29 November 2021

Progress has been made on the development of automated technology to diagnose blood disease from bone marrow cells, German researchers have reported.

Bone marrow cytology is important in the diagnosis of malignant and non-malignant blood diseases. However, it has been difficult to automate and is still mainly carried out by human experts, according to the research team led by Dr Carsten Marr from Helmholtz Zentrum München, Germany.

“Manual evaluation of bone marrow smears can be tedious and time-consuming and are highly dependent on examiner skill and experience, especially in unclear cases,” they write.

“Hence the number of high-quality cytological examinations is limited by the availability and experience of trained experts. Furthermore, examination of individual cell morphologies is inherently qualitative, which makes the method difficult to combine with other diagnostic methods that offer more quantitative data.”

The training of automated tools has been hampered by a lack of large numbers of good quality annotated images to work from.

So Dr Marr’s team created an open access dataset of 171,374 single-cell images from bone marrow smears from 945 patients with a variety of blood diseases. They also developed an artificial intelligence model and trained it using the image dataset.

They reported recently in the journal Blood that this method identified “a wide range of diagnostically relevant cell species with high precision and recall".

“This study is a step toward automated evaluation of bone marrow cell morphology using state-of-the-art image-classification algorithms,” they write.

Lead author Dr Christian Matek commented: “On top of our database, we have developed a neural network that outperforms previous machine learning algorithms for cell classification in terms of accuracy, but also in terms of generalisability.

“The analysis of bone marrow cells has not yet been performed with such advanced neural networks, which is also due to the fact that high-quality, public datasets have not been available until now.”

The team have made their database and the model freely available for further research and education.


Matek C, Krappe S, Münzenmayer C, Haferlach T, and Marr C. (2021) “Highly accurate differentiation of bone marrow cell morphologies using deep neural networks on a large image data set.” Blood 18 November 2021; doi: 10.1182/blood.2020010568

Link: https://ashpublications.org/blood/article/138/20/1917/477932/Highly-accurate-differentiation-of-bone-marrow

Disclaimer: The news stories shared on this site are used as a way to inform our members and followers of updates and relevant information happening in Haematology. The BSH does not endorse the content of news items from external sources, and is not in a position to verify the findings, accuracy or the source of any studies mentioned. Any medical or drugs information is provided as an information resource only, and is not to be relied on for any diagnostic or treatment purposes.

News service provided by Englemed News http://www.englemed.co.uk/