Cloud native EDA tools & pre-optimized hardware platforms
Posted on 21 September 2020 by Celia Butler
Personalized implants and devices offer a whole series of benefits to the patient. The biggest benefits are:
Creating these implants and devices all depends on accurate models of the patient’s individual anatomy. This has always been a time-consuming process that makes implants slower to develop and manufacture, resulting in high costs. With the enhancements of 3D printing and additive manufacturing, it is now possible to create a series of individual and incredibly complex and detailed parts with high accuracy. However, there are still delays at the design stage, whereby expert biomedical engineers with training and experience in anatomical segmentation are typically required to work on patient anatomies. This approach must be used for both straightforward cases and more complex medical conditions, as well as for injured anatomies.
Synopsys has tackled this challenge by harnessing Machine Learning (ML) techniques to develop a fully automated tool that takes the time and hassle out of segmenting knees. Knees are one of the most commonly replaced joints, and so presents an ideal example of how segmentation can be optimized to improve effectiveness in many key areas.
The result: the Simpleware AS Ortho module allows MRI scans of knees to be segmented in just a few minutes compared to an hour or more using manual segmentation! Simpleware ML algorithms are trained by experts and checked by clinical professionals, so you can be confident of accurate reliable results.
The model produced includes a series of masks of the bones and cartilage from the knee, all segmented from the image data available. Patient-specific landmarks are also automatically placed on the model.
Using Simpleware AS Ortho not only cuts down the time spent segmenting the image data, it is also a consistent approach, so there are fewer inconsistencies in the process that could be caused by different users’ interpretation of the features in the image data. This means the models do not need to go through multiple reviews between team members.
Automated knee segmentation (Femur, Tibia, and associated cartilage, Patella, Fibula) in Simpleware AS Ortho.
1. Import your data into Simpleware ScanIP using our dedicated DICOM importer or use the Image import if you have stacks of images or raw data. Currently this tool supports PD weighted Sag/Cor, and T2 Cor MRI scan data (with more to be released).
2. Open the Knee MRI tool in the Auto segmentation tab. The Knee MRI tool presents are a series of options:
Use these options to customize your model by removing parts or steps you don’t require to speed up your model generation even further.
The Knee MRI tool in the Auto segmentation tab.
3. Press Apply.
4. Review your model. Once the model has been generated, you can view it as a series of masks in the 2D slice views or in 3D. The landmarks placed on your model can be seen in the Measurements tools, where you can unlock and edit them as required. You can also use all the tools and features in Simpleware ScanIP as usual. This tool has just sped up the process by doing the initial segmentation for you.
Knee landmarks placed on Femur and Tibia in Simpleware AS Ortho.
The model can then be used in a range of ways:
Simpleware AS Ortho can work through batches of cases using automation and scripting capabilities, freeing up expert time for review, design and development. For more information on our customization tools, please read our blog Customize and Automate Workflows with Simpleware Software.
For a fully tailored solution to your individual workflows, speak to us to learn more about Simpleware Custom Modeler.
We hope this blog gave you a brief insight into our the Simpleware AS Ortho Module and the Knee MRI tool in particular. If you would like to know more, please contact us. Our technical specialists will be happy to help with problem-solving and taking on any unique challenges that you are facing.