Peterborough launches AI trial to help cancer patients

Peterborough City Hospital's Dr David Bailey says there is a "desperate shortage" of pathologists and AI will help to cut patient waiting times

Cutting NHS waiting times for cancer diagnosis will be difficult without artificial intelligence (AI), according to one of the region's leading pathologists.

Dr David Bailey has launched a trial of the technology for breast cancer patients at Peterborough City Hospital.

AI will be used to interpret scans and order follow-up tests before a consultant in involved.

The hospital is one of five to win government funding for the study.

Each year almost 56,000 people in the UK are diagnosed with breast cancer, according to Cancer Research UK.

As with other cancers, the earlier it is detected, the more likely it is that treatment is successful.

The government is funding a range of projects which explore whether AI - computer programmes which use data to predict patterns and "think" like a human - can be used to speed up that process.

• AI cuts treatment time for cancer radiotherapy

• What is AI and is it dangerous?

• Ex-cancer surgeon treated for disease for third time

For the next year, Peterborough City Hospital will work with a system known as Galen Breast, which is already licensed for use in other countries.

It will be used in two different ways to support the hospital's pathologists.

Firstly, it will analyse digital images of biopsies and, if it detects cancer cells, it will order extra tests before a consultant has been involved in the case.

Secondly it will be used as "the second pair of eyes".

Dr Bailey, the hospital's director of pathology who is leading the project, said "two pathologists will always review a patient's biopsy images as a quality measure" and using AI instead of one of them frees up staff.

"The AI technology will increases our productivity by 15% to 20% and offers us the best of both worlds - the reliability and reproducibility of machines with the intuition and insight of humans."

When asked whether patients could trust the AI's interpretation, Dr Bailey said: "Yes. It's been shown to be very sensitive so it flags any potential cancer even if it turns out to be nothing."

AI analyses the biopsy scan and colours it with a heat map - red indicates a high probability of cancer and blue is low probability

Patients with suspected cancer should have their diagnosis within 28 days according to government targets.

For patients with suspected breast cancer who are referred to the North West Anglia NHS Foundation Trust, which runs Peterborough City Hospital, 81% are diagnosed within 28 days, but Dr Bailey said it could take years to address the wider general pathology backlog without AI.

"Pathologists have a huge amount of work - there was a backlog of patients before the pandemic and Covid-19 made that worse," he said.

"AI can't put pathologists out of work because even with the technology there just aren't enough of us."

Pathologists across five NHS trusts will use AI to analyse a total of 10,000 biopsies as part of routine practice and evaluate how it affects the quality, speed and efficiency of diagnosis.

Cambridge University Hospitals NHS Foundation Trust, which runs Addenbrooke's Hospital is also taking part.

Follow East of England news on Facebook, Instagram and X. Got a story? Email [e-mail] or WhatsApp [phone]