Adrian Mendes, CEO of Perimeter Medical Imaging AI, is solving the problem of margin assessment in cancer surgery. Surgeons often have to perform a second surgery on breast cancer patients because they are unable to completely remove all of the cancerous cells the first time. The company has developed Optical Coherence Tomography (OCT) imaging technology that allows surgeons to see cells at a microscopic level in real-time during surgery, helping them remove the tumor and all surrounding cancerous tissue.

Adrian explains, "The surgeon is trying to ensure that when they extract the tumor and the cancer, there is a margin of healthy cells around it. Studies have shown that if they can achieve that, and with breast cancer, that margin needs to be two millimeters generally, then the likelihood of them having left cancer cells back in the body goes down drastically. This is every surgeon's objective for cancer treatment. We help the surgeons ensure that they've achieved what they call clean margins."

"Yes, it's quite significant. So, for breast cancer surgery alone, it's about one in every five surgeries are unsuccessful because the margins aren't clean, and then the patient has to come back for a second surgery. So, there are about 300,000 breast cancer surgeries per year in the United States. If you think about 20% of that, there are a lot of women that are having to go back for a second surgery every year. That's just in the United States. And it's a global problem."

"What's unique about our technology is we use an imaging tech called OCT or Optical Coherence Tomography that allows images to be created down at around 15-micron width level. And that's small enough to be able to distinguish cells. And so with that, it just gives the surgeon so much more ability to see what they're looking at."

"The next generation uses the same imaging modality, the OCT imaging, but it adds an artificial intelligence and AI layer to it. So what we've done is trained an AI algorithm to recognize suspicious areas inside images of breast tissue. And these suspicious areas are indicative of cancer. The way we train the model is we have a library of about 2 million images of breast tissue, both cancerous and healthy tissue, and we've got labels that have been provided by pathologists."

#PerimeterMedical #Oncology #BreastCancerSurgery #OCT #Tumors #MedAI

perimetermed.com

Download the transcript here