Using AI techniques to improve the diagnosis of myeloproliferative neoplasms (MPN)
Diagnosing myeloproliferative neoplasms currently relies on bone marrow biopsy and the human eye. In this project, researchers will look for new, more accurate ways to diagnose the disease.
The challenge
Myeloproliferative neoplasms (MPN) are a group of blood cancers affecting around 5,000 people every year in the UK. People with the disease are usually diagnosed after a bone marrow biopsy where blood cells in the spongy parts of our bone are taken away and reviewed by a specialist using a microscope. Relying on the human eye however can be prone to error and there’s a need to develop new techniques that can more accurately diagnose MPN.
The project
In this project, Dr Daniel Royston and his team will test a new technique to do this. He has developed a tool using artificial intelligence that can scan images of the bone marrow and describe what the cells look like. In initial testing, the tool has looked promising, and he now wants to test it further using large numbers of bone marrow samples donated by people with MPN.
The future
If this tool works, it could be used to guide the diagnosis of MPN in the future. It could also be used to help guide treatment decisions as the tool would be able to quickly identify whether a treatment is working. If it’s not, people can be switched onto something else. The team also want to use this tool to help people with MPN understand what’s happening in their bone marrow, so they can understand more about their disease.