Acute lymphoblastic leukemia treatment leads to elimination of blasts and stepwise regeneration of normal hematopoiesis. Several studies identified prognostic relevance of minimal residual disease (MRD) in bone marrow (BM) before achieving complete remission (Giuseppe Basso et al., J Clin Oncol, 2009). Crucial question is how to assess BM quality at day 15 (d15) of ALL BFM protocols. In ALL BFM 2009 protocol good quality sample is defined as containing more than 2% erythroid precursors (EP) of nucleated cells. EP were defined as CD19neg(orCD7neg)CD45neg.

Two pt cohorts were included in the study. First cohort (Coh2000) consisted of pts treated by AIEOP BFM ALL 2000, n=196 (177 BCP ALL, 19 T ALL, median follow-up 5.4 yrs, range 0.025-10). AIEOP BFM ALL 2000 study was a PCR MRD based protocol (assessment at day d33 and d78) and flow cytometric MRD (FC MRD) was assessed only on research basis at d15. Second cohort (Coh2009) consisted of pts treated by AIEOP BFM ALL 2009, n=331 (292 BCP ALL, 39 T ALL, median follow up 4.8 yrs; range 0.0027-7.6). In Coh2009, both PCR MRD (d33, d78) and FC MRD d15 were used for risk stratification.

We asked following questions:

  1. What is the specificity and viability of EP defined by CD45 negativity? Is a definition based on bright expression CD71 more specific?

  2. Is the amount of EP different between B and T ALLs and between risk groups defined by FC at d15? What is the overall frequency of low EP at d15?

  3. What is the relationship between amount of EP and FC MRD at d15?

  4. Is there any prognostic relevance of low EP?


  1. Population of EP was selected based on negativity of CD45 and a lineage marker (CD19 or CD7) among nucleated cells, which were defined as positive by a SYTO nucleic fluorescent dye. We found a high amount of non-viable cells defined by 4′,6-diamidino-2-phenylindole (DAPI) positivity (6.5-96%, median 55%). When we added bright CD71 into the EP definition (EP CD71++), the percentage of DAPI positivity was significantly lower (0-66%, median 9%) (p<0.0001 in both cohorts).

  2. There is no difference in amount of EP at d15 between B and T ALL in either of the cohorts. The treatment reduction in SR pts (FC MRD d15<0.1%) was in Coh2009 used only in BCP ALL and we focused in further analyses on BCP ALLs. Overall, the EP were below 2% in 16% and 18% in Coh2000 and Coh2009, respectively. Within risk groups, EP below 2% at d15 occurred more frequently in Standard Risk (SR; 27%) than in non-SR (non-SR; 12%) in both cohorts (p=0.0002). The frequency of low EP appears higher than the expected frequency of technically poor samples. Moreover, it is unlikely that quality of BM aspiration would depend on the risk group of the pt. This further supports the role of normal BM response to presence of leukemic cells on one hand and to therapy on the other one.

  3. In both cohorts we found significant positive correlation between amount of EP and FC MRD at day 15 (Coh2000 p-value= 0.0016 (R 0.23); Coh2009 p value <0.0001 (R 0.33)). The correlation was significant in BCP ALLs only (Coh2000 p value=0.008 (R 0.26); Coh2009 p value<0.0001 (R 0.39)). The same significant correlation is observed in BCP ALLs with more precisely defined population EP CD71++DAPIneg (Coh2000 p value=0.04 (R0.18); Coh2009 p<0.0001 (R 0.35)). Part of the Coh2000 was treated with prednisone and part of the pts with dexamethasone between d8 and d28, whereas the entire Coh2009 received prednisone only. However, the frequency of low EP was not different between dexamethasone and prednisone-treated pts, the correlation between FC MRD and EP was significant only in prednisone treated pts (p=0.0003, R=0.33).

  4. We focused on SR BCP ALL pts (FC MRD d15<0.1%). We did not find difference in event free survival (EFS) between pts with amount of EP below and above 2%. This result indicates that the pts with low MRD and low EP are not just pts with hemodiluted BM samples.

Conclusion: Sample quality is essential question in the assessment of MRD in BM. Although low EP may indicate poor BM aspiration quality, it may also result from other biological factors. At d15 BCP ALL, these factors include the interaction of normal BM cells with leukemia, patient's risk group, and type of corticosteroid used. EPs should be detected using an erythroid marker, such as CD71. However, new markers of BM quality, less influenced by leukemia treatment, are needed.

Supported by Ministry of Health of CR, grant nr. 15-28525A, NV18-07-00430 and NV18-03-00343; Czech Science Foundation nr. P302/12/G101.


Brüggemann:PRMA: Consultancy; Incyte: Consultancy; Pfizer: Speakers Bureau; Roche: Speakers Bureau; Affimed: Research Funding; Regeneron: Research Funding; Amgen: Consultancy, Research Funding, Speakers Bureau.

Author notes


Asterisk with author names denotes non-ASH members.