Case Study: Deploying OCT for Real-Time Medical Imaging Applications, Including Advanced Laser Surgery
January 25, 2022
Many medical imaging methods have been developed to provide visual representations of various body parts.
Optical Coherence Tomography (OCT) is a relatively new technology that’s used to obtain high-resolution cross-sectional images of the retina. As a noninvasive imaging technology, it uses infrared wavelength laser light rays to measure a patient’s retinal thickness. The key upside to the technology, which rose to prominence beginning in 2017, is that no radiation or X-rays are used in this process and, in most cases, there is no discomfort to the patient.
Using OCT, an ophthalmologist can see each of the retina’s distinctive layers, which lets him map and measure their thickness. These measurements help form the doctor’s diagnosis of the patient, which includes potential treatment for glaucoma and other retinal diseases like age-related macular degeneration (AMD) and diabetic eye disease.
One OCT system, designed with the help of Vecow, was manufactured by a team with deep experience in medical technologies, and they were able to adopt the techniques needed to develop advanced ophthalmic surgery devices. Based in Taiwan, the vendor’s core business is the design, development, manufacture, and service of medical instruments. The OCT system they developed is based on an excimer laser, which is a type of ultraviolet laser.
The specifications in this case called for a high-performance platform for real-time medical-image computing, with the ability to support AI imaging and real-time inspection. They also required mission-critical, high-speed (real-time) data transfer for use in surgical applications, which dovetailed with flexible (custom) expansion features. This included Power-over-Ethernet (PoE) and wireless capabilities. The ability to be used in operating rooms meant that some level of mobility was expected, as well as a multitouch display, and adequate storage. Finally, the ability to upgrade the system later was a requirement, as systems in medical applications tend to be in use for many years.
OCT is Not MRI
It’s important not to confuse OCT technology with what’s deployed in modern magnetic resonance imaging (MRI) systems. The MRI will still play a major role in medical diagnoses, but there are some key areas where the OCT can be a more viable and valuable technique. Currently, OCT is used exclusively as a way to scan the retina and measure the thickness and volume of its individual layers.
Specifically, OCT gives medical personnel the ability to measure posterior eye shape based on just a small part of the eye. The result is a more complete three-dimensional model of the eye. Eventually, all patients—young and old—will have their eyes scanned, and based on the findings, the doctor can determine whether a condition like nearsightedness is actually a symptom of becoming pathologic in the future, with an immediate need for intervention.
The theory of OCT is that after the light is emitted, it gets divided into two beams. One light hits the fundus or base of the eye, while the second light hits a reference mirror. The two lights are then reflected back to form an interference phenomenon and a spectrum, with the resulting signal collected for analysis. The resolution used in OCT instruments is quite high, up to 2 to 5 microns.
Some Level of Medical Experience Is Needed
To design a system for an application like OCT, there’s obviously some level of medical knowledge required by the medical solution provider. However, in most cases, the system’s design team is not expected to have the training of a doctor. For example, a supplier like Vecow would provide most of the base hardware and software to its customers. Those OEMs then configure that platform for the specific application, such as OCT. Such a would likely require the OEM to provide parts of the software, including the AI algorithms. After that, some specialized code may be added or some tweaks could be made.
For this particular application, the OEM would generally specify an eight-core Intel Core i7 processor (or higher), at least 768 NVIDIA CUDA cores to handle the AI algorithms, and 16 Gbytes of DDR4 memory. Based on these specifications, Vecow’s EVS-2000 fits the bill perfectly. Its feature set includes:
- a ten-core 10th Gen Intel Xeon/Core i9/i7/i5/i3 processor, plus a W480E chipset
- compact NVIDIA Quadro/GeForce MXM graphics that supports CUDA core computing and NVIDIA’s Turing architecture
- 9- to 50-V dc power input with 80-V surge protection
- a host of expansion ports, including PCI/PCIe, M.2, Mini PCIe, and SUMIT A and B
- storage that’s handled through SATA, mSATA, M.2, MicroSD, and/or a 2.5-in. hard-disk tray
While a system like one that performs the OCT application is generally plugged into an ac mains, it’s still important to keep the power dissipation down to a minimum. Having that high level of power efficiency can reduce the number of fans or the time that those fans are active, which in turn minimizes the noise coming from the platform.
With Vecow’s EVS-2000, fanless operation is achievable. The design team was able to accomplish this using a 35-W Intel eight-core i7 processor and an 896-core CUDA device for the graphics that draws just 50 W. The result is a total power consumption for the system that’s below 300 W. The final end system delivered by the OEM also included some “hospital-ready” features, like a robotic delivery arm and eye-tracking.
Keep Security at the Forefront
There’s no question that security must always be front of mind when designing an embedded system. One that is used for medical applications is an even higher priority. Relieving some of the pressure for the Vecow team was the fact that the OCT platform that employed its technology was not intended to be connected to the Internet. Rather, it connected to a private medical cloud though an internal intranet. That same intranet would be used to handle software upgrades, which could be significant in a system like the one described, as investment protection.
The OCT platform is an example of the Vecow design team providing its customers with the latest techniques and technologies to ensure that the final products withstand the test of time in the field. This includes the only fanless solutions based on Core i7 processors. Vecow also offers the latest GPU cards on its workstation-grade platforms, including outstanding thermal performance.
To ensure that your products stay on the road to success for today, tomorrow, and in the future, discuss your needs with the team at Vecow. You’ll find a one-stop shop for all your needs.