The Prognostic Value of Early Measurable Residual Disease Assessment in Patients with Acute Myeloid Leukemia Treated with Intensive Chemotherapy - Preliminary Results of Polish Adult Leukemia Group PALG-AML1/2016 Study

Background: Measurable residual disease (MRD) is an important biomarker in acute myeloid leukemia (AML). Among patients who achieve remission after standard chemotherapy, detection of MRD (MRD+) after two cycles of intensive chemotherapy, at the end of consolidation and before allogeneic stem cell transplantation (alloHSCT) is a strong prognostic factor for relapse and shorter overall survival (OS) (Short NJ, JAMA Oncol. 2020). The optimal time-points to asses MRD, and MRD cut-offs as well as whether eliminating of MRD due to further chemotherapy improves an outcome still remain open questions. PALG-AML1/2016 study aims to compare the safety and efficacy of two commonly used induction and salvage regimens in AML (NCT03257241). This is also the first international randomized trial in AML induction to prospectively evaluate the impact of MRD on overall survival, using multi-modality testing (flow-cytometry, FC; next-generation sequencing, NGS) of serial samples.

Study Design: In this study, 582 adult patients with newly-diagnosed AML will be randomized to DA-90 or DAC induction. Patients with complete remission (CR) or CR with incomplete hematologic recovery (CRi) receive further post-remission therapy adjusted to predefined risk group. The serial samples for multimodal MRD evaluation are collected at D14 (MRD-1), at time of CR/CRi after single or two induction cycles (MRD-2), and after each consolidation cycle (MRD-3, -4, -5).

Material and Methods. The aim of this preliminary analysis was to evaluate the prognostic value of bone marrow assessment at D14 both cytological and by FC (MRD-1) as well as the kinetics of MRD during post-remission treatment. Multivariable logistic regression models were developed with significant variables from univariate analyses, including MRD and blast count. The predictive power of the MRD level and blasts percentage at D14 was evaluated by receiver operating characteristics (ROC) and area under the curve (AUC) analysis to determine the ability of the biomarkers to accurately predict response to the induction treatment.

Results: The study group consisted of 284 patients (mean age: 47.1 ± 10.9) recruited until the end of May 2021. MRD-1 by FC was reported in 225 patients. CR and CRi either after single or double induction were achieved in 175 (61.6%) and 48 (16.9%) patients, respectively leading to an overall CR/CRi rate 78.5%. In univariate analyses, the factors significantly associated with achieving CR/CRi were: ELN high-risk group (OR 0.12, 95% CI: 0.05-0.33, p<0.0001), percentage of blasts at D14 (OR 0.95, 95% CI: 0.93-0.97, p<0.0001) and MRD-1 level (OR 0.94, 95% CI: 0.92-0.96, p<0.0001). A ROC analysis showed that AUC for the blasts count (0.71, 95%CI: 0.61-0.80) was significantly lower (p=0.0028) than AUC for MRD-1 levels 0.81 (95%CI: 0.74-0.89). Furthermore, two multivariate models, including ELN high-risk group and blasts or MRD biomarker, were evaluated. The model with blast percentage had an AUC of 0.83 (95% CI: 0.76-0.91). The model with MRD-1 level yielded higher AUC (p=0.0039) of 0.90 (95% CI: 0.84-0.95) and had a lower Akaike information criterion (116.9 vs. 127.1), indicating a better-fit model. In the analyzed group MRD-2 <0.1% did not have a predictive value for OS.

MRD-3 at the end of consolidation 1 was reported in 182 patients: MRD+ in 99 (54.4%) and MRD- in 83 (45.6%) patients. Results of MRD-4 at the end of consolidation 2 were available in 114 patients with 49 (43.0%) being MRD+ and 65 (57.0%) MRD-. Fifty-eight patients (35 MRD-2+ and 23 MRD-2-) were evaluated for MRD kinetics from the end of induction until the end of consolidation 3. Conversion rate from MRD- to MRD+ was 13% (n=3) and from MRD+ to MRD- 34%(n=12), respectively (p=0.038).

Conclusions: Our preliminary results suggest that flow cytometric bone marrow assessment on D14 has a predictive value for CR/CRi achievement and a potential prognostic utility for overall survival. The kinetics of MRD during consolidation requires further studies to better evaluate its prognostic significance.

References

1. Short NJ, JAMA Oncol. 2020; Zhou S, Fu C, et al. Association of Measurable Residual Disease with Survival Outcomes in Patients with Acute Myeloid Leukemia: A Systematic Review and Meta-analysis. 6(12):1890-1899.