Improving prognostication and treatment choices for patients with AML

Abstract: The treatment landscape of the aggressive haematological malignancy acute myeloid leukaemia (AML) has expanded but the prognosis is still unsatisfactory poor. Here, we aimed at improving prognostication and treatment choices in AML by addressing current clinical obstacles to successful AML treatment.Acute promyelocytic leukaemia (APL) is an AML subset characterised by a high rate of early death (ED). In Paper I, we developed a novel risk score for ED in APL. We identified three risk groups for ED based on regression analyses on first a training cohort from the population-based Swedish AML Registry (n=301) and later an external validation cohort from a hospital-based registry (n=129). The presented risk score included age, platelets and white blood cell (WBC) count. Importantly, already sub-normal to normal WBC counts conferred higher risks of ED.Molecular studies of elderly AML patients are sparse. In Paper II, we focused on patients ≥65 years to investigate the prognostic effect of molecular markers and to propose an algorithm for response to intensive chemotherapy (IC) in this patient group. We combined clinical data with targeted DNA- and RNA-sequencing of 182 patients. Notably, we identified and externally validated three risk categories for complete remission achievement after IC based on mutational status of TP53 and gene expression levels of ZBTB7A and EEPD1.Hypomethylating agents (HMAs) constitute a backbone for AML patients ineligible for IC. There are limited studies on their effectiveness in the real-world setting. In Paper III, we compared the utility of HMAs against IC and palliative care in all AML patients ≥60 years in Sweden (n=3135) during 2008-2018. Propensity score matching in this population-based cohort showed that HMAs are as effective as IC upfront when patient characteristics were balanced. Additionally, predictive factors for overall survival in HMA treated patients were different to IC treated patients.The HMA azacitidine combined with venetoclax is the current frontline option to AML patients unfit for IC. Few studies have addressed how this synergism arises. In Paper IV, we characterised the epigenetic and transcriptomic effects of azacitidine-venetoclax in vitro and elucidated potential survival/resistance mechanisms in AML blasts including the serine synthesis pathway and NTRK signaling. Furthermore, we utilised obtained RNA-seq data and in silico predictions to propose add-ons to azacitidine-venetoclax to further strengthen the synergy.In summary, the research presented herein contributes to improved personalised medicine in AML via real-world data, risk stratification algorithms and insights into potential novel therapeutic approaches.