Minimal residual disease after induction is the strongest predictor of prognosis in intermediate risk relapsed acute lymphoblastic leukaemia - long-term results of trial ALL-REZ BFM P95/96.

PURPOSE: This blinded prospective study was performed to optimise the risk assessment of children with a late isolated, combined or an early combined bone marrow (BM) relapse of precursor B-cell acute lymphoblastic leukaemia (ALL). The aim was to develop a reliable tool to identify patients with an intermediate risk relapse who are in need of haematopoietic stem cell transplantation (HSCT).

METHODS: Included were 80 children and adolescents with first intermediate risk BM relapse of ALL recruited in trial ALL-REZ BFM P95/96. We assessed the prognostic value of minimal residual disease (MRD) after induction therapy quantified by PCR using leukaemia clone-specific T-cell receptor/immunoglobulin gene rearrangements.

RESULTS: Molecular good responders (MRD < 10(-3), n=46) had a probability of event-free survival (pEFS) at 10 years of 76% standard error (SE) ± 6% and a cumulative incidence of second relapse (CIR) at 10 years of 21% SE ± 6%; pEFS of molecular poor responders (MRD ≥ 10(-3), n=34) at 10 years was 18% SE ± 7% and CIR 61% SE ± 9% (p<0.001). Cox regression analysis revealed MRD after induction to be the strongest independent prognostic parameter with a 6.6-fold increased risk (95% confidence interval 3.3-13.5, p<0.001) for molecular poor responders to suffer a subsequent adverse event compared to good responders.

CONCLUSION: In patients with intermediate risk BM relapse of ALL, low MRD after induction is associated with an excellent long-term prognosis with conventional chemo-/radiotherapy whereas patients with insufficient response have an extremely poor prognosis. Therefore, in the subsequent trial ALL-REZ BFM 2002, MRD is used to allocate molecular good responders to conventional post-induction therapy and molecular poor responders to allogeneic HSCT.