MedtecLIVE

My Account

Language

Interview MedtecLIVE 2024

Endoscopy diagnosis and treatment with AI implementation

Advancements in endoscopy technology play a pivotal role in cancer prevention and diagnosis. Today it is possible to address common challenges like bulky-sized systems, high latency in computing, and reliance on human analysis by adopting Advantech's MIO-5377R SBC and MIOe-UMXM expansion module.

Background

Cancer is a leading cause of death worldwide, with gastrointestinal organ cancers like those affecting the colon, stomach, and rectum, among the most serious. Since prevention is often more effective than treatment, the minimal incisions and minimal patient discomfort of endoscopy technology is an increasingly popular option. Indeed, the global endoscopy device market is predicted to reach approximately US$47.7 billion by 2028, registering a projected CAGR of 6%. Endoscopy technology is also gaining traction because it enables surgeons to perform polypectomy procedures instantly, reducing the risk of cancer development. Today, both surgeons and patients are increasingly inclined to use advanced endoscopy devices that provide high-resolution images and real-time displays through camera sensors.

Challenges in existing endoscopy

Traditional endoscopy relies on visual analysis and doctor diagnoses, but it can be prone to human-related issues like fatigue and lack of experience. As a result, there is high potential for overlooking subtle anomalies, leading to missed opportunities for early-stage treatment.

To avert this situation, some endoscope equipment developers look to improve display resolution and incorporate AI recognition by selecting large systems integrated with PCIe GPUs. However, endoscope manufacturers can tap into greater gains by adopting a smaller and more versatile solution. A compact single-board computer (SBC) with an expansion module, for example, offers a far better option for medical imaging in endoscopic analysis applications.

Performance and compactness

Compact embedded boards are playing an increasingly pivotal role in endoscopy machines, where size is a critical factor. The latest SBCs empower the development of more streamlined endoscopes featuring a small footprint, typically within a 3.5" form factor. This advancement has led to substantial gains in mobility and improved control for surgeons.

The latest-generation mobile processors are at the heart of modern SBCs, delivering an impressively high EU (execution unit) capability when it comes to graphics. Some offer three times greater EU than desktop processors of the same generation.

Enhanced processing power is a vital asset in the field of endoscopy, facilitating demanding tasks such as image enhancement and real-time image stitching. Processors of such capability also operate with a low power draw thanks to a fan-less and low-noise thermal design. These attributes translate into markedly improved diagnostics and treatments.

As a further point of note, the latest SBC capabilities take productivity a step further by supporting four simultaneous displays with resolutions up to 4K. Such impressive innovation means surgeons have the ability to examine patients with exceptional clarity and precision, considerably enhancing the diagnostic capabilities of endoscopy systems. Another tip for designers of endoscope systems is to seek an SBC that integrates USB Type-C, thus simplifying camera connectivity and enabling better lighting and the use of finer lenses.

High-performance GPU expansion for enhanced imaging and AI capabilities

In the realm of medical imaging, endoscopy demands elevated graphics performance. However, many of today’s embedded boards are purposely large in order to accommodate PCIe or MXM expansion. In contrast, the latest GPU expansion modules connect easily to SBC, using a USB4 Type-C connector. Certain expansion boards have the capacity to support MXM GPU modules with power ratings of up to 115W, resulting in a remarkable 460% increase in graphics performance.

By incorporating this same type of expansion module with an MXM GPU, the display quality is elevated from DP1.4 at 3840×2160 and 60Hz, to DP1.4a, with identical resolution but at a faster refresh rate of 120Hz. Such an upgrade delivers clearer, smoother images to enhance precision, aid the visualization of internal structures, and improve the effectiveness of medical procedures.

A GPU add-on can also enhance machine learning (ML) algorithms, further strengthening the endoscope’s AI capabilities. It assists medical professionals in swiftly identifying anomalies, lesions, and potential issues in real time, thereby improving both diagnostic accuracy and speed. Moreover, an expanded GPU can accelerate image processing, enabling real-time rendering of 3D medical images and volumetric data. This capability proves indispensable for tasks such as surgical planning and examinations, allowing surgeons to gain a comprehensive picture of patient anatomy.

Leverage AI deployment with design-in service

Of course, when consuming large amounts of energy for AI and ML algorithms, it can prove challenging to source hardware devices with an optimal price-performance ratio. Even if all data sets, algorithms, training and UI/UX are functioning, how is it possible to facilitate the easy deployment of an AI application? There is a clear need to manage AI models efficiently, in addition to all remote hardware devices, such as sensors.

Fortunately, end-to-end device management and edge orchestration software solutions are now available to help power and scale AI/ML applications. As an added benefit, certain vendors, such as Advantech, provide this as a free tool. Able to enhance CPU and GPU performance, the software makes object identification and AI performance evaluation even more efficient. With just one click, users can test the AI capabilities of their edge devices using pre-trained models, saving time and reducing the cost of AI deployment.

Conclusion

Advancements in endoscopy technology play a pivotal role in cancer prevention and diagnosis. Today it is possible to address common challenges like bulky-sized systems, high latency in computing, and reliance on human analysis by adopting Advantech's MIO-5377R SBC and MIOe-UMXM expansion module. These compact yet powerful devices, driven by 13th Gen Intel® Mobile processors, enhance diagnostic precision, mobility, and efficiency for medical professionals. Furthermore, the optional GPU add-on boosts graphics capabilities, enabling real-time imaging and enhancing AI applications. Advantech's commitment to streamlined design and cutting-edge technology, along with the company’s free Edge AI Suite software tool, underscores its dedication to pushing the boundaries of medical imaging for improved healthcare outcomes.