Workflow for Automated Digital Peripheral Blood and Bone Marrow Cytomorphology, With Consideration of Critical Cellular Misclassifications.

Authors

David Starostka, Richard Dolezilek, Hans Michael Kvasnicka, Barbora Sotkovska, Petra Miczkova

Published in

Journal of visualized experiments : JoVE. Issue 233. Jul 28, 2026. Epub Jul 28, 2026.

Abstract

Growing evidence supports the ability of AI-powered automated digital cytomorphology (ADM) of peripheral blood (PB) and bone marrow (BM) to assist expert decision-making. The ADM workflow for PB and BM smears is described. The core components of ADM technology include precision mechanics, advanced imaging systems, artificial intelligence (AI), and a user-friendly interface. The process begins with standardized smear preparation and staining using protocols adapted for automation and digitization. An automated microscope equipped with an immersion objective and digital camera captures high-resolution images of the specimen at multiple magnifications. AI-powered software then locates, records, and preliminarily classifies nucleated cells, including megakaryocytes when applicable. This is followed by mandatory expert review and, as needed, reclassification. Digital images and reports are subsequently stored on a high-capacity local server. Key benefits of ADM include increased speed, efficient data management, educational support, remote access for evaluation and consultation, digital archiving, and multimodal data integration. The most significant limitation of ADM is the risk of critical cellular misclassification, which can have a major clinical impact. Irrelevant misclassifications are tolerable and diagnostically neutral, whereas relevant misclassifications are unacceptable and may lead to serious diagnostic and clinical consequences. Critical misclassification of neoplastic lymphocytes, small lymphoblasts/myeloblasts, granular monoblasts, dysplastic promonocytes and monocytes, plasmablasts, and immature plasma cells-all of which may present with atypical morphology-carries a risk of misdiagnosis. Further advancements are required to achieve reliable diagnostic performance. Extensive, high-quality training data, combined with regular validation, are essential to ensure that AI functions as an effective supporting tool for diagnostic judgment.

PMID:
42612139
Bibliographic data and abstract were imported from PubMed on 19 Aug 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.