Inotuzumab Ozogamicin (InO) for Relapsed/Refractory Acute Lymphoblastic Leukemia (R/R ALL) in the Global Phase 3 INO-VATE Trial: Efficacy By MLL Status

Introduction: CD22, expressed in >90% of B-cell ALL cases, is an attractive target for B-cell malignancies. InO, an anti-CD22 antibody conjugated to calicheamicin, has demonstrated significantly greater efficacy vs standard care (SC) therapies in patients (pts) with R/R ALL in INO-VATE (Kantarjian et al, NEJM 375:740-53,2016). The MLL gene encodes Histone-lysine N-methyltransferase 2A, which exerts broad, positive effects on gene transcription. MLL rearrangements, including the canonical t(4;11) translocation, are associated with poor prognosis in B-cell ALL. Here we assess the potential impact of MLL status on efficacy of InO vs SC.

Methods: In this completed phase 3 trial (NCT01564784), pts with R/R ALL due to receive salvage 1 or 2 therapy were randomized to InO (1.8 mg/m²/cycle starting dose [0.8 mg/m² on day 1; 0.5 mg/m² on days 8 and 15 of a 21-28 day cycle for ≤6 cycles]) or SC (investigator choice of fludarabine/cytarabine [ara-C]/granulocyte colony-stimulating factor [FLAG], ara-C + mitoxantrone, or high-dose ara-C). Efficacy was assessed in the intent-to-treat (ITT) population; safety was assessed in all pts receiving ≥1 dose of study drug. MLL FISH analysis was done centrally using a "break-apart" probe mixture (hybridizes to the MLL locus within chromosomal region 11q32). Break-apart signals were scored as MLL rearrangements. Other MLL abnormalities eg, changes in 11q copy number were also detected. MLL translocation t(4;11) was assessed locally by karyotyping. Not all pts had samples evaluable by FISH. Final overall survival (OS) analysis occurred after 252 events (122 InO; 130 SC) were seen on 8-Mar-16; data as of this date are presented.

Results: ITT analysis included 326 pts (InO, n=164; SC, n=162). Baseline characteristics were balanced between arms. Of subjects evaluable for MLL status (InO, n=113; SC, n=114) , 20.4% (InO) and 23.7% (SC) exhibited any MLL abnormality (MLL+), including 9.7% and 7.0% with MLL rearrangements, 8.0% and 14.0% with MLL copy number gain (11q+/amplification), and 2.7% and 2.6% with MLL deletions (11q-), respectively. 12/23 MLL+ pts achieved CR/CRi with InO (5 had MRD negativity), while 7/27 MLL+ pts achieved CR/CRi with SC (3 had MRD negativity). ITT pts (2 arms combined) with MLL+ disease (n=50) appeared to have worse prognosis than pts with no MLL abnormalities (n=177) (median OS [mOS] 5.3 mo [95% CI: 3.6, 7.1] and 8.0 mo [95% CI: 6.0, 9.5], respectively). Pts with no MLL abnormalities had favorable OS with InO (n=90) vs SC (n=87) (unstratified HR 0.753, 97.5% CI: 0.509, 1.113; mOS 8.6 mo [95% CI: 5.8, 11.6] and 7.6 mo [95% CI: 5.2, 11.6], respectively). In contrast, this trend was not evident for MLL+ pts (InO, n=23; SC, n=27) (unstratified HR 1.331, 97.5% CI: 0.679, 2.609; mOS 5.3 mo [95% CI: 2.6, 7.1] and 5.5 mo [95% CI: 2.9, 9.4], respectively), but subgroup size was small. There were 13 pts (ITT) with t(4;11) per local laboratory (InO n=6; SC n=7). This small subgroup showed no apparent OS benefit for t(4;11) pts taking InO vs SC (unstratified HR 1.697, 97.5% CI: 0.388, 7.417; mOS 4.2 mo [95% CI: 2.2, 12.6] and 5.5 mo [95% CI: 1.0, 14.5], respectively). The potential relationship between CD22 expression and MLL status was explored. In ITT pts (2 arms combined) evaluable for MLL status, prevalence of MLL+, MLL rearrangements, and t(4;11) was significantly higher in those whose leukemic blasts were <90% CD22-positive than whose leukemic blasts were ≥90% CD22-positive (31%, 17%, and 14% respectively, vs 14%, 4%, and 2%; 2-sided P < 0.003 for each comparison). Given a potential association between CD22 expression and cytogenetics (CG), we further explored OS by CD22 in CG subgroups. For pts with normal CG and CD22 positivity ≥90%, OS was significantly improved for InO vs SC (unstratified HR 0.308 [95% CI: 0.130, 0.729]; 1-sided P= 0.0006). For pts with MLL+ CG, OS did not differ statistically for InO vs SC independent of CD22 positivity (≥90% or <90%). In general, recognizing smaller n, the MLL+ rearrangement and 11q+/amplification subcategories showed similar trends to MLL+. There were too few 11q- pts to assess trends.

Conclusion: While pts with MLL abnormalities responded to InO, study results align with published results demonstrating MLL translocation t(4;11) to be independently associated with poor survival in ALL, and indicate applicability to a wide range of MLL abnormalities. Further, baseline CD22 positivity <90% was associated with high prevalence of MLL abnormalities.