkey. A magnetic resonance imaging (MRI) protocol was developed and applied to quantify subject-specific CSF space geometry and flow and define the CFD domain and boundary conditions. An algorithm was implemented to reproduce the axial distribution of unsteady CSF flow by nonuniform deformation of the dura surface. Results showed that maximum difference between the MRI measurements and CFD simulation of CSF flow rates was <3.6%. CSF flow along the entire spine was laminar with a peak Reynolds number of ∼150 and average Womersley number of ∼5.4. Maximum CSF flow rate was present at the C4-C5 vertebral level. Deformation of the dura ranged up to a maximum of 134 μm. Geometric analysis indicated that total spinal CSF space volume was ∼8.7 ml. Average hydraulic diameter, wetted perimeter, and SAS area were 2.9 mm, 37.3 mm and 27.24 mm2, respectively. CSF pulse wave velocity (PWV) along the spine was quantified to be 1.2 m/s.

An MRI-Compatible Hydrodynamic Simulator of Cerebrospinal Fluid Motion in the Cervical Spine

GOAL

Develop and test an MRI-compatible hydrodynamic simulator of cerebrospinal fluid (CSF) motion in the cervical spinal subarachnoid space. Four anatomically realistic subject-specific models were created based on a 22-year-old healthy volunteer and a five-year-old patient diagnosed with Chiari I malformation.

METHODS

The in vitro models were based on manual segmentation of high-resolution T2-weighted MRI of the cervical spine. Anatomically realistic dorsal and ventral spinal cord nerve rootlets (NR) were added. Models were three dimensional (3-D) printed by stereolithography with 50-μm layer thickness. A computer controlled pump system was used to replicate the shape of the subject specific in vivo CSF flow measured by phase-contrast MRI. Each model was then scanned by T2-weighted and 4-D phase contrast MRI (4D flow).

RESULTS

Cross-sectional area, wetted perimeter, and hydraulic diameter were quantified for each model. The oscillatory CSF velocity field (flow jets near NR, velocity profile shape, and magnitude) had similar characteristics to previously reported studies in the literature measured by in vivo MRI.

CONCLUSION

This study describes the first MRI-compatible hydrodynamic simulator of CSF motion in the cervical spine with anatomically realistic NR. NR were found to impact CSF velocity profiles to a great degree.

SIGNIFICANCE

CSF hydrodynamics are thought to be altered in craniospinal disorders such as Chiari I malformation. MRI scanning techniques and protocols can be used to quantify CSF flow alterations in disease states. The provided in vitro models can be used to test the reliability of these protocols across MRI scanner manufacturers and machines.

Characteristics of CSF Velocity-Time Profile in Posttraumatic Syringomyelia

BACKGROUND AND PURPOSE

The development of syringomyelia has been associated with changes in CSF flow dynamics in the spinal subarachnoid space. However, differences in CSF flow velocity between patients with posttraumatic syringomyelia and healthy participants remains unclear. The aim of this work was to define differences in CSF flow above and below a syrinx in participants with posttraumatic syringomyelia and compare the CSF flow with that in healthy controls.

MATERIALS AND METHODS

Six participants with posttraumatic syringomyelia were recruited for this study. Phase-contrast MR imaging was used to measure CSF flow velocity at the base of the skull and above and below the syrinx. Velocity magnitudes and temporal features of the CSF velocity profile were compared with those in healthy controls.

RESULTS

CSF flow velocity in the spinal subarachnoid space of participants with syringomyelia was similar at different locations despite differences in syrinx size and locations. Peak cranial and caudal velocities above and below the syrinx were not significantly different (peak cranial velocity, P = .9; peak caudal velocity, P = 1.0), but the peak velocities were significantly lower (P < .001, P = .007) in the participants with syringomyelia compared with matched controls. Most notably, the duration of caudal flow was significantly shorter (P = .003) in the participants with syringomyelia.

CONCLUSIONS

CSF flow within the posttraumatic syringomyelia group was relatively uniform along the spinal canal, but there are differences in the timing of CSF flow compared with that in matched healthy controls. This finding supports the hypothesis that syrinx development may be associated with temporal changes in spinal CSF flow.

Barrier dysfunction or drainage reduction: differentiating causes of CSF protein increase

BACKGROUND

Cerebrospinal fluid (CSF) protein analysis is an important element in the diagnostic chain for various central nervous system (CNS) pathologies. Among multiple existing approaches to interpreting measured protein levels, the Reiber diagram is particularly robust with respect to physiologic inter-individual variability, as it uses multiple subject-specific anchoring values. Beyond reliable identification of abnormal protein levels, the Reiber diagram has the potential to elucidate their pathophysiologic origin. In particular, both reduction of CSF drainage from the cranio-spinal space as well as blood-CNS barrier dysfunction have been suggested ρas possible causes of increased concentration of blood-derived proteins. However, there is disagreement on which of the two is the true cause.

METHODS

We designed two computational models to investigate the mechanisms governing protein distribution in the spinal CSF. With a one-dimensional model, we evaluated the distribution of albumin and immunoglobulin G (IgG), accounting for protein transport rates across blood-CNS barriers, CSF dynamics (including both dispersion induced by CSF pulsations and advection by mean CSF flow) and CSF drainage. Dispersion coefficients were determined a priori by computing the axisymmetric three-dimensional CSF dynamics and solute transport in a representative segment of the spinal canal.

RESULTS

Our models reproduce the empirically determined hyperbolic relation between albumin and IgG quotients. They indicate that variation in CSF drainage would yield a linear rather than the expected hyperbolic profile. In contrast, modelled barrier dysfunction reproduces the experimentally observed relation.

CONCLUSIONS

High levels of albumin identified in the Reiber diagram are more likely to originate from a barrier dysfunction than from a reduction in CSF drainage. Our in silico experiments further support the hypothesis of decreasing spinal CSF drainage in rostro-caudal direction and emphasize the physiological importance of pulsation-driven dispersion for the transport of large molecules in the CSF.

Accuracy of 4D Flow Measurement of Cerebrospinal Fluid Dynamics in the Cervical Spine: An In Vitro Verification Against Numerical Simulation

Abnormal alterations in cerebrospinal fluid (CSF) flow are thought to play an important role in pathophysiology of various craniospinal disorders such as hydrocephalus and Chiari malformation. Three directional phase contrast MRI (4D Flow) has been proposed as one method for quantification of the CSF dynamics in healthy and disease states, but prior to further implementation of this technique, its accuracy in measuring CSF velocity magnitude and distribution must be evaluated. In this study, an MR-compatible experimental platform was developed based on an anatomically detailed 3D printed model of the cervical subarachnoid space and subject specific flow boundary conditions. Accuracy of 4D Flow measurements was assessed by comparison of CSF velocities obtained within the in vitro model with the numerically predicted velocities calculated from a spatially averaged computational fluid dynamics (CFD) model based on the same geometry and flow boundary conditions. Good agreement was observed between CFD and 4D Flow in terms of spatial distribution and peak magnitude of through-plane velocities with an average difference of 7.5 and 10.6% for peak systolic and diastolic velocities, respectively. Regression analysis showed lower accuracy of 4D Flow measurement at the timeframes corresponding to low CSF flow rate and poor correlation between CFD and 4D Flow in-plane velocities.

Inter-operator Reliability of Magnetic Resonance Image-Based Computational Fluid Dynamics Prediction of Cerebrospinal Fluid Motion in the Cervical Spine

For the first time, inter-operator dependence of MRI based computational fluid dynamics (CFD) modeling of cerebrospinal fluid (CSF) in the cervical spinal subarachnoid space (SSS) is evaluated. In vivo MRI flow measurements and anatomy MRI images were obtained at the cervico-medullary junction of a healthy subject and a Chiari I malformation patient. 3D anatomies of the SSS were reconstructed by manual segmentation by four independent operators for both cases. CFD results were compared at nine axial locations along the SSS in terms of hydrodynamic and geometric parameters. Intraclass correlation (ICC) assessed the inter-operator agreement for each parameter over the axial locations and coefficient of variance (CV) compared the percentage of variance for each parameter between the operators. Greater operator dependence was found for the patient (0.19 < ICC < 0.99) near the craniovertebral junction compared to the healthy subject (ICC > 0.78). For the healthy subject, hydraulic diameter and Womersley number had the least variance (CV = ~2%). For the patient, peak diastolic velocity and Reynolds number had the smallest variance (CV = ~3%). These results show a high degree of inter-operator reliability for MRI-based CFD simulations of CSF flow in the cervical spine for healthy subjects and a lower degree of reliability for patients with Type I Chiari malformation.