In childhood acute lymphoblastic leukaemia (ALL) overall event free survival exceeds 75% for most patients. Assessment of early responses to induction therapy by measurement of minimal residual disease (MRD) either by PCR or flow cytometry is a powerful independent prognostic factor. However, there is little data on whether the initial rate of bone marrow (BM) disease clearance is a contributory factor and if this varies for the cytogenetic subgroups, t(12;21) or hyperdiploid. Using 4-colour flow cytometry we examined the phenotypic characteristics at presentation of 29 cases (14 male and 15 female; age range 18–197 months; median 36 months) of B-lineage ALL. BM leukaemic cells were identified on the basis of light scatter characteristics, CD19 and CD45 expression. Geometric mean fluorescent intensities (MFI) were then assessed for a standard series of B, T and myeloid antigens. Novel phenotypic correlates were found (table 1) that allowed early prediction of the presence of t(12;21) or hyperdipoidy prior to confirmation by PCR, cytogenetics or FISH.

Antigent(12;21) (n=7) mean MFI (range)Hyperdiploid (n=15) mean MFI (range)P value (Mann-Whitney U)
CD10 3636 (1597-5562) 1516 (777-3788) 0.001 
CD20 9.0 (2.9-19.3) 75.9 (5.2-493.7) 0.062 
CD66c 6.0 (3.5-8.1) 97.5 (10.1-310.8) <0.001 
Antigent(12;21) (n=7) mean MFI (range)Hyperdiploid (n=15) mean MFI (range)P value (Mann-Whitney U)
CD10 3636 (1597-5562) 1516 (777-3788) 0.001 
CD20 9.0 (2.9-19.3) 75.9 (5.2-493.7) 0.062 
CD66c 6.0 (3.5-8.1) 97.5 (10.1-310.8) <0.001 

Using the unique phenotypic features of each subgroup, 4-colour combinations of antibodies were designed that would allow MRD monitoring of BM samples taken on days 8, 15 (optional) and day 28 (end) of induction therapy. These combinations would also be applied to BM samples throughout maintenance therapy. The t(12;21) cases showed rapid disease clearance with 2/7 patients becoming MRD negative by day 8 of induction and only one case MRD positive at day 28. In contrast, hyperdiploid cases showed a significantly slower initial clearance of leukaemic cells (P= 0.013 Mann Whitney U), with all cases MRD positive at day 8 and 4/15 MRD positive at day 28. Interestingly, MRD negativity at day 8 appeared to be a specific feature of t(12;21) disease, as in addition to the hyperdiploid cases, 7 other patients with non t(12;21) cytogenetic abnormalities were also MRD positive at day 8. Phenotypic shifts of residual leukaemic cells as measured by changes in MFI, were demonstrated throughout therapy by the hyperdiploid cases. They showed significant decreases in expression of CD34 and CD66c, whilst CD20 expression increased (All P<0.05). This phenomenon was not seen in the t(12;21) subgroup which retained a stable phenotype, though CD10 expression did fall, but not significantly (P = 0.07). As phenotypic changes are a consistent feature of hyperdiploid cases, they may indicate a preferential eradication of leukaemic subclones. Our findings have significant implications for flow cytometry based MRD studies of specific subtypes of B-ALL. The results raise doubts about the reliability of some antigens that are aberrantly expressed at presentation and therefore good candidates for MRD markers, as their expression changes significantly during therapy. Prospective monitoring of this cohort of patients is ongoing in order to assess the relationship between disease clearance and outcome at end of therapy.

Author notes

Corresponding author

Sign in via your Institution