The efficacy of induction chemotherapy in childhood acute lymphoblastic leukemia (ALL) is usually evaluated on day 35. However, at this stage, many patients have already begun to recover and present with a regenerative bone marrow (BM) where hematogones may make the identification of residual blast cells problematic both in morphology and in flow cytometry (FCM). In the FRALLE (French Acute Lymphoblastic Leukemia) trials, evaluation is proposed on days 8, 21 and 35. Here we evaluated whether FCM performed on day 21 (D21), when hematogones are still absent, would prove informative.

The cohort reported here was constituted of 45 children aged between 1 and 20 years old (median 6) treated for ALL according to the FRALLE recommendations since 2006.

There were 81% B-ALL, 17% T-ALL and 2% of mixed phenotype acute leukemia (MPAL, T/My). At diagnosis, the mean percentage of BM blasts was 50%. Classification according to the European Group for Immunophenotyping of Leukemia (EGIL) was 3 B-I, 21 B-II, 11 B-III, 2 B-IV and 1 T-I, 2 T-II and 4 T-III. Extensive immunophenotyping at diagnosis identified a median of 3 leukemia associated immunophenotypes (LAIP, range 1-5), defined as discriminant from hematogones.

Corticosensitivity was defined on a complete blood count (CBC) as less than 1 G/L of blast cells on day 8. Chemosensitivity was assessed on a bone marrow aspiration at day 21, both morphologically (< 5% blasts) and in FCM (MRD0). Molecular biology (according to Biomed2) was performed on BM samples collected on days 35 (MRD1) and 70 (MRD2). Follow-up median time was 59 months (3-276).

Corticosensitivity was observed for 39/43 patients (one had received corticosteroids for a tonsillitis before being referred and diagnosed with ALL and another one had less than 1 G/L of blasts at diagnosis). Five/44 patients were identified as chemoresistant by morphology on D21 (one aplastic sample). Enough cells were available for minimal residual disease (MRD) by FCM in 43 patients, on bone marrow collected on EDTA. As a mean, 586 328 total nucleated cells were acquired in FCM (range 9 616 - 1 751 000) thereby providing good sensitivity. Multiparameter FCM in 6 to 8 colors was performed on a single tube, customized according to each patient’s LAIP. Five MRD thresholds were defined as follows : level 1, >10-2 detected blasts; level 2, 10-3- <10-2detected blasts; level 3, 10-4- <10-3detected blasts; level 4, 10-5- <10-4 detected blasts or no event detected; level 5, <10-5detected blasts or no event detected. The table below indicates the partition of patients according to these MRD levels on D21.

Table 1
LevelPatient numbersDetectable MRDAbsence of MRD
LevelPatient numbersDetectable MRDAbsence of MRD

Event-free survival (EFS; Kaplan Meier Log rank test) was statistically significant (p=0.023) when comparing patients with level 4 or 5 MRD to the others. Level 3 patients had an intermediate EFS and OS.

In an independent cohort of 79 patients with evaluable FMC MRD on day 21 from a different center, the same good prognosis value of MRD lower than 10-4at that time point was confirmed for EFS (p=0.03).

Childhood ALL has become of relatively good prognosis with the progress in therapies. However, it is noteworthy that half the patients had detectable MRD at levels 1, 2 or 3, which probably prompted the clinicians to intensify treatment. However, below level 3, i.e.<10-4, FCM MRD on D21 is of excellent prognosis, even if detectable.

In summary, this work demonstrates the feasibility of FCM evaluation on D21, with the advantage of not being complicated by the presence of hematogones. Moreover, interpretation remains delicate in morphology at this time point where BM smears may be very poor. FCM thereby provides precious early information on chemosensitivity with a single patient-adapted antibody combination.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution