Minimal residual disease (MRD) testing performed on bone marrow (BM) samples has become a part of the risk group stratification procedure in several of the most progressive acute lymphoblastic leukemia (ALL) treatment protocols. MRD testing of peripheral blood (PB) instead of BM is not routinely performed, although PB sampling could cause less discomfort especially in children. It is well established that the level of MRD in PB and BM correlates well in T-ALL; the data concerning B-cell precursor (BCP)-ALL remain controversial, with most studies lacking sufficient number of samples taken during the early phase of treatment with quantifiable MRD in both compartments. We simultaneously evaluated MRD in BM and PB using immunoglobulin and T-cell receptor gene-based RQ-PCR. 221 paired samples from 47 children with BCP-ALL treated according to the Berlin-Frankfurt-Muenster (BFM) ALL IC-BFM 2002 protocol were taken at diagnosis (dg, n=47), day 8 (d8, n=39), day 15 (d15, n=44), day 33 (d33, n=34), week 12 (w12, n=31) and at the end of maintenance therapy (post-MT, n=26). As in the BM at d15, patients with lower MRD in PB at d15 were more likely to achieve MRD negativity in BM at d33 in the univariate analysis (p=0.01, Mann Whitney). Patients younger than 10 yrs had lower MRD in PB at d8 and at d15 than other patients (p=0.03 and p=0.01, respectively). Unlike in BM, patients with hyperdiploidy had lower MRD in PB at d15 than other patients excluding TEL/AML1 cases (p=0.05). There were no significant associations with diagnostic white blood cell count (WBC), sex, immunophenotype (cALL/prae-B ALL) or presence of TEL/AML1 fusion at any time point. Patients with MRD<1E-04 in PB at d15 had a 5-year relapse-free survival (RFS) of 100% vs. 62.5±9.9% for those with a higher MRD (p=0.0089). No such threshold could be set for dg, d8 and d33 PB MRD level. Low numbers of MRD-positive results at w12 and post-MT samples precluded statistical analysis. In 125 paired samples MRD was detected in both tissues, in 18 pairs in BM only and in 5 pairs in PB only. MRD levels in PB varied greatly and were a mean of 149-fold lower than in BM (range 0.04–8293). We next examined the prognostic impact of BM/PB MRD ratio. Patients having MRD levels in PB similar to those in BM (BM/PB MRD<10) at d8 and d15 were more likely to relapse (d8: RFS 88.1±6.4% vs. 61.5±13.5%, p=0.04; d15: RFS 89.5±5.7% vs. 54.5±15%, p=0.01). No such relationship was observed for dg or d33. In conclusion, our data show that in childhood BCP-ALL, MRD in PB is not simply proportional to the BM level and provides additional prognostic information. A higher relapse rate in patients with PB MRD level similar to that in BM suggests that leukemic blasts with the propensity for massive escape from BM to PB during the induction treatment have a great potential for giving rise to relapse.

Disclosures: No relevant conflicts of interest to declare.

Supported by GA UK 7393/2007, VZ MSM 0021620813 and MZ 000064203.

Author notes

Corresponding author