Why Collaboration Matters for the Adoption of AI and Vision Systems in Medtech

Adding greater intelligence via AI technology to vision systems is opening up a new era of productivity and quality management, offering huge potential to highly regulated sectors such as life sciences. Such cognitive vision systems are already found in markets like automotive, being offered by leading companies such as Bosch, Siemens, GE and others. The value of this technology is evident by the fact that the AI computer vision market is predicted to reach $73.7 billion globally by 2027, with a CAGR of 37.8%.  Soon, machines may be able to adapt to unexpected changes in a visual environment. 

Hurdles for Medtech

A particularly promising area for cognitive vision systems is medtech due to the promise of lowering costs, creating higher product yields, faster time to market and better quality. However, there are some hurdles in front of medtech companies committed to exploiting this important technology.
 
These include: the need for further capabilities such as enabling the cameras to “see” the very subtle elements like different finish textures; assuring the system can consistently choose correct answers; and creating systems acceptable to medicine’s regulatory bodies, according to Domhnall Carroll, CEO of Digital Manufacturing Ireland (DMI), a national industry-led organization accelerating access to new digital technologies for manufacturers.  
 
“The cognition link is making machines almost self-aware” comments Carroll. “It’s allowing the machine to think for itself in relation to the space that it works in.  Adding vision means that a machine can do a lot of quality checking to ensure dimensional correctness for items such as catheters, stents and medical devices with very specific dimensions and finishes.”
 
There’s another challenge on the road to greater use of cognitive vision systems in medtech that involves the optimal approach to creating such products. According to Carroll, the answer is “building the shortest route possible between describing the need and coming up with a solution for realizing it,” with the best path being through collaboration. Specifically, collaboration between government bodies, research institutions, academia and industry to focus on industry led solutions and applications.

The Importance of Collaboration

As Carroll describes it, “The fact that there’s five or six or maybe 10 or 20 companies interested in an AI/vision system solution, it means that it’s more likely to have a cohort of companies willing to input into what the final research looks like. You have a much better chance of solving things at scale from the people who are investing in the solution. Often, manufacturers don’t have the capacity or resources solve issues, however they create the exciting and important challenges. In the best collaborations, they look to academic and researchers for the solutions. And they’re looking to tech companies. They’re looking to organizations like DMI to understand their issues and bring their technical tools and capability together to solve it. Our ability to resource and address the scope of a challenge grows when there are more companies interested in the solution.”
 
One of the notable vehicles for such collaboration is the Visual Cognitive Manufacturing Group (VCMG), recently launched by DMI that will help companies take advantage of technologies such as AI and vison systems on their manufacturing lines, in order to boost their competitiveness globally. Demonstrating the demand for these types of solutions, the VCMG group already includes global companies such as West Pharmaceutical Services, Boston Scientific, Medtronic, Abbott and Johnson & Johnson, along with top academic researchers and others. 
 
The collaborative opportunity is high through programs and initiatives like VCMG and others. These efforts include: projects fueled by the €500 million Disruptive Technologies Innovation Fund (DTIF) established under Project Ireland 2040; The national AI strategy for Ireland entitled “AI – Here for Good;” the SFI Research Centre for AI Driven Digital Content Technology (ADAPT); a rich environment of ongoing research at top institutions such as the bioelectronics cluster at Tyndall National Institute; and active R&D undertaken by some of the many leading medtech companies with locations in Ireland, who widely collaborate with government, researchers, universities and other firms.
 
Taking advantage of this collaborative environment is Medtronic, which used AI and imaging for applications such as surgical navigation, robotic-assisted surgery and various kinds of medical imaging. In fact, the company just received the AI Breakthrough award for having the best deep-learning platform. Specifically, its GI Genius system, which is an intelligent endoscopy module that uses AI to assist physicians in detecting polyps in real-time during a colonoscopy procedure. GI Genius is the first FDA-authorized computer-aided detection (CADe) system using AI to help endoscopists identify colorectal polyps, regardless of shape, size or morphology.
 
Ireland’s collaborative focus on medtech and AI can be seen in many promising, important projects now underway. For example, DTIF has funded a project involving IBM, Royal College of Surgeons, a University College Dublin clinician and Dublin-based Deciphex for the early diagnosis of colorectal cancer using AI and imaging technology. Deciphex makes software to improve productivity in preclinical/toxicologic pathology. This effort starts with the vast array of imaging usually produced for cancer patients – MRIs, CTs, etc. – and uses machine learning to identify anomalies and respond much more quickly, producing early diagnoses.
 
Another company exploring smarter systems to empower its patients, customers and people is Boston Scientific. Medical device manufacturing as we know it today is changing rapidly and AI has the potential to improve efficiency.  For example, vision systems with AI cognition can offer significant benefits in manufacturing by enhancing product quality through real-time defect detection and production downtime reduction. As the systems progress, the company anticipates they will greatly enable manufacturing engineers to improve throughput and quickly adapt to changing production requirements, which ultimately help to realize significant operational efficiencies.
 
Given the volume of imaging performed in medicine today, the new work in intelligent vision systems can make a huge difference in accuracy, speed and cost. When vision systems are applied, they will take a picture or a scan or look at results as they’re being quality checked. The AI piece will identify and learn about anomalies like cracks, tears and weaknesses, and then find them quickly. This work is usually done manually, which makes it susceptible to fatigue and errors besides being labor intensive. At scale, it’s a challenge to process all the images normally produced in clinical settings so having new, accurate methods that take much less time is important.
 
Diagnostically, the overwhelming evidence around AI is it augments the diagnostic experience from a clinician’s perspective as well as from the patient’s perspective. In healthcare, this technology takes over routine tasks and allows medical professionals to focus on high-quality and complex care. It also speeds up diagnosis, which is critically important. Applying such methods to early diagnosis within the human body is essential given the high stakes involved.
 

---

Rachel Shelly is Global Head of Life Sciences & Food and Talent, Transformation & Innovation at IDA Ireland.