Standardized 4 Color Flow Cytometric Minimal Residual Disease Detection Failed To Overcome Regeneration Problems but Identifies Early Blast Clearence Predictive of Molecular Remission after Induction in Childhood B Lineage Leukemia.

Several studies showed a good correlation between minimal residual disease (MRD) obtained by flow cytometry (FC) and by PCR (Ig/TCR rearrangements). However, before FC is widely applied for therapeutic decisions we need exact and standardized criteria. Therefore, we pre-defined B-lineage cell subsets using 4 or 3-color FC and we established background levels for each subset in pre-defined time-points on therapy.

Methods. Children with ALL were treated by ALL IC BFM2002 protocol. This international flow cytometric trial MiniMini contained 4 laboratories (Croatia 36 patients, Israel 48, Hong Kong 33 and Czechia 136). 442 samples were measured simultaneously by PCR and 4color FC. Pre-defined subsets (28 values) were measured at 5 time points (d0, d8, d15, d33 and week 12). Cut off for positivity and negativity of individual subpopulations were assessed as follows: 1) negative samples by PCR for respective time point were randomly divided into training and testing cohort 2) 98.5 percentile of individual subpopulation in training cohort was selected as a cut off for positivity and negativity 3) cut off values were tested on testing cohort and PCR positive samples (total 337 samples), sample was considered positive when having at least one value above cut off (when more values were above cut off, highest value was considered as MRD).

Results. When all time points were analyzed together, sensitivity of FC MRD was 82% and specificity 88% There are no negative patients at day 8 in bone marrow by FC or by PCR. 5 samples at day 15 were positive by PCR (<=10^−4) and negative by FC, no sample was false positive by FC, sensitivity was 98%, no sample was negative by both methods. At day 33 and at week 12, sensitivity dramatically dropped to 34 and 13%, specificity was 80% and 100%. We next asked whether the early blast clearance assessed by FC MRD predicts positivity by PCR after induction. Univariate analysis showed that the blast clearance by FC both at days 8 and 15 predicts the PCR-based molecular response at day 33 and/or week 12. Patients with PCR-positive MRD at day 33 and at week 12 had more often over 50% blasts at day 8 BM (Fischer test p=0.0013 and p=0.04) and/or more than 5% blasts at day 15 by FC (p=0.0015 and p=0.0036).

Conclusion. With a minimized subjective bias and exact background values set for each cellular subset and timepoint, we did not reach satisfactory sensitivity at day 33 and at week 12. The FC-assessed leukemia clearance at days 8 and 15 predicts MRD by PCR at day 33 and at week 12. Since a non leukemic background influences sensitivity of FC MRD at day 33 and at week 12, complex approach using 6 and more color combinations is needed for precise distinction between leukemic and non leukemic B cells.