New tumour removal surgery using artificial intelligence developed at UCD

Technique allows surgeon to ‘see’ tumours in real-time and comes with better success rate

A surgical technique using artificial intelligence, a camera, and dyes, which enables surgeons to “see” tumours in real-time and have a far better chance of fully removing them has been developed at UCD.

"The surgical team can better perfect the right intervention to the individual patient," explains Ronan Cahill, Professor of Surgery at UCD and the Mater Hospital, who developed the new approach with Jeffrey Dalli, General Surgeon and Surgical Research Fellow at UCD.

The scientific work on this technique was done at UCD and also at the Mater Hospital, with technological collaboration from IBM Research. "The technology will help surgeons better discriminate during operations what is best for each individual patient," says Mr Dalli.

Some 2,800 people per year are affected by colorectal cancer in Ireland, says Cahill, who has a special interest in these cancers. When cancer is not fully removed initially it becomes much harder to cure and further operations and treatments are needed, he says. Some 900 Irish people per year still die of the disease, he adds.

READ MORE

“Your best chance is to get the operation right first time,” says Mr Cahill. “It’s the law of diminishing returns if you have to go back and operate again.” The effectiveness of the technique in doing that in real-world patients is reported on Monday in Nature Scientific Reports.

The new method is particularly effective for colorectal cancers, which are increasing in incidence, especially among younger people, says Mr Dalli. There has been growing interest, he says, in perfecting surgical methods to define the exact extent of cancers during a procedure to ensure the maximum cancer is surgically removed.

Using existing technology

Up to now, patients with suspected colorectal cancer will undergo a colonoscopy. A tiny camera is inserted into the patient, the tissue is examined, and, if the surgeon believes, based on his experience, that cancer may be present a biopsy on suspected tissue is performed.

The biopsy samples are sent to the lab, explains Mr Cahill, where it can take between 48 and 72 hours to prepare specimens for examination under a microscope. Next, a pathologist will decide if the tissue is cancerous, or if more tissue - and a second biopsy – will be required.

“This takes time,” says Mr Cahill. “Wouldn’t it be terrific if we could just understand what the disease is, in a moment?” he asks. And, that, essentially, is what the new technique can achieve, he adds.

The beauty of the new technique is that it uses existing technology, fits in with current workflows, and is not expensive, says Mr Cahill. A camera takes a video of the suspect area while a specialised dye is infused into the tissue. Then, based on how the tissue changes colour, an algorithm determines the chances of it being cancerous.

Meanwhile, motion tracking technology keeps the camera tightly focused on the tissue of interest, even as the tissue moves due to heartbeats or breathing. The tracking permits tissue to be looked at for long enough to allow an analysis of its cancer risk to be defined.

The motion tracker software and cancer prediction algorithm were developed by UCD with IBM Research, in a collaboration funded by the Disruptive Technologies and Innovation Fund, a €500 million fund established under Project Ireland 2040 and run by the Department of Enterprise, Trade and Employment along with Enterprise Ireland.

“A few minutes is enough to determine if a lesion is cancerous,” says Cahill. “If it is there is no need to wait for a biopsy, we can remove it straight away. We also have a better chance of getting all of the cancer out first time and increasing a person’s chances of a cure.”

The technique has been applied to 200 patients at the Mater Hospital and Waterford Regional Hospital. Mr Cahill would like to see more hospitals adopt the method, as it's cheap and effective. In addition, the more video footage taken, the smarter the algorithms get and the better they become at making their cancer predictions.

This is another step towards computers helping surgeons practice better surgery and achieve better outcomes for patients, says Mr Cahill, yet he doesn’t fear computers putting him out of a job.

“I’m quite happy to be bettered by a smart system, as we all benefit from better health outcomes,” says Mr Cahill. “For the near-term, it’s all about human-machine synergy; a marriage of the best of both.”