Only young age and genetics retain prognostic value for 2-year relapse-free survival in AML

Abstract Background: Intensive chemotherapy (IC) regimen is the standard of care and the only curative treatment for acute myeloid leukemia (AML) since 1970's (Döhner and al, NEJM 2015) that allows a five-year survival of less than 50% (Kantarjian and al, Cancer 2021). While most patients are eligible for allogeneic transplantation (ASCT), some do not receive it due to donor unavailability, comorbidities, or other limiting factors. Factors associated with prolonged 2-year progression-free survival (PFS) in AML patients—especially those with an indication for ASCT in first complete remission (CR1)—remain poorly understood (Bennett and al, Cancer 1997; Bloomfield and al, Cancer 1997; Heering and al, Eur J Haematol 2022). We focused on AML patients alive at 2 years, not transplanted in CR1, with or without an initial transplant indication. Methods: Between 2008 and 2018, among more than two thousand patients included in the AML Hauts-de-France observatory (n° 2214454v0 de la Commission Nationale de l'Informatique et des Libertés), 237 patients were homogenously treated by IC in Amiens and Lille University Hospital Centers. We included in our study 147 patients who were alive beyond two years, without ASCT. The correlation between clinical and biological characteristics (including cytogenetic and molecular landscape) and PFS were specifically analysed in univariate and multivariate following a step-wise backward algorithm. Secondary outcomes were predictive factors of CR and overall survival (OS). Results: Median age was 53 years old IQR [45-63] and sex-ratio was 1.13. Patients had mainly de novo AML (91.16%) with recurrent genetic abnormalities (80.95%) and 2022 ELN risk group was favourable in 72.41%, intermediate in 11.72% and adverse in 15.86%. CR was obtained after IC (one or 2 cycles) for all patients. ASCT was indicated for 31.03% but was not performed, mainly due to lack of donor. Among our cohort, 100 patients were relapse-free at 2 years (named relapse-free), and 47 relapsed within that timeframe (named relapse). Relapse patients had less transcription factor gene fusions or mutations than the relapse-free patients (42.55% versus 62%, p=0.042) and more adverse 2022 ELN risk AML (17.02% versus 15.31%, p=0.037). With a median follow-up of 86.77 months [62.6-124.5], median OS was not reached. In univariate analysis, the predictive factors of OS > 2 years for the 100 patients without relapse were: age < 65 years (HR 4.00 [1.85-8.64], p<0.001), absence of comorbidities (3.68 [1.65-8.20], p=0.001), abnormal cytogenetic (0.26 [0.10-0.68], p=0.006) and the absence of ASCT indication (3.31 [1.53 - 7.16], p= 0.002). In multivariate analysis, only age as continuous variable (1.08 [1.04-1.13], p<0.001) was predictive of OS > 2 years. Global median PFS was 66.6 months (95% IC [44.22-145.54]) (13.08 months [11.2-14.67] in the relapse group versus not reached in the relapse-free group). Relapse was observed for 47 patients (100%) in the relapse group and for 29 patients (29%) in the relapse-free group. In multivariate analysis, only age as continuous variable (1.10 [1.05-1.15], p < 0.001), abnormal karyotype (0.21 [0.08-0.58], p = 0.003) and adverse or intermediate 2022 ELN risk (7.96 [3.27-19.40] p < 0.001) were predictive of PFS > 2 years. Conclusion: In our homogenously treated population without ASCT, 2-year PFS was only influenced by younger age, cytogenetic and ELN risks.