By Ariel Grossman, NoCamels -
The weeks it takes to receive biopsy results for a cancer diagnosis can be debilitating.
Patients are plagued by anxiety, depression, stress, and sleep disturbances. And they worry that every extra day they wait could delay the start to their treatment.
But AI currently being tested in Israeli hospitals could change that, slashing the waiting time from four weeks to just two minutes.
The algorithm is able to make a near-instant diagnosis. And it has a major advantage over traditional laboratory processes – it identifies the type of mutation that has caused the cancer to develop.
That information allows doctors to prescribe precision treatments, based on the genetic makeup of the tumor to select the most safe and effective therapy.
That’s something pathologists cannot ascertain from looking through a microscope alone.
Imagene AI, the company behind the algorithm, was established after CEO Dean Bitan found his mother had been diagnosed with Stage 4 ovarian cancer.
He navigated the treatment and clinical process with her, trying to understand how physicians made certain decisions, and the options they were able to provide for patients.
She unfortunately passed away several years ago. But he knew that things should – and could – be better, for both the patients, and the physicians.
“I believe that this is the part that we’re not talking about enough,” he says. “Behind the numbers, and behind the statistics, there are real patients who are suffering while they wait, and we really need to assist them.
“We have to ask ourselves whether cancer patients receive the right treatment for them at the right time. And I believe that this is not always the case.”
It takes the AI two minutes to analyze a digital biopsy, compared with routine lab testing, which takes four weeks on average.
He says one patient in clinical trials started treatment the day after Imagene’s AI checked her biopsy.
The technology can currently identify mutations in the cells of eight different organs, and more than 28 different biomarkers – specific biological molecules produced by the body of a person with cancer. Identifying biomarkers can help physicians provide more precise treatments.
Imagene’s initial focus was on lung and breast cancer, but the AI they developed can also perform a rapid molecular analysis on thyroid, CNS (brain and spinal cord) colon, hematologic (cancer that begins in blood-forming tissue, like bone marrow), bladder, and ovarian cancers.
“We have basically developed an algorithm that can detect the patterns of those cancer mutations out of the biopsy image alone,” Bitan tells NoCamels, “and by doing so, we can recommend the right treatment.”
Imagene trained its AI to recognize mutations in a digitized biopsy image by showing it many pictures of tissue biopsies at different angles. It can conduct a full molecular analysis on a digitized biopsy image in real time and return a result in minutes, instead of needing several weeks to process it.
Pathologists and oncologists aren’t able to look at a slice of tissue on a glass slide and determine whether the cancer developed because of a certain mutation – but the AI can.
So if someone has cancer that has developed because of a mutation in the gene, the AI will be able to determine that and highlight it in the digital biopsy.
Oncologists can use the accurate diagnosis and initiate a precisely targeted therapy plan sooner than ever before, and the algorithm can be integrated into any pathologist’s current workflow without needing to change the way they operate.
Bitan says only around 15 per cent of patients will be diagnosed with a biomarker that can lead them to receiving precision medicine therapy. He’s hoping that Imagene will make these treatments more accessible and available.
“Unfortunately, the technologies that are currently being used over the start-of-care are complex, cumbersome, they take time, and they demand a certain amount of tissues in order to provide those assays (a laboratory test to find and measure the amount of a specific substance).
“And this is why we see that so many patients just don’t get the results – because of the time it takes, the complexity of testing, and because of the costs. This is where Imagene enters the picture.”
When a patient’s biopsy is sent to the lab today, pathologists use expensive machines and complicated processes to convert the tissue into a card that represents our DNA. Afterwards, there are biologists who try to analyze whether mutations are present within that card – and eventually provide a report that is sent to a physician.
“It’s changing the game in terms of the time, costs, processes, and workflows that currently happen in the pathology lab.
“In my opinion, that’s really important. The clinical outcomes will definitely improve with this – it will save their lives and prolong their lives to be treated earlier. But a second, and equally important part, is the emotional part.”
There are several companies using different methods to cut down cancer diagnosis times, but Bitan says that Imagene’s approach is different.
He says that having employees from different disciplines – like biologists, physicians, data scientists, software engineers – all working towards the same common goal, like his company has, helps Imagene solve complicated problems, and lets them operate differently compared with other companies.
“So many people from the company have some type of experience relating to cancer – and these experiences are really what drives us forward.”
Imagene’s AI has been implemented in hospitals under clinical trials, including Tel Aviv Sourasky Medical Center, and Sheba Medical Center in Israel, as well as several medical facilities and pharmaceutical companies worldwide.
The company is currently working on receiving regulatory approval from health authorities. Bitan hopes that his solution will be implemented within hospitals around the world within five years.
