Kevin Tangen

CNS wide simulation of flow resistance and drug transport due to spinal microanatomy.

Spinal microstructures are known to substantially affect cerebrospinal fluid patterns, yet their actual impact on flow resistance has not been quantified. Because the length scale of microanatomical aspects is below medical image resolution, their effect on flow is difficult to observe experimentally. Using a computational fluid mechanics approach, we were able to quantify the contribution of micro-anatomical aspects on cerebrospinal fluid (CSF) flow patterns and flow resistance within the entire central nervous system (CNS). Cranial and spinal CSF filled compartments were reconstructed from human imaging data; microscopic trabeculae below the image detection threshold were added artificially. Nerve roots and trabeculae were found to induce regions of microcirculation, whose location, size and vorticity along the spine were characterized. Our CFD simulations based on volumetric flow rates acquired with Cine Phase Contrast MRI in a normal human subject suggest a 2-2.5 fold increase in pressure drop mainly due to arachnoid trabeculae. The timing and phase lag of the CSF pressure and velocity waves along the spinal canal were also computed, and a complete spatio-temporal map encoding CSF volumetric flow rates and pressure was created. Micro-anatomy induced fluid patterns were found responsible for the rapid caudo-cranial spread of an intrathecally administered drug. The speed of rostral drug dispersion is drastically accelerated through pulsatile flow around microanatomy induced vortices. Exploring massive parallelization on a supercomputer, the feasibility of computational drug transport studies was demonstrated. CNS-wide simulations of intrathecal drugs administration can become a practical tool for in silico design, interspecies scaling and optimization of experimental drug trials.

Clearance of Subarachnoid Hemorrhage from the Cerebrospinal Fluid in Computational and In Vitro Models

Subarachnoid hemorrhage (SAH) mostly occurs following the rupture of cerebral aneurysm causing blood to leak into the cranial subarachnoid space (SAS). Hemorrhage volume has been linked to the development of secondary vasospasm. Therefore, eliminating blood contaminants from the cerebrospinal fluid (CSF) space after the initial hemorrhage could improve patient outcomes and prevent the development of vasospasm. A number of clinical trials demonstrate that lumbar drainage effectively clears hemorrhagic debris from the cranial compartment. The benefits of optimal lumbar drainage rate and patient orientation are difficult to determine by trial-and-error in live patients, because of the invasive nature, limited subject availability and ethical considerations. Therefore, there is a lack of consensus about clinical guidelines for the use of continuous lumbar drainage following the ictus of SAH. A realistic bench-top model which reproduces the anatomy and CSF dynamics of the human central nervous system (CNS) was built to experimentally study contaminant clearance scenarios under lumbar drainage. To mimic a hemorrhagic event, porcine blood was injected at the basal cistern level of the bench-top model and the efficacy of lumbar drains was assessed experimentally for different drainage rates and patient orientations. In addition, the efficacy of blood clearance was predicted with a computational fluid dynamics (CFD) model. Bench-top experiments and CFD simulations identify body position and drainage rates as key parameters for effective blood clearance. The study findings suggest the importance of treatment in upright position to maximize contaminant diversion from the cranial CSF compartment. The bench-top CNS model together with the validated CFD predictions of lumbar drainage systems can serve to optimize subject-specific treatment options for SAH patients.

Computational and in vitro experimental investigation of intrathecal drug distribution: Parametric study of the effect of injection volume, cerebrospinal fluid pulsatility, and drug uptake

BACKGROUND: Intrathecal drug delivery is an attractive option to circumvent the blood-brain barrier for pain management through its increased efficacy of pain relief, reduction in adverse side effects, and cost-effectiveness. Unfortunately, there are limited guidelines for physicians to choose infusion or drug pump settings to administer therapeutic doses to specific regions of the spine or the brain. Although empiric trialing of intrathecal drugs is critical to determine the sustained side effects, currently there is no inexpensive in vitro method to guide the selection of spinal drug delivery parameters. The goal of this study is to demonstrate current computational capabilities to predict drug biodistribution while varying 3 parameters: (1) infusion settings, (2) drug chemistry, and (3) subject-specific anatomy and cerebrospinal fluid dynamics. We will discuss strategies to systematically optimize these 3 parameters to administer drug molecules to targeted tissue locations in the central nervous system. METHODS: We acquired anatomical data from magnetic resonance imaging (MRI) and velocity measurements in the spinal cerebrospinal fluid with CINE-MRI for 2 subjects. A bench-top surrogate of the subject-specific central nervous system was constructed to match measured anatomical dimensions and volumes. We generated a computational mesh for the bench-top model. Idealized simulations of tracer distribution were compared with bench-top measurements for validation. Using reconstructions from MRI data, we also introduced a subject-specific computer model for predicting drug spread for the human volunteer. RESULTS: MRI velocity measurements at 3 spinal regions of interest reasonably matched the simulated flow fields in a subject-specific computer mesh. Comparison between the idealized spine computations and bench-top tracer distribution experiments demonstrate agreement of our drug transport predictions to this physical model. Simulated multibolus drug infusion theoretically localizes drug to the cervical and thoracic region. Continuous drug pump and single bolus injection were successful to target the lumbar spine in the simulations. The parenchyma might be targeted suitably by multiple boluses followed by a flush infusion. We present potential guidelines that take into account drug specific kinetics for tissue uptake, which influence the speed of drug dispersion in the model and potentially influence tissue targeting. CONCLUSIONS: We present potential guidelines considering drug-specific kinetics of tissue uptake, which determine the speed of drug dispersion and influence tissue targeting. However, there are limitations to this analysis in that the parameters were obtained from an idealized healthy patient in a supine position. The proposed methodology could assist physicians to select clinical infusion parameters for their patients and provide guidance to optimize treatment algorithms. In silico optimization of intrathecal drug delivery therapies presents the first steps toward a possible care paradigm in the future that is specific to personalized patient anatomy and diseases.

Large-scale subject-specific cerebral arterial tree modeling using automated parametric mesh generation for blood flow simulation

In this paper, we present a novel technique for automatic parametric mesh generation of subject-specific cerebral arterial trees. This technique generates high-quality and anatomically accurate computational meshes for fast blood flow simulations extending the scope of 3D vascular modeling to a large portion of cerebral arterial trees. For this purpose, a parametric meshing procedure was developed to automatically decompose the vascular skeleton, extract geometric features and generate hexahedral meshes using a body-fitted coordinate system that optimally follows the vascular network topology. To validate the anatomical accuracy of the reconstructed vasculature, we performed statistical analysis to quantify the alignment between parametric meshes and raw vascular images using receiver operating characteristic curve. Geometric accuracy evaluation showed an agreement with area under the curves value of 0.87 between the constructed mesh and raw MRA data sets. Parametric meshing yielded on-average, 36.6% and 21.7% orthogonal and equiangular skew quality improvement over the unstructured tetrahedral meshes. The parametric meshing and processing pipeline constitutes an automated technique to reconstruct and simulate blood flow throughout a large portion of the cerebral arterial tree down to the level of pial vessels. This study is the first step towards fast large-scale subject-specific hemodynamic analysis for clinical applications.

Backflow-free catheters for efficient and safe convection-enhanced delivery of therapeutics

Convection-enhanced delivery (CED) is an invasive drug delivery technique used to target specific regions of the brain for the treatment of cancer and neurodegenerative diseases while bypassing the blood-brain barrier. In order to prevent the possibility of backflow, low volumetric flow rates are applied which limit the achievable drug distribution volumes from CED. This can render CED treatment ineffective since a small convective flow produces narrow drug distribution inside the treatment region. Novel catheter designs and CED protocols are needed to improve the drug distribution inside the treatment region. This is especially important when administering toxic chemotherapeutics which could adversely affect other organs if backflow occurred and these drugs entered the circulating blood stream. In order to help elucidate the causes of backflow and to design backflow-free catheters, we have studied the impact that microfluid flow has on deformable brain phantom gels experimentally as well as numerically. We found that fluid injections into porous media have considerable effects on local transport properties such as porosity and hydraulic conductivity. These phenomena not only alter the bulk flow velocity distribution of the microfluid flow due to the changing porosity, but significantly modify flow direction and even volumetric flow distribution due to induced local hydraulic conductivity anisotropy. These studies led us to the development of novel backflow-free catheters with safe volumetric flow rates up to 10 µL/min. The catheter designs, numerical simulations and experimental results are described throughout this article.

Impedance Changes Indicate Proximal Ventriculoperitoneal Shunt Obstruction In Vitro

Extracranial cerebrospinal fluid (CSF) shunt obstruction is one of the most important problems in hydrocephalus patient management. Despite ongoing research into better shunt design, robust and reliable detection of shunt malfunction remains elusive. The authors present a novel method of correlating degree of tissue ingrowth into ventricular CSF drainage catheters with internal electrical impedance. The impedance based sensor is able to continuously monitor shunt patency using intraluminal electrodes. Prototype obstruction sensors were fabricated for in-vitro analysis of cellular ingrowth into a shunt under static and dynamic flow conditions. Primary astrocyte cell lines and C6 glioma cells were allowed to proliferate up to 7 days within a shunt catheter and the impedance waveform was observed. During cell ingrowth a significant change in the peak-to-peak voltage signal as well as the root-mean-square voltage level was observed, allowing the impedance sensor to potentially anticipate shunt malfunction long before it affects fluid drainage. Finite element modeling was employed to demonstrate that the electrical signal used to monitor tissue ingrowth is contained inside the catheter lumen and does not endanger tissue surrounding the shunt. These results may herald the development of “next generation” shunt technology that allows prediction of malfunction before it affects patient outcome.

Cerebrospinal Fluid Dynamics and Intrathecal Delivery

Abstract Intrathecal delivery of opiates directly into the cerebrospinal fluid (CSF) is the oldest technique for the delivery of anesthetics to the central nervous system (CNS). Despite the long empirical experience with spinal anesthetics, the relationships between CSF dynamics and biodistribution of intrathecally delivered drugs are complex and the mechanisms that lead to this complex distribution are poorly understood. First-principle models of fluid mechanics have been created to elucidate complex flow patterns in subject-specific computations. It has also been shown that complex CSF flow patterns are responsible for the vigorous mixing effects that govern the biodistribution of intrathecally delivered drugs. This chapter aims to link CSF flow patterns to expected drug dispersion. However, predicting the biodistribution of drugs is not an easy task. Due to the difficulty of access to the CNS, computational analysis capable of interpreting spatial and temporally distributed imaging data plays a vital role in elucidating pharmacokinetics and pharmacodynamics of intrathecal drug delivery. We review progress in our lab as well as the open literature to give a snapshot of this very active field. The chapter aims to understand better the relationship between infusion parameters and achievable drug distribution. In the near future we expect that the integration of imaging data with first-principle computational models such as those described here will enable us to design more effective drug administration strategies so that specific cells and tissue in any point of the CNS can be more effectively targeted than is the case with existing methods. We close by pointing out research directions that will pave the way for these imminent developments.

Systems engineers’ role in biomedical research. Convection-enhanced drug delivery

Abstract Convection-enhanced delivery (CED) is an emerging field focused on administering therapeutic drugs directly to the brain to treat a multitude of pathologies ranging from malignant tumors to degenerative diseases. CED methods include direct administration of therapeutics using a catheter to the target site or using other vectors, such as nanoparticles or retroviruses to localize the therapy. In this chapter, the authors review landmark in vitro experiments, including catheter design and brain phantom models, and address contemporary challenges in catheter design, including the backflow problem. Prominent animal studies are next discussed, including tumor, epilepsy, and degenerative models, and vectors for drug delivery, including nanocarriers, liposomes, antibodies, and viral vectors are explored. Contemporary clinical trials are then outlined, including current brain tumor, neurodegenerative, and metabolic disorder initiatives. The chapter concludes with a discussion of ongoing challenges and future directions in CED which systems engineering can advance.

Modeling the diffusion of D-2-hydroxyglutarate from IDH1 mutant gliomas in the central nervous system

Background Among diffusely infiltrative gliomas in adults, 20%-30% contain a point mutation in isocitrate dehydrogenase 1 (IDH1mut), which increases production of D-2-hydroxyglutarate (D2HG). This is so efficient that D2HG often reaches 30 mM within IDH1mut gliomas. Yet, while up to 100 µM D2HG can be detected in the circulating cerebrospinal fluid of IDH1mut glioma patients, the exposure of nonneoplastic cells within and surrounding an IDH1mut glioma to D2HG is unknown and difficult to measure directly. Methods Conditioned medium from patient-derived wild type IDH1 (IDH1wt) and IDH1mut glioma cells was analyzed for D2HG by liquid chromatography-mass spectrometry (LC-MS). Mathematical models of D2HG release and diffusion around an IDH1mut glioma were independently generated based on fluid dynamics within the brain and on previously reported intratumoral and cerebrospinal D2HG concentrations. Results LC-MS analysis indicates that patient-derived IDH1mut glioma cells release 3.7-97.0 pg D2HG per cell per week. Extrapolating this to an average-sized tumor (30 mL glioma volume and 1 × 108 cells/mL tumor), the rate of D2HG release by an IDH1mut glioma (SA) is estimated at 3.2-83.0 × 10-12 mol/mL/sec. Mathematical models estimate an SA of 2.9-12.9 × 10-12 mol/mL/sec, within the range of the in vitro LC-MS data. In even the most conservative of these models, the extracellular concentration of D2HG exceeds 3 mM within a 2 cm radius from the center of an IDH1mut glioma. Conclusions The microenvironment of an IDH1mut glioma is likely being exposed to high concentrations of D2HG, in the low millimolar range. This has implications for understanding how D2HG affects nonneoplastic cells in an IDH1mut glioma.