MRD is a powerful prognostic tool, increasingly adopted in adult ALL for a risk-oriented application of stem cell transplantation (SCT) in MRD-positive (MRDpos) patients. This policy gives the opportunity to assess the therapeutic value of allogeneic (or autologous) SCT in MRDpos patients, who are very unlikely to be cured by chemotherapy. In this study we reanalyze the long-term results of a prospective Northern Italy Leukemia Group trial (Bassan et al, Blood 2009;113:4153), aiming to determine which MRD levels were predictive of a better SCT outcome in MRDpospatients after induction/consolidation therapy.
MRD was evaluated molecularly using one or two patient-specific probe(s) with sensitivity of at least 10−4. In remission patients, bone marrow MRD was assessed after chemotherapy courses no. 3 (week 10), 4 (week 16) and 6 (week 22). In the original risk model, MRDpos was defined by a level of 10−4 or greater at week 16 and/or any positivity at week 22. For this analysis, MRD results from all three time-points were pooled, and the highest value registered in individual patients was used to identify the MRD-negative group (always MRDneg) and three other subsets characterized by increasing levels of residual leukemia (less than 10−4 [MRDpos1], 10−4 and greater but less than 10−3 [MRDpos2], 10−3 and greater [MRDpos3]. Survival, disease-free survival (DFS), and relapse incidence were assessed and compared among these MRD groups, as well as outcome following SCT in MRDpossubsets, and this in accordance with protocol design i.e. in patients having SCT at end of consolidation phase after completion of the MRD study.
Three-hundred and four patients with Ph- ALL were treated between 2000–2006 (median age 35 years, range 16–68; male gender 57%), of whom 258 (85%) entered complete remission. One or more sensitive probe(s)were available for 200 responsive patients (77.5%), of whom 141 completed consolidation and 59 did not because of elective early SCT in t(4;11)+ ALL (n=13), relapse (n=41) and toxicity (n=5). Of 141 evaluable patients, 136 completed the MRD study through weeks 10–22 (standard risk: 61 B-lineage and 14 T-lineage; high-risk: 41 B and 17 T; unspecified 3). Altogether, 64 patients were reclassified MRDneg (47%), 21 MRDpos1 (15.5%), 17 MRDpos2 (12.5%) and 34 MRDpos3 (25%). With a minimum observation time of 4 years and a maximum close to 12 years, estimated 6-year survival, DFS and relapse rates were 72.5%, 64% and 36% in MRDneg, 56%, 57% and 32% in MRDpos1, 48.5%, 46% and 50% in MRDpos2 and 23.5%, 15% and 76% in MRDpos3 cohorts, respectively (all P values <0.0001). Of 72 MRDpos patients, 44 (61%) underwent SCT as per protocol design (allogeneic 26, autologous 18). Although 6-year DFS rate was improved after allogeneic SCT (42% vs 20% with autologous SCT, P=0.09), the most influential factor for posttransplantation outcome was MRD level (DFS 48% in MRDpos1–2 group [n=25] vs 16% in MRDpos3 group [n19], P=0.025), and the best overall result was obtained with allogeneic SCT in MRDpos1–2 group (DFS 60% [n=15] vs 18% in MRDpos3 group [n=11], P=0.08).
In this study different MRD levels measured at weeks 10–22 were associated with a progressively worse outcome in the four patient subsets defined by the highest individual value at three study time-points. However, about one half of MRDpos1–2 patients could be salvaged by SCT, and even more with allogeneic rather than autologous SCT. Because of the very poor SCT results in MRDpos3 group, patients with MRD levels of 10−3 and greater should receive further therapy and not proceed to SCT until the MRD signal is below the 10−3 cutoff.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.