DOI: https://doi.org/10.36347/sjmcr.2026.v14i06.027

Pages: 1434-1440

Hematology analyzers of modern generation produce multidimensional and multiparametric datasets that cannot be effectively interpreted by traditional methods. AI and ML provide advanced computational tools that can help increase diagnostic accuracy and optimize the workflow in laboratories. To critically analyze the use of artificial intelligence in modern clinical hematology laboratories with a focus on automated hematology analyzers, digital morphology, prediction, and hemostasis integration. A structured narrative literature analysis of articles published between 2015 and 2026 was performed. The PubMed and Web of Science databases were used to identify scientific publications studying the application of machine learning and deep learning approaches to hematology. Various AI-based diagnostic models showed good discriminatory ability when classifying leukocytes, detecting blasts, stratifying different types of anemia, and diagnosing malaria. The area under curve of the models frequently exceeded 0.90 in validation datasets. The application of convolutional neural network helped increase the accuracy of digital morphology compared to conventional microscopy. Machine learning algorithms using CBC data allowed predicting sepsis and estimating the risk of malignant diseases. However, the vast majority of the studies used retrospective datasets with little external validation. Artificial intelligence helps optimize and make hematology analysis more accurate. Nevertheless, further research in the form of multicentric prospective validation and regulation is needed before applying the technology in practice.