Bioelectric Analyses of an Osseointegrated Intelligent Implant Design System for Amputees

By the year 2050, the projected number of amputees in the United States is expected to reach 3.6 million. Complications with traditional prosthetic limbs often result in asymmetric gait patterns and discomfort for amputees who hope to resume active lifestyles following injury. Current prosthetic options are generally inadequate.

This is primarily due to the high incidence of pressure sores from non-uniform loading patterns. One way to overcome the functional limitations of traditional prosthetics is to engineer direct skeletal attachment between an implant and the host bone of the amputee. This technology known as osteointegration depends on direct skeletal fixation and a tight epidermal and dermal seal to prevent infection when successful, osteointegration drastically improves activity levels for amputees with high proximal amputations.

This is shown in the videos provided by Dr.Ash off's team. In conjunction with the Esca organization. There are two major complications that prevent the widespread adoption of osteointegration.

The first is a risk of bacterial adhesion and biofilm formation on orthopedic implants. Second, there is a risk of inadequate skeletal fixation in vivo as shown here with radiolucencies. To overcome these complications, the exo prosthesis may be used as a cathode to deliver electrical stimulation.

This may result in positive bone remodeling, prevent bacterial attachment, and reduce the rehabilitation time required for full weight bearing. This video demonstrates how implants are designed using a novel OSSEOINTEGRATED intelligent implant design system, or OIID. Hi, I'm Brad Isaacson from the Bone and Joint Research Laboratory at the University of Utah in the Department of Bioengineering.

Today, we will show you how osteo integrated implant design can be tested and evaluated with hypothetical modeling using computational computer programs. Our goal is to provide proof of concept of a new osteointegration rehabilitation tool, which uses electrical stimulation. When implemented properly, the tool will be able to reduce periprosthetic infections and increase skeletal attachment.

With that in mind, let me show you how we can develop an integrated intelligent implant design system. CT scans are used for model reconstruction because they allow different tissue types to be clearly distinguished. The scans used for this were collected from the University of Utah Hospital and the Department of Veteran Affairs after obtaining IRB and HIPAA approval.

These CT scans were first visually inspected and selected for the study based on the absence of metal implants, which can introduce image artifacts. Here we show a case of an amputee who would be unacceptable to computer model based upon the presence of an orthopedic implant. The amputee has an orthopedic implant inside the medullary canal, and when we scan through the CT slices, we see image distortion.

However, we would advocate for a patient such as this. The amputee has clear distinction between tissue types crucial for anatomically accurate models. Excellent To begin.

Model generation CT scan files are downloaded as DICOM images and loaded into the SEG 3D program as a new volume. A median filter is then used to smooth the imported volumes prior to determining geometrically defined tissue structures. Once filtering is complete, the tissue boundaries of the bone, bone marrow, organs, and adipose tissue are generated by thresholding.

The CT images interactively and filtered once more. The musculature is obtained by manually setting seed points inside the threshold of muscle tissue, and by using a confidence connected filter to find all the tissue connected to the seed points. This step eliminates erroneous tissues that may be grouped together with the muscle based on similar absorbency from cts.

The skin, which is impossible to discern reliably from the CT images is generated by dilating the outermost tissue, two millimeters based on average skin thickness. This produces a layer of homogeneous thickness that surrounds the full model. Segmentations are manually inspected, corrected to ensure accuracy and combined in a hierarchy to a single label map.

Required for finite element analysis. Models are exported to ski run using a near raw raster data format or n NRD file developed by Gordon Delman at the Ski Institute. This is now the basic file type from many open source image processing software platforms.

This file is exported and used as the mesh for finite element analysis. So far, we have manually inspected RCTs and combined them in a hierarchy O model. In the SEG 3D program there, we're able to define tissue geometry and space.

Next, we'll load these models into ski, run our finite element package to define the electrode positioning and type. The reason for using finite element analysis is it provides numerical approximation for complex geometries such as our amputees residual limbs. To prepare for finite element analysis, the 16 centimeter implant is designed using custom software written in mat lab, then imported into ski run to serve as the orthopedic implant and cathode for electrical stimulation.

The electric field must be maintained below one to 10 volts per centimeter, and the current density must be lower than two milliamps per square centimeter. This will induce osteoblast migration, prevent localized tissue heating, and ensure patient safety. This is indicated with the red dashed line.

Small electrodes are designed to ensure patient comfort. Since the OIID device should not restrict movement or daily activities during rehabilitation. The ski run software package is used to design the electrode because it supports interactive electrode placement and simulation.

A network is created and modules organized with specific mathematical functions for defining boundary conditions, tissue conductivities and mesh refinements, as well as for generating histograms of electrical metric distributions and recording filled data. The configurations for the electrodes consist of one patch electrode, two patch electrodes, one continuous band, and two continuous bands. External electrode bands, which are 1.6 centimeters thick are applied to the residual limb of the models generated from patient CT scans.

Electrode patches, which are three centimeters thick, are placed in a strip covering approximately half the diameter of the residual limb. The inner cortical implant shown in red represents the osteo implant. This is set to diameter to allow for perfect implant fit and fill.

Now that the electrodes have been placed, we'll perform the finite element analysis. The goal here is to maintain any electric field between one to 10 volts per centimeter to be an effective tool while maintaining a current density below two milliamps per centimeter squared to prevent tissue necrosis To perform finite element analysis, simulations are generated assuming that the electrical metrics can be calculated using a quasi static approach with no time dependency. Numerical approximation is used to compute the electrical potential through the subject's residual limb.

The model is computed by solving poisson's equation and electrostatics for each tissue type generated from the SEG 3D segmentations. Boundary conditions are assigned in the model in which nodes are continuous between tissue types and to govern the path of current flow. Because the electrodes and the implant have much larger conductivities than the surrounding tissues, it is assumed that the implant or cathode is at a constant potential.

Likewise, the surface electrodes are modeled with a constant potential difference from the percutaneous implant. Numerical simulation is used to compute the electric potential through the subject's residual limb to evaluate the efficacy of electrode configuration and sizing. Patient specific models are developed and the electrical potential around the implant interface is used to determine localized field strengths.

The model is generated using a hexa hugel mesh that consists of approximately 1.8 million elements. The optimal element size for this experiment is selected based on a mesh sensitivity study that results in the less than 5%relative difference in voltage gradients. This ensures model accuracy.

The elements are treated as peace wise, homogeneous omic and isotropic elements. Electrodes are incorporated in the finite element meshing and assigned a constant potential difference between the skin electrode and the osseointegrated implant. This selection is based on expected tissue resistivity.

The potential is set between one and 4.5 volts based on the amputee geometry, resistivity of the tissues and electrode configuration using an iterative solver. The potentials in the finite element models are computed for four electrode configurations. This is done to determine the optimal design for amputees.

Validation of finite element analysis is being performed with an iacuc approved small animal in vivo model. An implant is inserted into the medullary canal of a rabbit and a second electrode is placed on the musculature. This generates a subcutaneous electric field at the bone implant interface.

All ski run finite element analysis as validated by needle electrodes placed between the two electrodes, which record the voltage gradients of the rabbit in vivo. The rabbit model study utilizes the same protocol as the OIID designated for human research. This is done by attaining computed tomography scans, developing 3D hierarchical models in SEG 3D and performing computational modeling with ski run.

The goal is to develop the electrical metrics required to prove the efficacy of controlled electrical stimulation. This aims to reduce periprosthetic infections and increase implant fixation strengths. So the technology can be FDA approved and used for rehabilitating amputees.

Successful completion of the detailed protocol will result in anatomically accurate amputee residual limbs developed from computed tomography scans. This will provide the initial guidelines for the OIID system. Users will develop an interactive interface to adjust electrode size and position for patient specific models.

Altered electric fields and current densities will be the result of tissue geometry and electrode position. Therefore, models will require individual alterations to help persons with limb loss seeking osteointegration technology. The osseointegrated intelligent implant design system that we showed you today will help alleviate the classic problem with electrical stimulation, the inability to define the current pathway in the human body.

Preliminary results indicate that the electric field at the bone implant interface may have the capabilities of reducing periprosthetic infections, increasing skeletal attachment, and be used as an effective tool for our clinicians in rehabilitation. So that's it. Thank you for watching and good luck with your osseointegration research.