International Research Journal of Oncology

While genetic mutations are well-known triggers for leukaemia, inflammation plays a big part by messing with cell growth and immune responses, yet its complex patterns are hard to track with traditional methods. This review looks at how artificial intelligence (AI) and machine learning (ML) uncover new ways to study inflammation-driven leukemogenesis, helping predict shifts to malignant tumours and improve diagnosis, treatment, and prevention. A thorough check of recent studies from PubMed, Scopus, and other biomedical databases was done, focusing on AI uses in haematological malignancies like acute myeloid leukaemia (AML) and myelodysplastic syndromes (MDS). Results show that ML tools, such as convolutional neural networks and random forests, spot inflammatory markers like elevated IL-6 and NF-κB pathways with high accuracy, in predicting MDS progression to AML, as seen in one study on cytokine profiling. AI in single-cell analysis has revealed NF-κB upregulation in leukemic clones, linking inflammation to clonal evolution. In niche modelling, AI simulates how cytokines like TGFβ1 inhibit healthy cells by 50% while protecting leukemic ones. These tools also enable early detection of clonal haematopoiesis with 99.5% accuracy in blood smear analysis and predict progression to acute leukaemia at 90%. AI combines multi-omics data to boost prediction of tumor shifts by spotting inflammation-tumor interactions. Still, issues like data differences across regions and hard-to-understand models remain. Current work explores ways to make AI more explainable and adaptable to low-resource areas. Looking ahead, single-cell studies and AI in rare blood cancers could make precision medicine more accessible. Teamwork between researchers, doctors, and tech experts is needed to turn these findings into better care for leukaemia patients.