In spite of the excellent outcome of children with ALL, those who relapse fare poorly. To date, time to relapse is the most powerful predictor of outcome. Most studies indicate that children with early marrow relapse (ER), occurring within 36 months of diagnosis, have a long-term event-free survival (EFS) of 10–20%, while those with a late marrow relapse (LR) have an EFS of 40–45%. MRD, detected either by flow cytometry or polymerase chain reaction is an important prognostic factor in patients with newly diagnosed ALL, and recent studies indicate that it is prognostic of outcome following relapse. However studies in relapse are more limited and generally only one time point at end induction has been tested. COG AALL01P2 is a pilot study designed to administer three sequential blocks of intensive therapy to children and adolescents with first marrow relapse of ALL with the goal of improving remission rates and achieving a lower level of disease burden prior to post induction therapy such as bone marrow transplantation. We measured MRD in 52 patients on COG AALL01P2 by flow cytometry using the 4-color combinations CD20-FITC/CD10-PE-CD45-PerCP/CD19-APC; CD34/9/45/19; and CD38/58/45/19. We detected MRD in 35 patients (67%) after the first block, a much higher figure than the 28% seen at end induction in more than 2000 patients with newly diagnosed ALL, consistent with the poor prognosis of relapse patients. Of 32 ER patients, 26(81%) were MRD+ compared to only 45% (9/20) of LR patients (p<.01). Serial MRD samples measured after each block of therapy were collected in 32 patients, and several patterns of kinetic changes in MRD levels were noted. Twenty-two patients were classified as responders: 7 patients had no detectable disease at any time point(R1); 9 patients were initially MRD positive, but became negative(R2); and 6 additional patients showed at least a 1 log decrement in MRD(R3). No patient negative at any time point became positive at a later time point. Six patients had >1% MRD at all time points tested, while 4 patients with a level <1% on at least one timepoint had an increase of at least 1 log in MRD burden between 2 timepoints. All 10 of these latter patients were ER while all 7 R1 patients were LR. While it is too early to correlate MRD with EFS, our results show that the incidence of MRD is different in ER and LR patients in a pattern that correlates well with expected outcome. Serial measurements of MRD can be used to identify different patterns of treatment response and may be useful to dissect out effective components of therapy. Moreover, it may be possible to assess efficacy of novel agents added to this backbone of therapy by measuring changes in MRD response.

Author notes

Corresponding author

Sign in via your Institution