Key Takeaways
- A new AI system, CytoDiffusion, can diagnose diseases such as leukemia with greater accuracy and consistency than human specialists.
- The system uses generative AI to analyze blood cell appearance in detail, studying subtle variations in how cells look under a microscope.
- CytoDiffusion can identify abnormal cells with higher sensitivity than existing systems and quantify its confidence in its predictions.
- The system is designed to assist clinicians, not replace them, by quickly flagging concerning cases and automatically processing routine samples.
- The researchers are releasing the world’s largest publicly available collection of peripheral blood smear images to empower researchers worldwide to build and test new AI models.
Introduction to CytoDiffusion
A new artificial intelligence system, CytoDiffusion, has been developed to examine the shape and structure of blood cells, which could significantly improve the diagnosis of diseases such as leukemia. The system uses generative AI, the same type of technology used in image generators such as DALL-E, to analyze blood cell appearance in detail. As Simon Deltadahl from Cambridge’s Department of Applied Mathematics and Theoretical Physics, the study’s first author, notes, "We’ve all got many different types of blood cells that have different properties and different roles within our body… But knowing what an unusual or diseased blood cell looks like under a microscope is an important part of diagnosing many diseases."
Moving Beyond Pattern Recognition
Unlike many existing medical AI tools, CytoDiffusion does not focus only on obvious patterns. Instead, it studies subtle variations in how cells look under a microscope, allowing it to recognize the full range of normal blood cell appearances and reliably flag rare or unusual cells that may signal disease. As Deltadahl explains, "Humans can’t look at all the cells in a smear — it’s just not possible. Our model can automate that process, triage the routine cases, and highlight anything unusual for human review." This approach has the potential to reduce missed or uncertain diagnoses, which can have serious consequences for patients.
Training on an Unprecedented Dataset
To build CytoDiffusion, the researchers trained it on more than half a million blood smear images collected at Addenbrooke’s Hospital in Cambridge. This dataset, described as the largest of its kind, includes common blood cell types, rare examples, and features that often confuse automated systems. As co-senior author Dr. Suthesh Sivapalaratnam from Queen Mary University of London notes, "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 analyze… As I was analyzing them in the late hours, I became convinced AI would do a better job than me." The researchers’ use of this large dataset has allowed them to develop a system that can model the entire range of how blood cells can appear, making it more resilient to differences between hospitals, microscopes, and staining techniques.
Detecting Leukemia with Greater Confidence
When tested, CytoDiffusion identified abnormal cells associated with leukemia with much higher sensitivity than existing systems. As Deltadahl notes, "When we tested its accuracy, the system was slightly better than humans. 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." This ability to quantify its confidence in its predictions is a key advantage of CytoDiffusion, as it allows clinicians to make more informed decisions about patient care.
The Future of Blood Analysis
The researchers are releasing the world’s largest publicly available collection of peripheral blood smear images, totaling more than half a million samples, to empower researchers worldwide to build and test new AI models. As Deltadahl explains, "By making this resource open, we hope to empower researchers worldwide to build and test new AI models, democratize access to high-quality medical data, and ultimately contribute to better patient care." The team notes that additional research is needed to increase the system’s speed and to validate its performance across more diverse patient populations to ensure accuracy and fairness.
Supporting, Not Replacing, Clinicians
Despite the strong results, the researchers emphasize that CytoDiffusion is not intended to replace trained doctors. Instead, it is designed to assist clinicians by quickly flagging concerning cases and automatically processing routine samples. As co-senior author Professor Parashkev Nachev from UCL notes, "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… 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." The team’s work has the potential to improve patient care and outcomes, and its release of the large dataset of blood smear images will likely accelerate the development of new AI models for blood analysis.
https://www.sciencedaily.com/releases/2026/01/260112214317.htm
