Streamlining Innovation : Thinking in 3-D

Streamlining Innovation

At Interpore Cross International, a new software package was instrumental in helping engineers create an expandable spinal implant.

By Mark E. Apgar, Project Engineer, Interpore Cross International, Dublin, OH

Interpore Cross International had been designing spinal implants for years when it was presented with a new challenge: designing an expandable spinal implant. A neurosurgeon with a patent for a cage-type device had approached the corporation about developing his design. After careful consideration, Cross decided to accept the challenge and, in order to meet it, employed a variety of software programs to optimize the device's development.

The Challenge

Surgeons may opt to use an expandable spinal implant when performing certain types of spinal fusions. When spinal disks collapse, they can press on nerves, resulting in pain and loss of nerve signal to the limbs. In a spinal fusion procedure, a surgeon moves the vertebral bodies apart to relieve pressure on the nerves, then places bone graft material in the disk spaces between the vertebrae. An expandable vertebral device is used in place of a strut graft to hold the vertebral bodies apart and stabilize the spine while the bone graft placed inside the device takes hold, a process that usually takes several months.

The difference between an expandable vertebral device and a standard spinal cage is that after the expandable device is installed (in a compressed position) it can be lengthened to move vertebral bodies apart. The device is then locked in the expanded position. One of the challenges Cross faced in designing an expandable device was devising a locking mechanism that would hold securely. The device needed to be able to resist the dynamic loads placed on it—that is, be strong enough to support the weight of a person's head when used for treating cervical (neck) tumors.

Using the neurosurgeon's sketches, Cross created a solid model with Solid Modeling from Unigraphics Solutions Inc. (St. Louis). When the model was ready, Cross invited the surgeon to its offices. Under the surgeon's supervision, the engineers went through five different iterations of the design on the computer. Because Cross had parameterized the model, it was easy to make changes in response to the surgeon's suggestions. For example, when the surgeon wanted to change the angle of a flange to better meet the angle of a vertebral body, engineers simply changed a number and regenerated a model with the new flange angle. The surgeon was impressed with the speed at which Cross was able to incorporate his suggestions. Prior to this, he had worked out bugs by building and evaluating prototypes, which often took months to create.

How Did They Do It?

A spinal implant system consists of bone screws, hooks that attach to the vertebra, lateral connectors, and rods, as well as the tools needed to install these components. To be effective, a spinal implant must be strong, small, and easy to install. These were the guiding factors in the design of Cross's original Synergy spinal implant system, which has been on the market since 1995. For that design, Cross used an earlier generation of 2-D computer-aided design (CAD) technology to create drawings.

Engineers at Interpore Cross used Unigraphics CAD/CAM software to redesign the threads of a spinal implant screw. The software allowed design engineers to easily vary the model's parameters and run an analysis on the updated models until they found a thread design that met their requirements.

Although the 2-D CAD allowed Cross to create drawings quickly and to revise them easily, its effectiveness was limited. This is because Cross also used a finite-element analysis (FEA) program, Nastran, from MacNeal Schwendler Corp. (MSC), to simulate the performance of implant components. Nastran requires the use of a 3-D computer model to replicate real-world conditions. Working from the 2-D drawings, the Cross analyst had to recreate designs in 3-D in Nastran, a process that could take weeks and that limited the use of FEA. Despite its limitations, the system helped Cross fine-tune several components of the first implant system. It demonstrated the effect on stiffness and strength of changing bone screw variables such as pitch, ratio between major and minor diameter, and minor diameter taper. Using FEA as a guide, Cross created a new screw design in which the first five threads below the screw head are tapered for optimal performance.

Although these insights were valuable and helped Cross create a better product, the design of the original implant required a great deal of prototype testing. Here, too, someone had to work from the 2-D drawings—in this case, to program a CNC machine for cutting the metal prototype parts. Each prototype took 2 to 8 weeks to make.

Because Cross wished to reduce reliance on metal prototypes and allow greater use of FEA, the engineers decided to move from 2-D CAD to solid modeling. The benefits would be twofold: solid models could be transferred to Nastran, sparing the analyst the task of recreating designs for FEA, and it would allow Cross to replace metal prototypes with rapid prototyping techniques such as stereolithography. Cross also wanted parametric modeling capabilities, which would allow them to vary designs by plugging in new values for design variables. Because Cross creates a large number of screws, finding a program to simplify this task was a high priority. In a program such as 2-D CAD, laying out the threads was tedious and time-consuming. Cross wanted something more efficient and wanted the ability to export files directly into Nastran.

Cross found that Unigraphics' Solid Modeling met these requirements. The program's law curve capability proved to be an efficient way of modeling screws. Creating a path that a helix follows, a designer can simply attach the helix to a sketch of a screw thread, and the program removes the thread pattern from the solid mass of the screw. In addition, because MSC is a Unigraphics Solutions Alliance Program member, the company's products have a tight integration that allows solid models created in Unigraphics to be transferred to Nastran without data translation for the finite-element mesh.

Taking a Test Drive

The "test drive" of the new system was not as beneficial as it could have been because Cross did not have the new software from the very start of the project. The plan was to convert the corporation's original implant system from stainless steel into titanium. The engineers hoped they would be able to convert their original implant design to titanium without much reengineering; the fact that titanium is twice as flexible as stainless steel, however, complicated the matter. Because the engineering team did not have a solid modeler initially, they were required to do extensive prototype testing to investigate how the titanium threads that are inserted into bone would perform under physiological loading. This is when engineers decided to use the Unigraphics software to look for a solution.

The first step was to create the original screw design as a Unigraphics' solid model. Cross created a sketch, constrained parametrically, so that modifications would be easy to make later. Some of the parameterized values included major diameter, minor diameter, and pitch. The law curve function helped speed the process.

Next, Cross exported the solid model to Nastran to simulate the loads the screw would be subjected to inside a patient. Nastran results allowed engineeres to compare the performance of different thread designs. After analyzing one design in Nastran, Cross went back to Unigraphics to change the screw model by varying one or more of the model's parameters. Engineers then ran the analysis with the updated model, repeating the process until they had found a thread design that met their requirements.

New Product, New Process

Using the new software on the titanium conversion helped the engineering team complete the transition sooner than it normally could have. By allowing Cross to use FEA to solve the problems caused by the more flexible material, the software reduced the time-consuming prototyping process. Following this experience, it was clear to the engineers that solid modeling had a lot to offer, and that they had tapped only a fraction of its potential.

Designing the expandable vertebral cage would enable the engineers to use the full range of the software's capabilities. On this project, developed entirely in the new system, Cross also made good use of the ability to turn solid models into rapid prototypes. Cross used a stereolithography (STL) process provided by an outside service. When Cross wanted an evaluation part, the engineers simply converted a Unigraphics model to STL format (the conversion routine is included as part of Unigraphics), which is then read by the stereolithography machine. Cross would send an STL file to the service bureau via modem and in 1 or 2 days would receive a part. This was especially helpful when dealing with the surgeon, who liked to have solid model parts to examine.

As in other projects, Cross used Nastran to evaluate stresses on the various parts of the new device, which enabled design engineers to work out problems while the device was still in software, meaning fewer prototypes were needed. Although the engineers still had to build prototypes for final testing, by the time they built them, they were confident the prototypes would work.

Cross estimates use of the software saved more than $100,000 in prototyping and testing costs and reduced development time on this project by a year. Purchasing the new equipment (software and computers) cost $250,000, but Cross estimates it paid for itself after just two projects, including the expandable spinal implant.

Software Focus is a new feature. If you are interested in having your product appear in an upcoming issue, please contact MPMN's editors by fax, 310/392-4920 or e-mail, [email protected].

Return to the MPMN home page