Association Between Measurable Residual Levels and Infection Mortality in Acute Myeloid Leukemia

Abstract Background Acute myeloid leukemia (AML) outcomes have improved over the years. Better AML disease risk classification, achieved by measuring minimal residual disease (MRD) levels, has contributed to improved treatment and survival. Even though the role of MRD in survival is still preliminary, results are promising. However, infections are still a problem to circumvent. In our previous protocol LMA-GATLA 2007 we found a 30% incidence of sepsis related death (SRD), and they were more frequent during or after first induction. The overall survival of that study was 51.4 % (0.38). In this study, we aimed to analyze the mortality rates due to infection; and determine if there is an association of the MRD levels and infection mortality rates; also, we will analyze overall outcomes of AML. Methods Patients were accrued form the National Registry GATLA LMA 2021 protocol. Disease risks were classified based on cytogenetics and molecular abnormalities as standard risk (SR), high risk (HR), and very high risk (VHR). MRD determinations were performed using flow cytometry after 1st induction with mitoxantrone, anthraciclines and etoposide (MAE) protocol, at day 22 and 29 and after 2nd induction with high dose citarabine and mitoxantrone (HAM). We used descriptive statistics (analysis by frequency), compared categorical variables using chi-square or χ2 test and Fisher test, and time to event calculation techniques. Results 95 patients have been accrued, 51.6 % were males. The median age was 9 years of age (IQR 3-13), and 31% were SR, 22% HR, and 47% VHR. MRD levels < 0.1% were seen in 55.5% post first induction, 92% post intensification and 85% post consolidation. Estimated probability of overall survival was 65% (ES 0.6) at 45.7 (2.7) months, (CI 95% 40.3-51,1). 23 pts (24%) died; 11 (48 %) for progression, 7 (30%) for sepsis, for non-sepsis toxicity 3 (13%), other causes 2 pts (9%). Hazard ratio of death with MRD > 0.1% was 2.35 (p=0.06, 95% CI 0.043-1.756). The incidence of SRD was 7.4%, 4 patients had sepsis with MRD < 0.1% (57 %), after HAM /consolidation phase; and 3 patients with > 0.1% (43%) when in relapse (p= 0.028). The SRD episodes were independent from MRD levels, however patients on remission showed SRD as late events, because no patient developed SRD during or after first induction. Conclusion Overall survival results have improved with current protocol. MRD seems a valuable tool to estimate response, levels >0.1 % are associated with increased risk of death. However, better infection prevention and care are required to improve further the outcomes, especially those associated with higher MRD levels.