Fighting blood diseases with AI

Nov. 24, 2021

How can blood diseases be better diagnosed? A research group headed by Helmholtz Munich is investigating this question. Its aim is to facilitate the time-consuming microscopic examination of bone marrow cells using artificial intelligence (AI), according to their press release.

To this end, researchers published the largest publicly accessible database to date with microscopic images of bone marrow cells. They use this as the basis for an artificial intelligence (AI) model with high potential for use in routine diagnostics.

To diagnose blood diseases, hematology laboratories around the world manually classify bone marrow cells thousands of times a day, a method that has been established for more than 150 years. Trained staff analyze stained specimens of bone marrow cells under a light microscope. This is a complex and time-consuming process - especially when looking for rare but diagnostically relevant cells. Artificial intelligence could become an important cornerstone of diagnostics. However, so far there has been a lack of quantitative and qualitative sufficient data for training the corresponding algorithms. The database consists of more than 170,000 single-cell images from more than 900 patients who suffer with various blood diseases. It is the result of a collaboration from Helmholtz Munich with the LMU University Hospital Munich, the MLL Munich Leukemia Lab (one of the largest diagnostic providers in the field) and Fraunhofer Institute for Integrated Circuits.

Until now, looking for rare but diagnostically important cells has been a laborious and time-consuming task. The researchers believe artificial intelligence has the potential to improve the process. What has been needed, however, is enough high-quality data to train an AI algorithm.

The researchers aim to further expand their bone marrow cell database to capture a broader range of findings and validate their model.

Helmholtz Munich release

Photo 106117755 © Jakub Jirsak |
Photo 89319446 © Mohamed Ahmed Soliman |