Minimal Residual Disease (MRD) and Risk-Oriented Therapy in Adult Acute Lymphoblastic Leukemia (ALL).

In adult ALL standard prognostic models do not predict relapse with great accuracy, which implies under-/over-exposure to a given therapy, especially transplants. A study is exploiting MRD as “best” predictive factor and decisional aid for therapy in unselected patients.

Adults with ALL underwent induction-consolidation with IDR/V/ASP/P/CY (cycles 1–3,5,6,8), HD-MTX/Ara-C (cycles 4,7), and CNS prophylaxis. Imatinib was added in Ph+ ALL. For MRD, 3 marrow samples were taken before cycles 4, 6, 8, to be analyzed by PCR/RQ-PCR with case-specific probe(s) targeting fusion genes or Ig-H/TCR rearrangements. MRD negativity was defined as negative/low-positive (<10e-4) timepoint (TP) 2 and negative TP3. MRD- patients received standard maintenance (1st year 6MP/MTX and V/P or CY pulses; 2nd year 6MP/MTX); MRD+ ones were eligible to allogeneic SCT (related/unrelated donor) or alternatively, to autologous stem cell-supported “hypercycles” with MEL/VP/6MP (cycles 1,3) and HD-MTX/Ara-C (cycles 2,4) plus Rituximab if CD20+, and maintenance. Patients with Ph/t(4;11)+ ALL were eligible to early SCT regardless MRD study, and those with undefined MRD were treated by clinical risk class. MRD monitoring was planned at 6, 12 and 24 months since TP3. The trial was activated 05/’00, enrolling 215 ALLs and 15 lymphoblastic lymphomas. Diagnostic characteristics were: age 16–66 years (median 37), B-lin 182, T-lin 48, Ph+ 49, t(4;11)+ 11. CR rate was 84%. So far, 109 patients completed consolidation and were eligible to MRD-oriented therapy. The MRD study was informative in 80 (73%), 14 lacked a probe(s) and 15 were poorly sampled. Altogether, 36/80 were MRD− (45%), with few variations among subsets: B-lin low-risk 38%, high-risk 44%, Ph+ 25%, T-lin 54%. Only a WBC count >100 and adverse cytogenetics predicted to some extent an MRD+ result (P=0.05), while MRD negativity was always confirmed as early as TP1 (week 10). Only 6 MRD− cases relapsed (17%), 1 with a new clone and 4 with suboptimal MRD study, compared to 20/44 (45.5%) MRD+ (P=0.000). DFS at 4 years was 76% in MRD− vs. 24% in MRD+ vs. 54% in the unknown MRD group. In multivariate analysis MRD was stronly predictive for relapse (RR 5.51, P=0.000) followed by a WBC count >30 (RR 3.64, P=0.001), the latter effect being almost exclusively seen in MRD+ patients. As regards MRD monitoring, conversion rate from MRD+ to MRD− was similar after SCT and “hypercycles” in the few cases evaluable (4/6 in each group), with a strong correlation in every prognostic/treatment group between overt relapse and prior positive MRD test (2/35 MRD− vs. 9/20 MRD+; P=0.000).

MRD is an invaluable prognostic tool for risk-oriented trials. Both an early analysis and monitoring support optimal therapeutic choices (with some reported exceptions), identify patients at greatest risk of relapse, and could allow comparative trials testing specific treatment elements in different ALL subsets.