Diagnostics, Artificial Intelligence

AI-powered medtech company aims to transform breast cancer surgeries

Toronto-based Perimeter Medical Imaging wants to make breast conservation surgeries more efficient by combining an optical coherence tomography device with an AI engine to reduce the number of repeat surgeries required.

breast cancer

The possibility of repeat surgeries to remove cancer from women diagnosed with breast cancer is a reality of breast conservation surgery. In the U.S. about 25 percent of women need additional surgery after a lumpectomy, which is done to preserve breast tissue after an initial breast cancer diagnosis.

This is because, following a lumpectomy after which the patient is sent home, a pathologist’s report days or weeks later may find cancer at the margins of the excised breast tissue. And that positive margin means the patient has to undergo another surgery to fully remove cancer or undergo a full mastectomy. All of which is emotionally taxing for patients

sponsored content

A Deep-dive Into Specialty Pharma

A specialty drug is a class of prescription medications used to treat complex, chronic or rare medical conditions. Although this classification was originally intended to define the treatment of rare, also termed “orphan” diseases, affecting fewer than 200,000 people in the US, more recently, specialty drugs have emerged as the cornerstone of treatment for chronic and complex diseases such as cancer, autoimmune conditions, diabetes, hepatitis C, and HIV/AIDS.

“They trust their surgeon that they are going to remove it and it’s a very difficult phone call that the surgeon has to make to the patient to say ‘I missed some of the cancer, so you have to come back for a surgery.'” said Andrew Berekely, co-founder of Perimeter Medical Imaging a Toronto startup that is developing an AI-powered solution to the problem of repeat breast cancer surgery. “That’s not a phone call that surgeons like to make.”

Perimeter Medical went public in a reverse takeover of a public company this week and is expected to start trading on the TSX Venture Exchange in Canada on July 6. The company raised private capital of C$10 million (about $7.3 million) from investors including Roadmap Capital, shareholders of the public company it took over and high net worth people. Perimeter Medical has also received a $7.4 million grant from the Cancer Prevention and Research Institute of Texas, according to Berkeley. The company has its U.S. offices in Dallas.

The technology that Perimeter hopes to commercialize includes a medical device that can perform optical coherence tomography (OCT) on excised breast tissue and then feed that imaging data into its AI engine to help the surgeon understand whether an area of the breast still retained in the body should be further investigated — in other words, does the tissue in the margins contain cancer. That real-time information can help to determine whether the surgeon needs to take out more of the breast tissue thereby perhaps reducing the chance of another surgery down the road.

“The OCT is used to scan the back of your eye which is a 1 cm area. We have adapted the technology to scan very large complex surfaces like a removed tissue specimen in a very fast amount of time and give information back in the operating room where it is a time-sensitive situation …,” Berkeley said.

In other words, the goal is to use the AI engine to speed up the turnaround instead of having to wait for pathology reports.

“We have custom-built machine learning algorithms specifically for OCT that are specifically trained on breast tissue and this is years worth of working and selecting clinical data. What we do is that we image the tissue and then we take the post-operative pathology – what the pathologist looks at under the microscope — and we do what’s called a correlation,” he explained. “So we take the exact same area [that a pathologist finds] when they see tissue under the microscope, [and] we can go back to the same area in the images and we label in our images what disease looks like and we feed that into the ML algorithm and the more we do that the better it gets at identifying new diseases.”

With the $7.4 million grant, the OCT device called Otis will be placed in pathology labs at MD Anderson Cancer Center, Baylor University and the University of Texas at San Antonio to collect the imaging data needed to train the AI algorithm. Following that, a clinical trial will test the AI’s ability to predict accurately the status of the margins of breast tissue against the results of the pathology reports for those same patients. The trial will be done in 10 sites mainly in Texas but also in Mount Sinai in New York and the University of Wisconsin, he said.

“We have FDA clearance on the imaging hardware but we don’t have the clearance for the algorithm yet,” Berkeley said. “So the study is to get the algorithm approved.”

The problem of repeat cancer surgeries has been something that other companies have also taken a shot at. Most notably Israel-based Dune Medical with its MarginProbe product has been at the forefront of this effort though the company uses RF technology and not OCT. There isn’t an AI element in the technology developed by Dune Medical, which was acquired by Dilon Technologies in April for an undisclosed sum. It appears that when the company’s assets were put up for sale, one goal was to pay back $5 million in a loan that was still outstanding since 2015 to a debtor. The FDA approved Dune Medical’s MarginProbe in 2013 and $100 million was invested in developing the technology.

A breast cancer surgeon who has used MarginProbe and has also appeared in a YouTube video for the company said she used the technology sparingly.

“We use Dune RF technology but because of the cost, I usually only use it on intraoperative radiation patients where the margin is more important for me because I am radiating the breast at the same time,” said Dr. Michele Carpenter, program director for the breast program at St. Joseph’s Center for Cancer Prevention and Treatment in Orange, California, in an email forwarded by a representative. “The largest reason I have problems with it is its utility in very dense breasts seems to have more false [positives] as well as the cost of the disposable portion of the device.

And that is where Perimeter Medical Imaging — assuming it’s AI tool gets the nod from the FDA — hopes to make inroads. Berkeley said he expects the hardware piece of his company’s technology – the Otis device – would be priced similar to a specimen radiograph which ranges between $130,000 and $180,000. While that is is more expensive than Dune Medical’s hardware, the per-use cost based on a single-use consumable would be lower than that of Dune Medical, he said.

“At a medium volume center over a five-year period, Perimeter’s technology will be substantially lower cost,” Berkeley noted.

The technology might also be helpful when surgeons don’t have that much experience. Carpenter said that at her institution, repeat cancer surgeries are around 10 percent or less, which she attributed to training and experience. That’s far lower than the national average of nearly 25 percent per data published in JAMA Surgery referenced earlier. 

Higher rates have a financial ramification in addition to the emotional toll on patients having to return for a second surgery. According to a study published in 2017 in JAMA Surgery entitled Beyond the Margins—Economic Costs and Complications Associated With Repeated Breast-Conserving Surgeries, the average cost of additional surgeries is $16,072 in added healthcare costs.

“Indeed, reexcision after BCS owing to margin status has been deemed ‘the other breast cancer epidemic,'” the study noted.

Perimeter Medical Imaging has a ways to go before it can prove that its Otis device and companion AI engine can actually reduce the number of additional surgeries but the economic future appears to be clear.

“Value-based medicine is going to come much faster than any of us anticipated and what’s going to happen with value-based medicine for breast cancer is that surgeons are not going to get paid for that second time,” Berkeley declared. “So this will be shifted from the payer to provider — to the hospital — and the provider is going to have to take a look at the tools they have internally and say, “Are we able to make sure that this patient is going to be treated  pretty much  with the dollars that the payer is providing?'”

Photo: belchonock, Getty Images

CORRECTION: Due to incorrect information provided, the article has been updated.