AI cancer breakthrough set to save thousands of lives | UK | News

A groundbreaking AI-powered cancer testing system could help save lives by showing doctors which drugs are most likely to work before treatment begins.

The breakthrough technology allows scientists to grow tiny living copies of a patient’s tumor in the laboratory, then bombard them with dozens or even hundreds of different drugs and drug combinations to see what works best.

The AI ​​then quickly analyzes the results, giving doctors a much clearer idea of ​​which treatment is more likely to treat the cancer in question.

PreComb, the UK-funded company at the forefront of the research, says it is currently working on cancers including small cell lung cancer, bowel cancer, bone cancer and adult and child brain cancers.

Experts believe this could transform the care of thousands of patients, freeing up time or saving the lives of patients with potentially fatal tumors.

Around 412,000 people are diagnosed with cancer in the UK each year; This corresponds to one person every 75 seconds. More than 100,000 continue to receive chemotherapy.

But despite their widespread use, chemotherapy drugs are still given without doctors knowing in advance which drug or combination of drugs will work for each patient.

Because cancer is not a single disease. Even seemingly similar tumors can behave very differently depending on their mutations, resistance mechanisms and biology.

Many patients endure toxic side effects of drugs that ultimately do little to stop cancer.

Scientists now say a new approach known as ‘functional precision oncology’ could change this.

Instead of choosing drugs based primarily on what worked on average in previous patients, researchers can now test treatments directly on live tumor material taken from the patient himself.

Professor Javad Nazarian, an oncology scientist at the University of Zurich, is now using this technology for children with aggressive brain cancer.

“These are very lethal tumors and in many cases they are still incurable,” he said.

“Until now, we haven’t found a reliable way to see and measure how a patient’s cancer responds to drugs. We now have a new monitoring and AI-powered drug screening platform that allows us to test treatments directly on the patient’s tumor cells.”

Doctors remove a small piece of tumor tissue, and scientists grow live cancer cells from that tissue in the laboratory, effectively creating a miniature version—or “avatar”—of the patient’s cancer.

These ‘miniature tumors’ are then exposed to a wide range of drugs selected based on the genetic profile of the tumor.

Professor Nazarian said his team could test more than 120 drugs tailored to the patient’s tumor biology.

Unlike standard chemotherapy, which attacks all fast-growing cells, the goal is to more precisely target treatments that target the tumor.

The platform tracks what happens to cancer cells over two to three weeks.

Artificial intelligence analyzes thousands of images to see if they are growing, stabilizing or dying.

Professor Nazarian said: “Once the system is fully developed it will give doctors a much clearer view of what is going on.

“We can tell whether the cells are expanding or falling,” he said. “The AI ​​platform helps us see whether tumor cells are dying, proliferating, or becoming dominant. This can greatly improve the way we treat patients.”

For brain tumors, where clinicians often have only a narrow window to decide treatment after surgery, this speed can be crucial.

“We only have a few weeks to think about how to act in many situations,” he said. “This could be a game changer.”

Professor Nazarian revealed that a child who was expected to live just three months survived for 19 months after doctors used a screening approach to guide treatment.

“The tumor eventually recurred, but part of it disappeared, prolonging survival,” he said.

The PreComb system allows multiple drug tests to be performed simultaneously using automated robotics and “well plates” containing miniatures of the original tumor.

“Our platform produces approximately 5,500 images per test, and with the advent of artificial intelligence and machine learning, we can now process this data and derive meaningful results,” said Hal Bosher, CEO of PreComb.

He said the tests were performed by robots and the drugs were administered into hundreds of small wells containing tumor cells or tissue.

The company says it currently runs 371 scans producing more than 1.1 million tumor images across 272 different drugs.