| "Revolutionising Flow Cytometry: AI-Powered Diagnostics for Haematologic Malignancies." |
| Paper ID : 1203-IGA |
| Authors |
|
Adela Perolla * University of Medicine , Service of Hematology , UHC " Mother Teresa"
Tirana , Albania |
| Abstract |
| Background Flow cytometry has transformed the diagnosis and monitoring of haematologic malignancies, providing a high-throughput method for analysing cellular DNA, immunophenotyping, and minimal residual disease (MRD) tracking. Studies have demonstrated its superior sensitivity over conventional cytology in detecting small malignant populations (1,2). The integration of artificial intelligence (AI) and machine learning (ML) represents a paradigm shift, bringing automation, precision, and predictive insights to haematologic diagnostics. Objective This study evaluates the clinical impact of AI in flow cytometry, focusing on haematological malignancy detection, CNS malignancy assessment, MRD monitoring, and treatment response prediction. Methodology A comprehensive literature review was conducted using PubMed, Nature, Frontiers in Medicine, and Clinical Chemistry, analysing peer-reviewed papers (1987–2024) on AI-driven flow cytometry, leukaemia classification models, MRD tracking, and CNS haematologic malignancy diagnostics. The review assessed automation, accuracy, efficiency,of AI’s integration with emerging cytometry technologies. Data were analysed based on methodological rigour, clinical impact, and AI model validation challenges. Results & Discussion AI-driven flow cytometry demonstrated higher diagnostic accuracy, surpassing manual gating in AL, CLL, MM, and MDS (5). Improved MRD monitoring, enabling earlier relapse prediction and treatment response tracking (6). Enhanced CNS malignancy detection, outperforming cytomorphology in cerebrospinal fluid (CSF) analysis (7). Advanced real-time data interpretation, integrating mass spectrometry and multidimensional analysis (8). Despite its benefits, challenges persist, including data heterogeneity, lack of standardisation, and variability in antibody panels. AI models function as black-box systems, limiting clinician trust and interpretability. Regulatory validation and high computational requirements remain key barriers to clinical adoption, especially in low-resource settings. Conclusion AI-powered flow cytometry is revolutionising haematologic diagnostics, enabling faster, highly accurate, and reproducible detection of AL, CLL, MM, and MDS. However, ensuring standardisation, regulatory compliance, and clinical validation is critical for its successful integration. Future research should focus on model transparency, equitable access, and AI-driven standardisation of flow cytometry protocols. |
| Keywords |
| AI in hematology, flow cytometry, acute leukemia, CLL, multiple myeloma, MDS, minimal residual disease, machine learning, cerebrospinal fluid diagnostics |
| Status: Accepted |
