AI blood analyser outperforms humans detecting leukemia
- World Half Full
- Nov 25
- 3 min read
SCIENCE/HEALTH
An AI system that can examine the shape and structure of blood cells with higher accuracy and consistency than human specialists may significantly reshape how conditions such as leukemia are diagnosed. The tool, known as CytoDiffusion, built on generative AI, is designed to study the appearance of blood cells in detail.
Many current AI models focus mainly on pattern recognition, but the team of researchers from the University of Cambridge, University College London, and Queen Mary University of London have demonstrated that CytoDiffusion can recognise a broad range of normal blood cell variations and detect rare or unusual cells that may signal disease.
Identifying small variations in the size, shape, and overall look of blood cells is essential when diagnosing many blood disorders. However, developing the expertise required for this work takes extensive training, and even highly experienced clinicians can disagree when evaluating difficult samples.
“We’ve all got many different types of blood cells that have different properties and different roles within our body,” says the study’s first author, Simon Deltadahl from Cambridge’s Department of Applied Mathematics and Theoretical Physics. “White blood cells specialise in fighting infection, for example. But knowing what an unusual or diseased blood cell looks like under a microscope is an important part of diagnosing many diseases.”
However, a typical blood ‘smear’ contains thousands of cells — far more than any human could analyse. “Humans can’t look at all the cells in a smear — it’s just not possible,” says Deltadahl. “Our model can automate that process, triage the routine cases, and highlight anything unusual for human review.”
“The clinical challenge I faced as a junior hematology doctor was that after a day of work, I would face a lot of blood films to analyse,” says co-senior author Dr Suthesh Sivapalaratnam from Queen Mary University of London. “As I was analysing them in the late hours, I became convinced AI would do a better job than me.”
To develop CytoDiffusion, the researchers trained the system on over half a million images of blood smears collected at Addenbrooke’s Hospital in Cambridge. The dataset — the largest of its kind — included both common blood cell types and rarer examples, as well as elements that can confuse automated systems.
By modelling the full distribution of cell appearances rather than just learning to separate categories, the AI became more robust to differences between hospitals, microscopes, and staining methods, and better able to recognise rare or abnormal cells.
During tests, CytoDiffusion could detect abnormal cells linked to leukemia with far greater sensitivity than existing systems. It also matched or surpassed current state-of-the-art models, even when given far fewer training examples, and quantified its own uncertainty.
“When we tested its accuracy, the system was slightly better than humans,” says Deltadahl. “But where it really stood out was in knowing when it was uncertain. Our model would never say it was certain and then be wrong, but that is something that humans sometimes do.”
CytoDiffusion can also generate synthetic blood cell images that are indistinguishable from real ones. In a test with ten experienced haematologists, the human experts were no better than chance at telling real from AI-generated images.
“That really surprised me,” says Deltadahl. “These are people who stare at blood cells all day, and even they couldn’t tell.”
As part of the project, the researchers will make that massive dataset of peripheral blood smear images publicly available. “By making this resource open, we hope to empower researchers worldwide to build and test new AI models, democratise access to high-quality medical data, and ultimately contribute to better patient care,” says Deltadahl.
The researchers note that CytoDiffusion won’t replace trained clinicians, but rather support them by rapidly flagging abnormal cases for review and handling more routine ones automatically.
“The true value of healthcare AI lies not in approximating human expertise at lower cost, but in enabling greater diagnostic, prognostic, and prescriptive power than either experts or simple statistical models can achieve,” says co-senior author Professor Parashkev Nachev from UCL. “Our work suggests that generative AI will be central to this mission, transforming not only the fidelity of clinical support systems but their insight into the limits of their own knowledge. This ‘metacognitive’ awareness — knowing what one does not know — is critical to clinical decision-making, and here we show machines may be better at it than we are.”
The researchers say further work is needed to make the system faster and to test it across diverse patient populations to ensure fairness and accuracy.
Comments