AI could be used to detect blood diseases more accurately

How can we improve our ability to diagnose blood disorders? This question is being addressed by artificial intelligence (AI) by a research group headed up Helmholtz Munich. Their aim is to ease the time-consuming analysis of bone marrow cells under a microscope. Researchers created the world’s largest open-source database of microscopic images of bone cells. It is the basis for an AI model that has a great potential for routine diagnostics.

Every day, cytologists around the globe use optical microscopes to analyze and classify samples of bone marrow cells thousands of times. While this method of diagnosing blood diseases was developed over 150 years ago, it’s still a complex process. Looking for rare but diagnostically important cells is both tedious and time-consuming. Artificial intelligence can boost this method however, it requires lots of high-quality information to create an AI algorithm.

The largest open-source database of bone Marrow cell images

The Helmholtz Munich researchers created the largest open-access database of microscopic images taken from bone marrow cells. The database consists of more than 170,000 single-cell pictures of more than 900 patients suffering from different blood diseases. It is the result of a partnership between Helmholtz Munich with the LMU University Hospital Munich and the MLL Munich Leukemia Lab (one of the biggest diagnostic providers in this field globally) and Fraunhofer Institute for Integrated Circuits .

Utilizing the database to enhance artificial intelligence

“On top of our database, we have developed a neural network that outperforms the previous machine learning algorithms used for cell classification, not only in terms of accuracy, but also in terms of generalizability” says Christian Matek, lead author of the new study. The deep neural network is a machine-learning concept that specifically is able to process images.

Analysis of bone marrow cells is not feasible with these sophisticated neural networks. This is due to the deficiency of high-quality datasets.

Christian Matek, lead author

Researchers are planning to expand their bone marrow cells database to capture more findings and validate their model. The model and database are free for research and training purposes. They can be used to educate professionals and as an aid for AI-based strategies, e.g. in blood cancer diagnostics,” says study leader Carsten Marr.

Journal reference:

Matek, C., and. (2021) Highly accurate differentiation of bone marrow cell morphologies using deep neural network on a large image data set. Blood.

Content Source:

Gemma Wilson

Gemma is a journalism graduate with keen interest in covering business news – specifically startups. She has as a keen eye for technologies and has predicted quite a few successful startups over the last couple of years.

Related Articles