Key points are not available for this paper at this time.
BACKGROUND: Clinical researchers and laboratory specialists are striving to explore artificial intelligence (AI) to facilitate and optimize haematological diagnostics in response to the growing demand for more efficient and accurate diagnoses. SUMMARY: This review summarizes current approaches integrating AI into blood and bone marrow cytomorphology, flow cytometry (FC), genetics, and haemostasis. Efforts include automated cell differentiation in peripheral blood and bone marrow aspirates, algorithms for identifying causes of anaemia, tools for rapid diagnosis of acute leukaemia, and other haematological entities. AI in FC may reduce subjectivity and variability, while in genomics, machine learning is increasingly implemented for processing high-throughput sequencing data and may enable automated detection of karyotypes in the future. In haemostasis, AI allows for automation in quality control, the establishment of personalized reference ranges, and potentially automated result interpretation. AI has, however, limitations such as cross-platform compatibility and often lacks sufficient validation. Ethical concerns include risks of bias and regulations are lagging behind the rapid developments. KEY MESSAGES: AI shows promise for automating and improving many steps in haematological diagnostics, though final interpretation still needs expert haematologists.
Schnegg‐Kaufmann et al. (Fri,) studied this question.