Early Peripheral Blast Cell Clearance Predicts Response to Induction Chemotherapy in Acute Myeloid Leukemia

Early peripheral blast cell clearance (PBC) assessed by leukocyte differential is well established as a predictor of leukemia induction outcome, but measuring PBC by flow cytometry also can predict complete remission (CR) and survival after induction treatment for acute myeloid leukemia (AML), according to findings from a study published in Blood Advances. Investigators also reported that the early PBC rate correlated with responses to induction intensity, suggesting that this assay may be useful for the personalization of therapy in patients with AML.

The minimally invasive PBC approach allows for functional quantification of AML burden reduction while identifying early high-risk patients who might benefit from more intense chemotherapy regimens, the authors wrote. “For decades, induction therapy … response has been assessed after the fact, at hemopoietic recovery,” corresponding author Francesco Mannelli, MD, from the University of Florence in Italy, told ASH Clinical News. “Our data support the personalization of treatment dependent on individual chemosensitivity provided by the rate of PBC, which could … identify patients who actually need intensification.”

The present analysis was embedded in the Northern Italy Leukemia Group (NILG) AML 02/06 phase III trial, which randomized adults with treatment-naïve AML to one of two treatment arms:

  • standard CR induction with idarubicin 12 mg/m2 on days 1-3, etoposide 100 mg/m2 on days 1-5, and cytarabine 100 mg/mm2 twice daily on days 1-7 (ICE; n=82)
  • high-dose sequential therapy with cytarabine 2,000 mg/mm2 twice daily on days 1-2 and 8-9 and idarubicin 18 mg/mm2 on days 3 and 10 (HDS; n=69)

To determine the prognostic value of PBC, researchers compared PBC responses and related therapeutic outcomes in the two randomization cohorts and across subsets of patients with different AML risk. Eligible patients had ≥100 circulating cells/mL with a leukemia-associated immunophenotype (LAIP+) at diagnosis. Individual LAIP was quantified via flow cytometry prior to treatment, as well as after 3 days of treatment.

Overall, the median absolute LAIP count decreased from 4.198/μL (range 120-148.86) at baseline to 41.8/μL (range = 0.07-44.80) by 4days following the start of induction chemotherapy. More than two-thirds of patients experienced a CR after 1 course of induction therapy (71.5%), while 22.5% had non-responsive disease, and 6% died early during induction.

The investigators reported that patients with non-responsive disease had higher counts of residual LAIP+ cells on day 4, compared with patients who experienced a CR. Non-responders also had a significantly lower median PBC (1.0 log vs. 2.3 log; p<0.001).

Using receiver operating curve methodology, the authors then identified 1.5 log as the PBC cutoff value with the highest predictive power for CR. Based on this PBC value, 96 patients were considered in the high PBC range (>1.5 log; PBChigh) and 55 were considered in the low PBC range (≤1.5 log; PBClow).

The two distinct high and low PBC ranges were then analyzed as potential predictive markers for therapeutic response. In this analysis, 8788% of patients in the PBChigh group achieved CR after a single chemotherapy course, compared with only 44% of patients in the PBClow group (p<0.0001).

“PBC analysis can provide a very early prediction of outcome, correlates with treatment intensity and disease subset, and may support studies of customized AML therapy.”

—Francesco Mannelli, MD

Rates of CR and PBChigh status were significantly higher in patients randomized to the high-dose induction chemotherapy arm. PBChigh patients also experienced better long-term outcomes than patients in the PBClow category:

  • overall survival: 63% vs. 25% (p<0.0001)
  • disease-free survival: 56% vs. 29% (p=0.0002)
  • event-free survival: 50% vs. 18% (p<0.0001)
  • cumulative incidence of relapse: 37% vs. 62% (p=0.0024)

Multivariate risk models also revealed that PBChigh rates correlated strongly with genetic/cytogenetic risk but retained independent predictive power for long-term outcomes.

Taken together, these results suggest that “PBC analysis can provide a very early prediction of outcome, correlates with treatment intensity and disease subset, and may support studies of customized AML therapy,” the authors concluded.

However, the implications of this study are limited by the lack of individual LAIP re-assessment in bone marrow samples at the time of response examination and MRD analysis and by the lack of clinical action taken based on LAIP results. Another limitation, according to Dr. Mannelli, concerns the correlation of PBC with MRD, which is widely recognized as an essential post-treatment parameter for measuring response in AML. “We could not assess this issue because MRD data were not included in the patient cohort of the NILG trial,” he said. “We are working now with data from a different series, and we hope to publish these findings soon.”

Dr. Manelli added that researchers also are planning to launch a multicenter trial that will evaluate outcomes of induction treatment modified on the basis of PBC results; if a patient has PBC findings predictive of poor response to standard induction, they will immediately be switched to an intensified induction regimen consisting of high-dose cytarabine.

The authors report no relevant conflicts of interest.


Gianfaldoni G, Mannelli F, Intermesoli T, et al. Early peripheral clearance of leukemia-associated immunophenotypes in AML: centralized analysis of a randomized trial. Blood Adv. 2020;4:301-311.