AJA provided by Baran used the Corvid 44 12G card in an Artificial Intelligence (AI) project to facilitate cancer diagnosis. Stepping into the health sector by crossing its borders, AJA showed how different sectors can work in harmony in its interview.

Artificial Intelligence (AI) achieves unimaginable successes. AI technology is starting to show great promise in healthcare, when paired with live, high-quality video, it helps practitioners identify health risks earlier and more accurately than the human eye. Realizing the endless potential of this technology, Oslo-based medical technology startup Augere Medical began developing an artificial intelligence-driven solution that alerted doctors to polyps that could develop into colon cancer by analyzing live video images from colonoscopies. We interviewed Andreas Petlund, CEO of Augere Medical, about the life-saving technology with an AJA Corvid 44 12G card at its center and compiled the highlights for you.

Can you tell us a little about yourself?

I’ve always had a passion for technology and this led me to pursue a PhD in Computer Science at one of Norway’s highest ranked research institutions. Like our CTO, who has a background in the film industry, I also specialize in multimedia data processing. Much of my work has focused on how I can apply the latest advances in space to an industry like healthcare. For several years, I led a research group that focused most of the work on an Artificial Intelligence solution used to detect colorectal cancer, one of the most common cancers today. In 2018 our findings inspired the launch of Augere Medical and we are now going through regulatory processes to make our product available in Europe and North America.

What prompted Augere to focus on colorectal cancer detection?

The medical community diagnoses 2 million cases of colorectal cancer each year, and the disease claims a million lives. It is also one of the most expensive and difficult to treat cancers and is often diagnosed at a late stage. Detecting colon cancer risks early can make the difference between life and death. Therefore, colonoscopies are a vital preventive measure.

The procedure is largely used to identify polyps or adenomas (cancer precursors) that are spotted in the colon. These precancerous abnormalities that are caught early can be easily removed to prevent the cancer from spreading. However, modern colonoscopy technology and the technique required to move the camera can make it difficult for doctors to detect every potentially problematic growth. For example, flat and sessile polyps are particularly difficult to detect during routine colonoscopies. Color and blood vessel patterns can help doctors understand different types of polyps. But even then, not many flat and sessile polyps are detected in colonoscopy procedures each year. This, This is a huge challenge facing the medical community, and one that we know an AI-driven video analytics tool can help solve. Our goal with this solution is to help doctors identify polyps more accurately and achieve more positive outcomes for patients.

So how does the solution work?

Typically, colonoscopy procedures move very quickly and certain images are available in mere seconds. That’s why we used machine learning to train our solution to detect as many hard-to-detect polyps as possible during routine colonoscopy procedures. Our solution consists of a physical box that connects to the facility’s colonoscopy equipment and monitor, as well as our proprietary video analysis software. Video is transmitted from the colonoscopy camera to the inside of the box and to our software, which analyzes each frame in real time. When the software flags a potential polyp, an alarm is activated in the box and a graph is drawn over the area. The feed is recorded so the doctor can further review the post-procedure images and revisit any areas of concern.

What technology powers the solution?

The facility’s AV needs vary depending on the colonoscopy equipment. That’s why we designed the solution to support both 4K and 1080p video. AI software powers the core functionality of the solution. But a standard PC with few additional components is supported by industry standard video hardware, including an NVIDIA Quadro RTX 4000 GPU card to support 50 video frames per second, and an AJA Corvid 44 12G I/O card for audio and video playback.

Corvid was a natural choice for the solution. Highly intuitive, it can reliably pass through 4K and HD video without signal dropouts, and the SDK is robust. Unlike other solutions on the market, Corvid’s SDK works independently of our software running the video analytics. This provides an additional layer of security against an unexpected software bug that could ultimately freeze the entire solution and interrupt the procedure. In general, it avoids the need to restart the box during colonoscopy and protects the patient from delayed examinations. Corvid’s video passthrough functionality also allows us to overlay the colonoscopy video signal with the lowest possible latency. Even a 100 millisecond delay can negatively affect the procedure, so this is huge.

What makes your technology unique?

Our video resolution processing capability is a key differentiator, and we are able to achieve such a high level of image accuracy through years of research into efficient processing. We also taught the system to analyze video with a temporal approach. This allows the system to look at polyps from several different angles and identify abnormalities in a few frames like a doctor would.

How do you see AI and video impacting the medical community in the future?

AI-driven video capture and analysis is one of the most groundbreaking developments in modern medicine. It allows medical professionals and institutions to use and analyze multimedia data like never before. We can now teach a machine to look for details that the human eye cannot always catch. This will be a major improvement in the prevention of colorectal cancer. AI is also used in radiology, genetics and other medical fields outside of our business. Of course, with this research comes privacy issues and other challenges that require problem solving. But the steps taken have the potential to fundamentally change the way we approach medicine.

What awaits Augere in the future?

We are currently focused on navigating this solution through regulatory approvals and getting it into the hands of doctors. However, we look forward to applying our findings to other areas of medicine that will benefit, such as cystoscopy and other endoscopic procedures such as laparoscopies. We are also building a cloud platform on top of our solution that will allow doctors to extract even more information from video images and provide patient-tailored treatment that takes examination history into account.