Breast cancer women spared needless chemo after AI breakthrough

THOUSANDS of women with breast cancer could be spared unnecessary chemotherapy thanks to a new technique using artificial intelligence.

Scientists believe the new method could save the NHS millions of pounds each year because it offers a quicker and cheaper way of determining whether the disease is likely to return after surgery.

At the moment, when a woman has a breast lump removed, thin slices of tumour are sometimes sent to California for analysis to gauge how aggressive the cancer is.

The result is a ‘score’ for how likely a woman’s cancer is to return within ten years. If she has a high risk score, her oncologist will recommend she has chemotherapy to kill any remaining cancer cells.

But sending the samples abroad is expensive for the NHS and slow for the women, who face weeks of stressful waiting for the results.

In addition, a decision is often made without a sample being sent. In such cases, the oncologist will often err on the side of caution and recommend drug treatment.

While chemotherapy can be a lifesaver, it can trigger unpleasant side-effects including nausea, fatigue and hair loss, as well as damaging long-term heart health.

The new method, called Digistain, aims to give more women a thorough assessment of whether they need chemotherapy by providing a quicker, cheaper service to the NHS.

Dr Hemmel Amrania, its chief executive, said: ‘We estimate around 30 per cent of patients with hormone-positive early-stage breast cancer who get chemotherapy unnecessarily, could safely forego it thanks to this method.’

That could mean up to 4,000 women a year avoiding unnecessary chemotherapy, he added.

Each time a woman undergoes chemotherapy for early-stage breast cancer it costs around £1,000 in drug costs alone, according to a 2020 study.

This means the new test could produce annual savings of £4million, excluding labour.

The NHS’s total chemotherapy budget is around £1.5billion.

On top of that, Dr Amrania said the new test was far cheaper than the £2,000 to £4,000 cost of sending a sample to the United States for analysis.

The method, based on research undertaken at Imperial College London, automates the process of assessing how aggressive cancers are by scanning microscopically thin biopsy samples mounted on slides, using a machine no bigger than a desktop printer.

Just as an office scanner digitises a photo – recording the information as thousands of pixels – so the Digistain machine turns the biopsy slice into a digital file.

But rather than recording the colour and intensity of spots on the digital photo, it records patterns that show the presence of proteins associated with aggressive cancers.

By comparing thousands of digitised pictures, researchers have developed a computer program which ‘learns’ which images denote aggressive cancers.

A Midlands NHS hospital trust is to start using Digistain for breast cancer soon, Dr Amrania said, while the method is also being rolled out in several hospitals in the US and India.