The detection of minimal residual disease (MRD) is highly predictive of outcome in pediatric acute lymphoblastic leukemia (ALL) and has now been integrated into the majority of treatment regimens. Early studies demonstrated the strong prognostic value of post-induction MRD measurement, using both molecular and flow cytometric techniques, allowing treatment intensity to be tailored to response. More recent studies have examined the value of MRD measurement at earlier time points and using peripheral blood (PB) in place of bone marrow (BM) in an attempt to reduce invasive procedures.
Within the United Kingdom (UK), previous national trials have shown that post-induction MRD assessment, using molecular techniques, is highly predictive of relapse, allowing treatment stratification based on MRD levels. We have previously demonstrated the value of flow cytometric MRD assessment at early time points in response prediction (Motwani et al, Blood 2007a), however, as yet there has been no assessment of MRD in peripheral blood at early time points using flow cytometric methods in the context of a UK national trial.
To address this we assessed MRD by flow cytometry using Day 8 (D8) and/or D15 paired PB and BM samples taken from children being treated for ALL at our unit on the UK MRC UKALL2003 trial or ALLR3 relapse trial. Samples were analyzed by 4-color flow cytometry using a large panel of monoclonal antibodies to identify all leukemia-associated immunophenotypes (LAIP). For follow up samples a minimum of 2 markers were identified for each patient and sequential gating was used for analysis. 500 000 to 106cells were acquired for each antibody combination for analysis to be sufficiently sensitive (0.01%). The number of residual leukemic cells was calculated as the percentage of blasts present within total nucleated cells counted. Sensitivity experiments were performed using dilutions of leukemic blasts in normal bone marrow or peripheral blood and demonstrated a sensitivity of 0.01%.
Fifty paired BM and PB samples from 46 patients (41 newly diagnosed, 5 relapse) were included in the study (Age 2–22 years, mean 7.8 years). Of these, 22 were taken at D8 and 28 at D15. Forty-two samples were from patients with B-cell ALL, whilst 8 were from patients with T-cell ALL. The most frequent LAIPs used to identify blasts in B-ALL were CD58 (90% of cases), CD45 (75%), CD38 (86%), CD123 (77%). In T-ALL the most common LAIP combinations were CD34 (75%), CD5/CD56 (62.5%), CD5/CD2 (50%) and CD5/CD99 (50%).
In keeping with previous reports, B-ALL MRD levels in PB blood were approximately 1.3-log lower than BM MRD levels (D8 mean 2.27% vs. 27.8%; D15 levels 0.63% vs. 7.99%). T-ALL levels were approximately 0.5-log lower. Importantly, the level of MRD in BM and PB in B-ALL showed a strong correlation (r=0.77).
Whilst there were insufficient relapses to allow correlation with long-term outcome, PB MRD levels at D8 showed good prediction of post-induction molecular MRD, the current time point used for risk stratification in the UK trial. In patients with PB MRD >1% at D8, 100% (5/5) were in the high-risk group post-induction, significantly more than those with PB MRD <1% at D8, of which only 36.4% (4/11) were allocated to the high risk group post-induction (p=0.023). MRD levels in BM at D8 did not show similar predictive value. Importantly, this early prediction of risk could offer an early opportunity to intensify treatment in this high-risk group, potentially enhancing disease clearance during the induction phase.
This study further supports that early assessment of MRD using flow cytometric techniques enhances treatment stratification in pediatric ALL. Furthermore, it shows that PB sampling may provide a markedly less invasive means of MRD assessment in these patients. Validation in a larger cohort of patients is warrented.
No relevant conflicts of interest to declare.
This icon denotes a clinically relevant abstract
Asterisk with author names denotes non-ASH members.