Digital and computational morphology in hematology: current platforms, clinical evidence, and future requirements.

Abstract

INTRODUCTION: Morphologic examination of peripheral blood and bone marrow remains central to the diagnosis and classification of hematologic disorders. Conventional optical microscopy, however, is labor-intensive, dependent on operator expertise, and affected by interobserver variability. Digital morphology has developed from automated image acquisition and cell pre-classification into a broader field that includes whole-slide imaging, remote review, quantitative morphometry, and artificial intelligence-based analysis. CONTENT: This review examines current applications of digital morphology in peripheral blood, bone marrow aspirates, malaria detection, and body-fluid analysis. Commercial platforms are evaluated with particular attention to the distinction between raw automated pre-classification, expert digital post-classification, and comparison with independent optical microscopy. Digital systems generally perform well for common mature leukocyte populations but remain less reliable for rare or diagnostically critical cells, including blasts, abnormal lymphoid cells, plasma cells, and intermediate maturation stages. Research systems increasingly extend analysis from individual-cell classification to whole-slide, specimen-level, and patient-level assessment. SUMMARY: Digital morphology can improve standardization, image traceability, remote consultation, education, proficiency testing, quality assurance, and selected aspects of laboratory workflow. Its clinical value depends on appropriate validation, transparent reporting of reference methods, recognition of algorithm-specific failure modes, and clearly defined criteria for expert review and conventional microscopy. Human expertise remains essential not only for validating results but also for adapting cell taxonomies and interpretive rules to evolving classifications of hematologic diseases. OUTLOOK: Future progress will require representative multicenter datasets, harmonized morphologic terminology, external validation, interoperability with laboratory information systems, and continuous monitoring after software or hardware updates. Integration of morphology with quantitative hematology, flow cytometry, cytogenetics, genomics, and clinical data may support more comprehensive computational diagnosis. Digital platforms may also broaden access to specialist expertise, training, and quality programs in resource-limited institutions and regions, provided that infrastructure, governance, and professional competency are adequately supported.