Real Life Feasibility of a Combined Clinical and Biological Fitness Assessment in Patients with Acute Myeloid Leukemia

Introduction: Fitness assessment plays a crucial role in the process of treatment allocation for adult patients (pts) with acute myeloid leukemia (AML). Among the available scores, the SIE (Italian Society of Hematology), SIES (Italian Society of Experimental Hematology) and GITMO (Italian Group for Bone Marrow Transplantation) criteria [Ferrara F. et al, Leukemia 2013] were intended to describe novel definitions of unfitness to intensive chemotherapy (IC) or non-intensive chemotherapy (N-IC). Although being retrospectively validated, these criteria were not designed to include disease-related genetic/cytogenetic status, thus failing to predict the impact of the available treatment modalities on the different genetic/cytogenetic risk groups. Aim: We investigated whether an integrated clinical fitness (according to SIE/SIES/GITMO criteria) and biological (according to ELN2017 risk stratification) assessment could help re-defining risk-category assignment of our pts, and better discriminating their outcome both in terms of complete remission (CR) probability and Overall Survival (OS). Methods: This retrospective analysis includes 285 consecutive pts affected with non-promyelocytic AML, diagnosed at our institution between January 2013 and December 2021. SIE, SIES, GITMO operational criteria were retrospectively applied through medical files backtracking. Molecular risk was attributed according to ELN2017 risk stratification. Results: Median age was 68 years (range 21-93). According to SIE/SIES/GITMO criteria, 142 (49.8%) pts qualified as fit (Fi-P), 57 (20%) as unfit (UF-P) and 86 (30.2%) as frail (Fr-P). According to ELN2017 risk stratification, 46 (16.1%) pts were classified as favorable [ELN2017-FR], 107(37.5%) as intermediate [ELN2017-IR] and 69 (24.2%) as adverse risk [ELN2017-AR]. Overall, 63 (22.2%) pts were not classifiable due to the lack of cytogenetic and/or molecular data. One hundred-32 (46.6%) pts received IC, 65 (22.8%) N-IC, 88 (30.9%) best supportive therapy (BST), with an overall 93% concordance rate between clinical fitness status and treatment received (90.1% of Fi-P received IC, 91.2% of UF-P received N-IC and 97.7% of Fr-P received BST, respectively). Such a concordance improved over the time (from 87.8% in 2013 up to 95.2% in 2021). We then focused on pts receiving active therapy (UF-P submitted to N-IC and Fi-P to IC, respectively). In the UF-P group, 12-months CR incidence was similar between ELN2017-IR (37.5%) and ELN2017-AR (28.6%) pts, with no difference in terms of median OS between the 2 groups (9.1 months vs. 6.6 months for ELN2017-IR and ELN2017-AR pts, respectively) [Figure 1A]. Five ELN2017-FR unfit patients submitted to N-IC were excluded from these analyses because of low numbers. Among Fi-P, 12-months CR incidence was proportional to ELN2017 risk (87.5% vs. 69.7% vs 51.4% for ELN2017-FR, ELN2017-IR, and ELN2017-AR, respectively; p=0.0012), with an OS advantage for ELN2017-FR pts over ELN2017-IR and ELN2017-AR ones (median: not reached for ELN2017-FR vs 12.2 months for ELN2017-IR and ELN2017-AR, respectively; p=0.0051) [Figure 1B].Conclusions: We confirmed that a real-life therapeutic allocation relying on a comprehensive clinical (SIE/SIES/GITMO criteria) and biological (ELN2017) risk assessment is feasible. Based on their superimposable outcomes, pts unfit for N-IC may be candidate to an active, non-intensive approach, even if belonging to the ELN2017-AR category. For Fi-P belonging to the ELN2017-IR and ELN2017-AR categories, there is room for further improvement of clinical outcome by taking advantage of new strategies (targeted agents, biomarker driven approaches). Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal