Assessment of Peripheral Blood (PB) Blast Decrease Rate by Multiparameter Flow Cytometry: A New Prognostic Factor in Acute Myeloblastic Leukaemia (AML). A Goelams Study.

Abstract

In spite of the tremendous progresses of therapeutic approaches in the past decades, the prognosis of AML remains very poor. Developing new biological tools for disease monitoring is therefore a constant challenge, to adapt therapeutic protocols as early as possible. It has been suggested (Leuk Res,1992) that a morphologically assessed rate of decrease of medullary blasts during induction could be a prognostic indicator. Increasing sophistication of flow cytometry (FC) has considerably improved the sensitivity of this technique, allowing for a precise and reliable detection of low blast percentages (<3%) in bone marrow and even PB samples.We devised such a FC 4colours/single-tube protocol, combining CD14-FITC/CD11b-PE/CD45-ECD/CD16-PC5 for whole PB staining, analyzed with a strategy of sequential subtractive gating. A biparametric histogram combining SSC and CD45 was used to devise a bitmap containing blast cells and remaining lymphocytes (LY), monocytes (MO) and polymorphonuclears (PMN). From this gate, LY were excluded by their SSC/CD45 properties, MO by CD14/CD11b gating and PMN by CD16/CD11b gating. The resulting population of blast cells was appreciated as % of all leukocytes. This approach was applied, in 4 different centres, to 130 patients with AML (excluding M3) (66 males, 64 females, mean age 55 y.o.) enrolled in therapeutic trials of the GOELAMS group (LAM2001, LAMSA02) or treated according to these strategies from 02/2003 to 05/2006. PB was collected daily for up to 10 days from D1 of chemotherapy. PB blasts assessed by FC on D1 ranged from 97.8 to 3.1% (med. 41%). This D1 percentage was normalized to 100, and subsequent values were adjusted accordingly. Each slope of blast decrease was calculated between D1 and the first day when at least 90% of PB blasts had disappeared (D90%). The same strategy was applied for decrease slopes of WBC and blast counts (BC), over the same period. The slope of blast cell percentages decrease provided the most significant information about the patients' outcome. Analysis of CR (n=111, 85%) and failure (n=19, 15%) showed significantly different mean slopes. Failed patients had the lowest slopes (mean: −7.56±1.07). CR patients who relapsed (n=47) had a mean slope of −18.3±1.22 while those who did not (n=64) displayed the steepest mean slope (−27.6±1.09, ANOVA p<0.001). DFS were compared by Log Rank test using a slope threshold of 20 (whole group's mean value), thus significantly (p<0.0001) discriminating two groups. Use of both BC (p = 0.09) or WBC (p = 0.97) slopes failed to show clinical correlations. According to D90%, 3 groups were constituted as follows: D90% <D5, D90% = D5 and D90%>D5, between which DFS differed significantly (p<0.0001). In conclusion, this report indicates that a simple, universal FC single-tube protocol provides a new prognostic factor in AML, based on the accurate determination of PB blasts' decrease during induction chemotherapy. It could be directly applied to new therapeutic protocols. [Display omitted]