Margherita Bodini

Acute Myeloid Leukemia (AML) is a cancer of the myeloid lineage of blood cells. Despite the high rates of complete remissions achieved after treatment (60-80% in young adults), the number of patients that will result cured after induction and consolidation therapy is low (~12%). The molecular basis of relapsing disease is still unclear and the few identified predictive factors has small predictive power. Currently, chemotherapy induction treatment consists in the administration of mainly three drugs (fludarabine, cytarabine, and idarubicin). In this thesis, endowed of the NGS technologies advancement, we decided to delineate the possible process of relapse formation in order to be able in the future to predict which patients are more susceptible to relapse. Our experimental plan includes the whole exome analysis of 30 pairs of primary/relapsed AML samples using NGS to identify relapse-specific mutations, the bioinformatics analysis of the clonal evolution of the disease and the identification of pathways that correlate with the relapsing disease. Our analysis shows that the genomic landscapes of primary and relapse AMLs are similar and in the majority of the patients (76%) some relapse clones were already present in the primary tumour and reappeared after chemotherapy at similar or augmented cellular frequencies. In 4 out of 29 patients (14%) we were able to identify driver mutations in the blood sample of the complete remission at low frequency; we hypothesize that more sophisticated diagnostic tools, based on NGS analysis, would help in driving the treatment to obtain better outcomes for patients.