Israeli Scientists Developing System to Detect Cancer Remnants During Surgery

Tel Aviv University-Ichilov Hospital collaboration harnesses AI, engineering and DNA sequencing to indicate if surgeon caught it all during the operation, rather than days later

Send in e-mailSend in e-mail
Send in e-mailSend in e-mail
Operating room at Kaplan, seen through porthole window
Operating room Credit: Moti Milrod

Cancer surgery comes with its own set of dangers, and one is whether any of the tumorous tissue has been left behind. Current methods to detect cancer remnants or even metastasis following removal of the primary tumor are slow, and the days or even weeks of delay between the operation and the pathology results can be critical for preventing a recurrence.

Now an interdisciplinary team of Israeli biologists, computer scientists, engineers and doctors are collaborating on the development of a new system to detect any cancer cells remaining in the abdominal cavity in real-time, during surgery.

The system is already undergoing initial testing but it could take some years before it reaches hospitals. Still, it is good to know that it’s in the works.

A team from Tel Aviv University’s Zimin Institute for Engineering Solutions Advancing Better Lives, led by Prof. Noam Shomron and doctoral student Artem Danilevsky, joined forces with Dr. Guy Lahat and Dr. Shelly Loewenstein at the Tel Aviv Sourasky Medical Center, better known as Ichilov Hospital.

To be clear, the team aspires to help quickly detect cancer remnants in general during surgery, not just in the case of abdominal cancers, but this is where they’ve started, to achieve their proof of concept.

That concept is based on the ability to quickly identify genetic differences between normal and cancer cells in the saline solution used to wash the internal abdominal organs during surgery, Shomron explains to Haaretz. The saline sample is put through a DNA sequencing device called MinION, which is manufactured by Oxford Nanopore Technologies.

The result – yes/no to “cancer sequences” in the sample,stemming from vestiges of the primary tumor to metastases – can take from minutes to several hours. Even at the longer end of the range, that’s a lot better than days or weeks.

And how is that result achieved? This is where the computer scientists come into the picture, as Shomron explains: The AI system is fed with data from cancer and normal cells.

It is similar, he says, to how Google knows whether that photo you posted is of a cat or a dog. Once the AI has learned by itself the differentiating and distinguishing features of catness and dogness, it can tell them apart in other photos it has never seen before. This is similar to how kids learn at an early age: Teach them a few times the difference between animals and then wherever they go, they will know whether the furroid before them is cat or dog.

“That’s what we’re doing with the computer system in this case: showing it cancer DNA and normal DNA, asking it to find the differences, and then it teaches itself to distinguish between normal and sick,” Shomron says.

Theoretically, the best forcomparison, in the case of treating abdominal cancers, is the same patient’s normal DNA, taken from white blood cells, or DNA found in the saliva. But there are global differences between cancer/not cancer, which the MinION machine will be able to identify.

Not to confuse the issue, but sometimes the difference between cancer/not cancer boils down to epigenetics.

“The difference might not be in the genetic code itself but in the methylation of the DNA, a kind of chemical marker along the DNA sequence. Our system can say, hey, the marking on this DNA looks like the marking of other cancer cells. This approach is the exact theme driven by the Zimin Institute for Engineering Solutions – to think outside the box in order to revolutionize current medical procedures,” Shomron adds.

Asked about the rate of potential false positives or negatives, Shomron drives home the point that the machine isn’t designed to be diagnostician, judge and jury. It is solely and only being developed to assist doctors, to help them reach decisions, to alert them that non-benign tissue may have been left behind. Then it is for the doctor to decide what to do, including, in the case of a positive result in real time during surgery, to wash out the patient’s innards some more.

It may take some more years until this cancer-detecting MinION reaches an operating room near you, but one comfort is that when it does arrive, it’s both mobile and relatively cheap. It’s early days to discuss pricing but the base system should cost a fraction of today’s genetic sequencing machinery, Shomron predicts.

Meanwhile, the pilot at Sourasky, in the operating room of Dr. Lahat using samples of abdominal cavity fluid collected by Dr. Loewenstein, has begun. If a sample tests positive for cancer-related DNA, Lahat employs abdominal cavity chemotherapy and saline washing in the operating room; then the staff take another sample and repeat the process – until they get a negative result. Which in this case is a good indicator.