Early peripheral blood blast clearance is a prognostic marker for early treatment response and long-term outcome in pediatric Acute Myeloid Leukemia

Abstract Background: Bone marrow (BM) aspirates during and after induction chemotherapy are established but cumbersome prognostic tools for assessing treatment response in pediatric acute myeloid leukemia (AML) (Creutzig 2014, Abrahamsson 2011, Kern 2003). Peripheral blood blast clearance (PBBC) assessment by morphology has shown prognostic value in adult AML and might offer a cost-effective, easily accessible alternative, especially in low-resource settings (Manabe 2008, Arellano 2012). Its utility in pediatric AML is not well-established. This study investigates early PBBC as a predictor of treatment response and outcomes in pediatric AML. Methods: Data on pediatric patients with de novo AML treated according to the NOPHO-DBH AML-2012 protocol were prospectively obtained. Participating countries included Belgium, Denmark, Israel, Latvia, The Netherlands, Norway and Sweden. The first 5 days of induction therapy consisted of etoposide, followed by 7 days of low-dose cytarabine and mitoxantrone or liposomal daunorubicin. Peripheral blood (PB) was assessed by morphology at the time of diagnosis, daily in the first week, and thrice weekly until D22 or until complete clearance of blasts occurred. BM at day 22 (D22) was assessed by morphology and flow cytometry. A receiver operating characteristic (ROC) analysis was performed to evaluate the predictive power of the number of days from starting chemotherapy to complete PBBC. The time to PBBC and rate of PBBC, defined as the daily percentage of peripheral blood blast count cleared from diagnosis to the last assessment, were correlated with BM response on D22, complete remission status (CR), refractory disease (RD), event-free survival (EFS), and overall survival (OS). Flow cytometry based measurable residual disease (flow-MRD) negativity was defined as the absence of detectable leukemic cells in the BM (0.01%). Results: Between March 2013 and February 2023, a total of 319 patients were included, 241 of whom were eligible for analysis, due to missing or incomplete data (n=78). Day 9 after initiation of chemotherapy was identified as the most discriminating cut-off point for predicting BM D22 response (predicted probability of 89.3%; P0.001). Patients who achieved PBBC ≤9 days were classified as early PBBC (n=159; 66%), while those with blast clearance 9 days were classified as delayed PBBC (n=82; 34%). For 217 patients (due to missing data in 24 cases), the rate of PBBC could be calculated, and a discriminative cut-off value of 11% was identified (predicted probability of 84%, P0.001). Patients were categorized into a high rate of PBBC (11%; n=132) and low rate of PBBC (≤11%; n=85). Early PBBC and a high PBBC rate were associated with favorable BM responses (5% leukemic cells) on D22 (86.7% vs. 68.4%, P0.001; and 88.1% vs. 67.1%, P=0.001, respectively). No poor responders with ≥15% leukemic cells in BM on D22 were observed in the early PBBC group. Early PBBC and high PBBC rate were associated with higher flow-MRD negative rates on D22 (53.7% vs. 37.5%, P=0.027 and 51.9% vs. 38.4%, P=0.028). There was no significant association with RD or EFS in either group. Five-year OS was significantly higher in the early PBBC and high PBBC rate group (84.2% vs. 72.6%, P=0.02; and 86.5% vs. 73.7%, P=0.021, respectively). Conclusion: Early PBBC is a promising, non-invasive prognostic marker of treatment response and outcome in pediatric AML. Its simplicity and cost-effectiveness make it particularly advantageous for low-resource settings.