Molecular Response to Treatment Predicts Outcome in Isolated, but Not in Combined Bone Marrow Relapses of Childhood Acute Lymphoblastic Leukaemia: Interim Results of Trial ALL-REZ BFM 2002

In the current trial ALL-REZ BFM 2002 for treatment of children with relapsed acute lymphoblastic leukemia (ALL) assessment of minimal residual disease (MRD) has been applied to quantify response to induction therapy at a submicroscopic level in children with intermediate-risk relapsed ALL aiming to stratify treatment, and improve survival. Time point of measurement of molecular response to therapy in bone marrow (BM) and the MRD cut-off used are based on a prospective blinded MRD-study performed during the previous trial ALL-REZ BFM 96. Further, in order to gain new experiences about a possible control of post-induction treatment, MRD reduction kinetics have been measured during treatment until stem cell transplantation (SCT) or end of maintenance therapy. MRD-quantification has been performed using quantitative real-time PCR with clone-specific T-cell receptor/Immunoglobulin gene rearrangements. Between 01/2002 and 08/2008, in 161 patients with intermediate risk (S2) including late isolated BM relapses and early or late combined BM relapses of B-cell precursor ALL, MRD after the second induction course (block F2) was used for post-induction treatment stratification. About half of intermediate risk patients with BM involvement showed a poor response to therapy (MRD ≥10⁻³ after F2) and were thereupon allocated to SCT. Probability of event-free survival of these patients is 48% (SE [standard error] ±8.5%) in the current ALL-REZ BFM 2002 trial compared to 18% (SE±6.5%) in the previous trial ALL-REZ BFM 96. MRD reduction kinetics until SCT were very heterogeneous in the poor response group. About 10% of this group presented as molecular non-responder with a persisting MRD-level of ≥10⁻² after the 5^th treatment block. In comparison to the previous trial, in the current trial the proportion of early relapses which includes only combined relapses among S2 patients was higher (20%) in the group with a molecular good response (MRD <10⁻³ after F2) than in the group with MRD poor response (7%, p=0.018). Subsequent relapses in patients with MRD good response occurred mainly among combined relapses. Therefore, the probability of event-free survival of molecular good responders among isolated BM relapses was 86% (SE±6.6%), but among combined bone marrow relapses only 41% (SE±13.5%). Concerning the extramedullary compartment involved the ratio between central nervous system and testes changed from 1:1 in the ALL-REZ BFM 96 trial to 2.8:1 in the ALL-REZ BFM 2002 trial. In conclusion, the hypothesis that the prognosis of S2-patients with poor MRD-response can be improved by allogeneic SCT seems to be confirmed in the current trial. MRD is a reliable marker for quantification of response to therapy in the BM, but not in case of extramedullary involvement. In future trials, this observation might result in a modification of clinical decisions. Furthermore, we are analysing the dynamic of treatment-response including all quantitative MRD-values from diagnosis until SCT in order to establish a more comprehensive individual risk-score.