0:08
NHS AI blood test could reduce invasive womb cancer checks
Several NHS hospitals are preparing to use an AI-powered blood test to help assess women referred for possible womb cancer before invasive checks are carried out.
According to The Guardian, around 90,000 postmenopausal women in England are referred by GPs each year for checks after experiencing heavy bleeding. About 10,000 women are diagnosed with womb cancer annually, and around 2,700 die from the disease.
The test, developed by Leeds-based PinPoint Data Science, uses machine learning to assess cancer risk from blood markers. It classifies patients as low, elevated, or high risk based on an analysis of around 30 markers.
PinPoint said the test costs around £30 and gives clinicians a risk score for use within existing cancer referral pathways. The score can help inform whether a patient is monitored, referred for further investigation, or prioritised for faster assessment.
PinPoint describes the tool as a multi-cancer test. The company said it has been used across gynaecological, lung, upper gastrointestinal, head and neck, and lower gastrointestinal cancer pathways.
The test is being introduced after a trial involving 16,481 patients referred through urgent suspected cancer pathways across Yorkshire. The trial included women referred with symptoms that raised concern about possible womb or gynaecological cancer.
About one in 10 women referred because of heavy bleeding were found to have cancer, according to the reported trial results.
PinPoint said the test correctly identified 99.1% of cancers as elevated or high risk and delivered a negative predictive value of 99.8% for women in the lowest-risk group.
Mid Yorkshire NHS Teaching Trust plans to use the test for six types of gynaecological or upper gastrointestinal cancer. Leeds Teaching Hospitals NHS Trust plans to use it for gynaecological cancer.
Under the current pathway, women referred for suspected reproductive system cancers usually undergo a pelvic examination that includes a transvaginal ultrasound scan. The procedure involves inserting an ultrasound probe into the vagina to measure the thickness of the womb lining, and some women find it uncomfortable or painful.
If doctors continue to suspect cancer, patients can then be referred for further checks, including a biopsy and hysteroscopy, an examination of the inside of the womb. PinPoint said its test is intended to identify women at very low risk before those procedures are used.
The company said the test could spare about one in five referred women from needing a transvaginal ultrasound scan. That would amount to around 18,000 women a year in England.
Professor Sean Duffy, chief medical officer at PinPoint Data Science and a former NHS England national clinical director for cancer, said the test’s value lies in ruling out women at very low risk.
Dr Jacinta Walsh, a GP at King’s Medical Practice in Normanton, West Yorkshire, said patients can require up to six GP visits before cancer is ruled out. She said the test could shorten that process and free up capacity for other patients.
Tracy Jackson, a consultant gynaecologist and cancer unit lead at Leeds Teaching Hospitals NHS Trust, said most women seen through the current referral route do not have cancer, while the investigations can be uncomfortable or distressing.
Jackson said the test could help clinicians triage patients before hospital-based investigations. She said low-risk patients could be ruled out in primary care, while higher-risk patients could be prioritised for further checks.
Recent NHS AI deployments include MEMORI at Kent and Canterbury Hospital, an AI triage tool in the NHS App, and AI-powered chest X-ray tools for suspected lung cancer pathways.
East Kent Hospitals University NHS Foundation Trust is using an AI system called MEMORI at Kent and Canterbury Hospital to assess infection risk from routine patient data. The system analyses information already included in patient records, including blood tests, blood pressure, temperature, observations, medications, and demographics.
NHS England said an AI triage tool in the NHS App is expected to reach more than 200,000 patients within 12 months and become available to all NHS App users by April 2028.
The government has also committed £20 million to roll out AI-powered chest X-ray tools to all NHS trusts in England by 2029. The tools are already available in about half of NHS trusts in England and have supported assessment for more than four million patients being investigated for lung cancer.
Further evidence will be needed to assess how the test affects patient outcomes, referral decisions, and NHS diagnostic capacity.
Cancer Research UK described the PinPoint test as promising but said more research is needed to understand its benefits for patients and the NHS.
4:11
Large Tabular Models Excel Where LLMs Fail
The large language models (LLMs) that form the basis of generative AI chatbots such as ChatGPT, Claude, and Gemini can generate uncannily human-like text and images. But these models still struggle with a skill that, ironically, looks at face value to be right in their wheelhouse: analyzing structured data. A new type of generative AI is set to change this situation.
Although you can get your favorite chatbot to solve intractable math problems, review dense legal documents, compose a catchy pop song, or put together some slick PowerPoint slides, give it anything more than a small table and it doesn’t have a clue what to do.
For most companies and organizations, the most important data sits in spreadsheets. Whether it’s a bank’s transaction logs, a marketing agency’s website metrics, clinical trial participants’ vital signs, or the vast amount of proton collision information produced at atom smashers like the Large Hadron Collider, structured, row-and-column data runs the world, and LLMs can’t deal with it.
AI startup Fundamental is pioneering a new type of AI foundation model, known as a large tabular model (LTM), to fill the gap. Fundamental came out of stealth mode on 5 February 2026 with US $275 million in funding and a model called NEXUS, purpose-built for tabular data. Now, the model is being adopted by companies such as Amazon Web Services, while others race to build their own LTMs.
Part of why structured data has garnered less attention is a very human bias, argues Boris van Breugel, a senior AI researcher based in Amsterdam. “People like to see images, videos, and ChatGPT responses,” he says. “But tabular data really lags behind because it’s not fun to look at numbers.”
Different tabular datasets are also difficult to compare, explains van Breugel, who co-wrote a prescient position paper on this topic in 2024. Whereas most language has similar semantics, making LLMs well-suited to being trained on vast amounts of text data, van Breugel argues that it is much harder to train a single tabular model on tables with very different variables.
Additionally, language is sequential by nature (as are music, images, and video). Changing the order of words in a sentence may change or completely destroy its meaning. But the structured data you find in spreadsheets isn’t sequential. You can swap the order of columns or play around with rows, but the underlying factual meaning of the data remains the same.
This independence from linear order is incompatible with an LLM’s fundamental purpose of predicting the next value in a linear sequence. “With LLMs, even slightly changing the input, you get a different output,” says Jeremy Fraenkel, CEO of Fundamental. “That’s fine, and actually often desirable for LLMs, but when you’re making a prediction of whether a transaction is fraudulent or not, you want to make sure that the prediction is the same, or deterministic, no matter what.”
Current tabular data solutions are limited to machine learning algorithms, such as XGBoost, that have been around for more than 15 years and are used by organizations globally. These algorithms—called gradient-boosted decision trees—have to be trained and optimized by data scientists over the course of months for each and every use case. In contrast, NEXUS and other emerging LTMs are foundational, leveraging learning amassed from pre-training on diverse databases so that they can be applied across a range of different predictive tasks with minimal bespoke feature engineering or task-specific model building.
And unlike LLMs, which primarily model sequences of tokens, LTMs model the structure of tabular data directly. They jointly learn from each entry’s numerical value, what it represents, and how it relates to other entries. For example, imagine an entry in a grocery stock inventory table for bananas: The LTM can take in not just the magnitude—say, 500—but the fact that the entry represents the current banana stock quantity, its category (produce), and the statistical properties that link the entry with the rest of the column. This contextual understanding enables more accurate reasoning and prediction over structured data.
According to Fraenkel, one of Fundamental’s biggest challenges in developing NEXUS was obtaining the right training data. Unlike natural language, which is abundant and broadly uniform in structure, tabular data is relatively hard to find—much of the data is sensitive or proprietary—and diverse. There are very few similarities between, for instance, a biology dataset and a financial one.
8:03
OpenAI's AI beats every human at AtCoder, a top competitive programming contest
At the AtCoder World Tour Finals 2026, an OpenAI system crushed all human competitors in an exhibition match, solving all five problems in the Algorithm Division. Two of those problems were rated exceptionally difficult by observers.
The article OpenAI's AI beats every human at AtCoder, a top competitive programming contest appeared first on The Decoder.
8:27
Databricks makes Chinese open-source model GLM 5.2 its default coding engine after it matched Opus at lower cost
Databricks benchmarked coding agents on its own multi-million-line codebase and found that the Chinese open-source model GLM 5.2 matched Anthropic's Opus 4.8 at $1.28 per task versus $1.94. The company plans to roll it out as a daily coding workhorse.
8:47
Where Does an AI’s Personality Actually Come From?
They aren’t designed, you can’t help perceiving one anyway, and that makes them an engineering problem almost no one is solving.
The post Where Does an AI’s Personality Actually Come From? appeared first on Towards Data Science.
9:03
Grok 4.5 is so cheap compared to Fable 5 and GPT 5.5 that benchmark gaps may not matter much
xAI releases Grok 4.5, trained on tens of thousands of Nvidia GB300 GPUs. In coding benchmarks, the model trails Fable 5 and GPT-5.5 but needs 4.2 times fewer tokens than Opus 4.8. At $2 per million input tokens, it costs a fraction of the competition. EU availability is expected in mid-July.
The article Grok 4.5 is so cheap compared to Fable 5 and GPT 5.5 that benchmark gaps may not matter much appeared first on The Decoder.
