Minimal Residual Disease (MRD)-Oriented Therapy of Adult Acute Lymphoblastic Leukemia (ALL): Final Results of Northern Italy Leukemia Group Study 09/00.

Clinical risk classifications fail to predict relapse (REL) in about 25% and 50% of high- and standard-risk (HR, SR) ALL patients (pts.), respectively, with potential under-/over-exposure to chemotherapy or allogeneic stem cell transplant (SCT, with intrinsic mortality ≥20%). Between 2000–5, we studied MRD as major predictive factor for REL and key decisional tool for SCT or less aggressive therapy in individual pts. with SR/HR ALL. Induction/consolidation consisted of Ida/V/P/Asp/CY (blocks 1–3,5,6,8), HD-MTX/Ara-C (4,7), CNS phase and imatinib (Ph+). Risk classes were SR (pre-B, WBC<30; cortical T, WBC<100), HR (pro-B, early/mature T; late CR; adverse cytogenetics; WBC>30 [B]/100 [T]), and very HR (VHR) i.e. Ph/t(4;11)+. All cases were screened for fusion genes or IgH/TCRβγδ/Kde rearrangements (23/22 PCR amplifications for B/T subsets, respectively). Marrow cells obtained before blocks 4,6,8 (week 10,16,22 i.e. timepoints [TP]1-3) were analyzed by RQ-PCR using 1 or more pt.-specific probes. MRD negativity (M^neg) was defined by negative/low-positive (<10⁻⁴ ) TP2 and negative TP3. MRD-based therapy consisted of maintenance in M^neg, SCT (family-related/unrelated) in M^pos (positive), and 2–4 autologous stem cell-supported Hypercycles (H/C: L-PAM/VP/6MP; HD-MTX/Ara-C) plus maintenance in M^pos without donor. VHR pts. were always eligible to SCT. Pts. with undefined MRD were treated by risk class. Of 280 total pts. (median age 38 years [range 16–66], SR 96, HR 90, VHR 89, unknown 5), 236 achieved CR (84%). 279 total probe(s) were obtained in 199 CR pts. (84%), with sensitivity ≥10⁻⁴ in 89%. 142 pts. completed consolidation and 94 did not (relapse 19%, early SCT 12%, violations/toxicity 7%); 18 (13%) lacked a probe and 12 (8%) were not correctly sampled at TP1-3. 112 pts. were exactly classified as M^neg (n 58, 52%) or M^pos (n 54, 48%), with some influence from VHR class (P=.044). With minimum FUP 1 year, 117 (50%) pts. relapsed and 28 (12%) died in CR (13 after SCT, 22%). For CR pts., 5-year overall survival (OS), REL-free survival (RFS) and cumulative REL incidence (CIR) were 54/42/56% in SR, 31/28/65% in HR and 28/24/65% in VHR. In M^neg pts., OS/RFS/CIR were 75/72/28% vs 33/14/79% in M^pos (P<.001), with no difference by clinical risk class, vs 39/37/55% in pts. with unknown MRD. The dominant predictive role of MRD for REL was confirmed by Cox regression (Haz. Ratio 5.72, P<.001). Also, when TP3 MRD results were compared to TP1, 36/39 (92%) and 41/47 (87%) pts. were confirmed M^pos and M^neg, respectively, while only 9 changed MRD risk class (P<.001); MRD remained the most significant prognostic factor, along with WBC count (B subsets: OS 79% in M^neg vs 33% in M^pos/WBC<30 vs 0% in M^pos/WBC>30; P<.001). Finally, M^pos pts. could be rescued by SCT (3-year RFS 40%) and H/C (53%), particularly those with WBC<30 (58%) who became M^neg (79%). Early MRD analysis with ≥1 probe(s) supports optimal therapeutic choices in ALL. M^pos ≥10⁻⁴ confers the highest risk, mandating for early SCT/experimental therapy. Confirmed M^neg status allows standard therapy with minimal mortality and expected RFS ≥70%.