Flow Cytometric Measurement of the Initial In Vivo Response to Prednisone Provides New Prognostic Information in Childhood Acute Lymphoblastic Leukemia (ALL)

Background The initial in vivo response on corticosteroid therapy (<1000 blasts/mmc at day 8 after exposure to Prednisone) is considered a prognostic predictor to qualify patients at higher risk of relapse (Riehm H et al, Klin. Padiat. 199, 1986). This parameter has been prospectively validated in several BFM-oriented trials providing evidences to be a robust predictor of disease recurrence even in specific subgroups of ALL (Schrappe M et al, Klin. Padiat. 2013). Evaluation of day8 prednisone response is assessed by morphological evaluation of peripheral blood (PB) smears, however this approach may have several limitations such as poor quality of the smears, operator- dependent variability and accuracy, poor reproducibility between centers. Methods We studied a total of 543 (of 648 eligible) patients enrolled in the AIEOP-BFM ALL 2000 study between January 2001 and December 2011 in the centers of Monza and Padova (275 and 268 patients respectively). PB collected at day+8 of induction phase was analyzed by multiparametric flow cytometry (FCM) to assess blast immunophenotyping by standard methods developed in the context of the I-BFM Flow network (Dworzak MN et al, Cytometry part B, 2008). Results Five hundred-seven patients (93.4%) had <1000 blasts/mmc as assessed by FCM, whereas 36 (6.6%) had ≥1000 blasts/mmc. Five-year Event Free Survival (5 yrs EFS) was 84.8% and 76.4% respectively (p-value =0.01). We further investigated the absolute count by sub-grouping the patients in three levels: i) 0<100 blasts/mmc (n 416), ii) ≥100<1000 blasts/mmc (n 91), and iii) ≥ 1000 blasts/mmc (n 36). 5 yrs EFS was 86.8%, 75.6% and 76.4%, respectively (p-value group a vs group b was 0.001), see Figure. Five years Cumulative Incidence of Relapse (5yrs CIR) was 11.4%, 21.1% and 17.9%, respectively (p-value group a vs group b was 0.003). When we excluded the high-risk (HR) patients from the analysis, those in the group with ≥100<1000 blasts/mmc (n71) maintained a significant worse outcome as compared with the group of patients with 0<100 blasts/mmc. Conclusions We identified a group of 71 patients (13% of total) with non-HR characteristics and with ≥100<1000 blasts/mmc at day8, showing significant worse outcome as compared to patients with <100 blasts/mmc, that could benefit of a more intensive treatment. These Analyses will be validated on a larger cohort of patients. FCM is more accurate and reproducible as compared to morphology and should be considered as an alternative method for the evaluation the initial in vivo response to corticosteroid therapy and possibly for a further refinement of treatment stratification of children with ALL.