Thousands of women could be spared painful cancer exam by new NHS AI blood test

An NHS trial of a Leeds-developed blood test could spare thousands of women from painful and invasive cancer investigations if the results hold up in wider use.

The test, made by PinPoint Data Science, uses machine learning to analyse 30 blood markers and is being positioned as a triage tool for women referred with suspected gynaecological cancer. In the reported trial, which involved 16,481 patients referred by 170 GP practices in Yorkshire, the test was said to have identified and ruled out gynaecological cancers with 99% accuracy in a subgroup of 3,313 women referred because bleeding raised concern about womb cancer.

How the test works

The goal is not to replace specialist care, but to help doctors sort lower-risk patients from those who need faster investigation. That matters because women with symptoms that could indicate womb cancer are currently sent on to pathways that can include transvaginal ultrasound, biopsy and hysteroscopy.

The blood test is designed to spot very low-risk women earlier, potentially reducing the number who go on to those examinations. The current diagnostic route can be uncomfortable for patients and adds pressure to already stretched services.

Why the result matters

The Guardian reported that about 90,000 postmenopausal women a year in England are referred for suspected womb cancer because of heavy bleeding. It also said the test could spare about 18,000 women a year from transvaginal ultrasound scans if adopted more widely.

The story is therefore not just about a technology claim. It is about whether a blood test could shorten the route to reassurance for some patients while helping clinicians focus resources on people more likely to have cancer.

The trial and the claims

The reported trial covered 16,481 patients referred through primary care in Yorkshire, making it a substantial study by diagnostic-testing standards. The strongest headline result was the reported 99% accuracy figure in the subgroup of women assessed for possible womb cancer.

That number is promising, but it is not the same as a full explanation of the test's performance. The public reporting does not yet spell out whether the 99% figure refers to sensitivity, specificity or another measure in the underlying study report.

Trusts preparing to use it

The test is already moving beyond a purely research setting. Mid Yorkshire NHS teaching trust is planning to use it for gynaecological cancers or upper gastrointestinal cancer, while Leeds teaching hospitals NHS trust plans to use it for gynaecological cancer.

That makes the development more immediate than a distant proof-of-concept. If the test performs as reported in routine use, it could affect waiting lists as well as patient experience by trimming the number of unnecessary scans.

Caution and next steps

Cancer Research UK said the test appears promising, but more research is needed to understand the benefit for patients and the NHS. That caution is important because diagnostic tools can look strong in a trial and still need further validation before they become routine practice.

The key unanswered questions are straightforward. It is not yet clear whether the underlying study has been published or remains an internal or conference-stage report, and the public record still needs a fuller breakdown of the methods behind the headline accuracy figure.

For now, the case for the test rests on a large reported trial, an immediate use case for women referred with suspected womb cancer, and early adoption plans at two NHS trusts. The next step is to confirm how it performs in wider NHS use and whether it can reliably cut invasive investigations without missing cancers.