Inotuzumab Ozogamicin Versus Standard Care for Acute Lymphoblastic Leukemia

Study Design and Patients

In this open-label, 2-arm, randomized phase 3 trial (NCT01564784), eligible patients were aged ≥18 years with relapsed or refractory (≥5% bone marrow blasts based on local morphologic analysis) CD22-positive, Philadelphia chromosome (Ph)–positive or Ph-negative ALL, who were due to receive their first or second salvage treatment. See Supplementary Appendix for additional eligibility criteria. Patients were randomized (1:1) to inotuzumab ozogamicin or investigator’s choice of standard therapies; no crossover between arms was allowed. Stratification factors at randomization were duration of first remission (<12 or ≥12 months), salvage treatment phase (1 or 2), and age (<55 or ≥55 years). Responders could undergo stem cell transplant at the investigator’s discretion.

The protocol was approved by the Independent Ethics Committee and/or Institutional Review Board at each study center. Informed consent was obtained in accordance with the Declaration of Helsinki. The study was designed through a collaboration of the sponsor and lead investigators. The sponsor held the data included in this manuscript. Generated data tables were freely accessible to all authors, who together with sponsor representatives, were responsible for analyses. All authors vouch for data/analysis accuracy and adherence to the protocol, which is available at NEJM.org, and contributed to drafting and critically reviewing the manuscript, approving the final draft, and the decision to publish it. Editorial support for this manuscript was provided by Johna Van Stelten, PhD, and Simon Slater, PhD, of Complete Healthcare Communications, LLC, who developed the first draft under direction from the authors, and was funded by Pfizer Inc.

Treatments

Patients in the inotuzumab ozogamicin arm received a starting dose of 1.8 mg/m² intravenously per cycle (0.8 mg/m² on day 1; 0.5 mg/m² on days 8 and 15 of each cycle [cycle 1, 21 days; subsequent cycles, 28 days; ≤6 cycles]). Once complete remission (CR) or CR with incomplete hematologic recovery (CRi) was achieved, the day 1 dose was reduced to 0.5 mg/m² for the duration of the study. Patients in the standard therapy arm received (per investigator’s choice) either FLAG (cytarabine 2.0 g/m²/d on days 1–6; fludarabine 30 mg/m²/d on days 2–6; granulocyte-colony stimulating factor 5 μg/kg/d [or per institutional standard] during a 28-day cycle [≤4 cycles]), cytarabine plus mitoxantrone (cytarabine 200 mg/m²/d on days 1–7; mitoxantrone [Mitox] 12 mg/m²/d on days 1–3 during a 15- to 20-day cycle [≤4 cycles; dose reduction to ≥8 mg/m² allowed based on age, comorbidities and prior anthracycline use]), or high-dose cytarabine (HIDAC; 3 g/m² every 12 hours [≤12 doses; 1.5 g/m² recommended for patients aged ≥55 years]). These 3 regimens are commonly used for the treatment of relapsed and refractory ALL and were chosen for the comparator arm to provide investigators latitude to tailor the regimen to the patient’s condition, treatment history, and standard practice. Dose schedule modifications are detailed in the Supplementary Appendix.

Outcomes

This study has 2 primary endpoints (see Statistical Analysis section): (1) CR/CRi and (2) overall survival. Secondary endpoints included safety, remission duration, progression-free survival, stem cell transplant rate, and MRD rate among responders.

For all patients, marrow aspirate (or biopsies if clinically indicated) and disease assessments were performed at screening; days 16 to 28 of cycles 1, 2, and 3, then every 1 to 2 cycles; at the end-of-treatment visit; during planned follow-up visits; and as clinically indicated. CR and disease progression status were assessed using modified Cheson criteria³¹ (Supplementary Appendix); CRi was defined as for CR except with absolute neutrophil count <1000/μL and/or platelets <100,000/μL. MRD-negativity was assessed as <0.01% marrow blasts and was analyzed by a central laboratory using multicolor, multiparameter flow cytometry. Definitions of duration of remission among patients with CR/CRi, progression-free survival, and overall survival are provided in the Supplementary Appendix.

Treatment-emergent adverse events (from all causes) were defined as any event occurring between the first dose and ≤42 days after last dose, all treatment-related adverse events occurring after last dose, and all veno-occlusive liver disease/sinusoidal obstruction syndrome events (any cause) occurring within 2 years after randomization. Veno-occlusive disease/sinusoidal obstruction syndrome was assessed and diagnosed by the investigators and evaluated according to previously defined clinical criteria (Supplementary Appendix).

Statistical Analysis

The sample size was calculated to adequately assess differences in CR/CRi and overall survival independently by splitting the 1-sided alpha of 0.025 evenly between the 2 primary endpoints. With 218 patients and 1-sided alpha of 0.0125, the study had ≥88.5% power to detect a difference in CR/CRi probabilities of 61% in the inotuzumab ozogamicin arm versus 37% in the standard care arm; with 248 overall survival events and 1-sided alpha of 0.0125, the study had 80% power to detect a difference in overall survival (median of 6.45 months in the inotuzumab ozogamicin arm and 4.30 months in the standard arm; hazard ratio [HR], 0.67). All reported P values are 2-sided.

Patients and Study Treatment

A total of 279 patients (inotuzumab ozogamicin, n=141; standard, n=138) from 18 countries were randomized between August 27, 2012 and a data cutoff date of October 2, 2014 (Fig. S1), among whom 259 received ≥1 dose of study treatment and were included in the safety population (inotuzumab ozogamicin, n=139; standard, n=120). The remaining 20 patients were randomized but not treated by the cutoff date. An additional 47 patients were then randomized after this cut-off date for a total of 326 patients to obtain further survival data. The prespecified requirement for ≥248 events to trigger the final overall survival analysis was achieved on March 8, 2016 when 252 events were observed. Therefore, survival data as of March 8, 2016 are presented for 326 patients included in the ITT population. Data based on an unlocked trial database.

Among the safety population, a total of 369 treatment cycles were initiated with inotuzumab ozogamicin versus 152 with standard therapy (FLAG, n=106; cytarabine + mitoxantrone, n=29; HIDAC, n=17). Patients in the inotuzumab ozogamicin arm received a median (range) of 3 (1–6) treatment cycles versus 1 (1–4) cycle in the standard arm. Compared with inotuzumab ozogamicin, few patients received ≥2 cycles of standard therapy, as was expected (73% vs 22%).

Dose reductions were more common over the duration of study treatment with inotuzumab ozogamicin versus standard therapies (12% vs 3%), whereas dose interruptions were less common (3% vs 15%). Compared with standard therapy, more patients receiving inotuzumab ozogamicin discontinued treatment due to achieving a complete response (35% vs 15%), whereas fewer patients discontinued inotuzumab ozogamicin because of resistant disease (10% vs 40%; Fig. S1). Currently, no patients remain on active study treatment.

Efficacy

The final CR/CRi primary analysis was prespecified to include the first 218 patients randomized in the intention-to-treat population (ITT218: inotuzumab ozogamicin, n=109 and standard, n=109) and was assessed by an independent, blinded, central endpoint adjudication committee (CR/CRi in subsequently randomized patients was not adjudicated). The demographics and baseline characteristics of the ITT218 population were well-balanced across treatment arms (Table 1). In the ITT218 population, the CR/CRi rate was significantly higher with inotuzumab ozogamicin versus standard therapy (80.7% [95% CI, 72%–88%] vs 29.4% [21%–39%]; P<0.001). Thirteen patients randomized to standard therapy in the ITT218 population refused to start treatment; CR/CRi rate with the standard care arm would have been skewed by their inclusion in the denominator. In an “as-treated” analysis excluding these 13 patients (inotuzumab ozogamicin, n=109; standard, n=96), the CR/CRi rate was again significantly higher with inotuzumab ozogamicin (80.7% [95% CI, 72%–88%] vs 33.3% [24%–44%]; P<0.001). In a subgroup analysis of the ITT218 population defined by randomization stratification factors and baseline patient characteristics, CR/CRi rate per endpoint adjucation committee was significantly higher (P≤0.004) with inotuzumab ozogamicin for all factors examined, except baseline Ph+ or t(4;11) karyotype (Fig. 1). In both arms, most patients achieving a CR/CRi did so at the end of cycle 1 (inotuzumab ozogamicin, n=64/88 [73%]; standard, n=29/32 [91%]). Among patients with CR/CRi, the MRD-negativity rate was significantly higher with inotuzumab ozogamicin versus standard therapy (78.4% [95% CI, 68%–87%] vs 28.1% [14%–47%]; P<0.001; Table 2).

Among patients in the ITT218 cohort with CR/CRi (per investigator’s assessment: inotuzumab ozogamicin, n=85; standard, n=31), with or without follow-up stem cell transplant, the median (95% CI) remission duration was 4.6 (3.9–5.4) months with inotuzumab ozogamicin versus 3.1 (1.4–4.9) with standard therapy (overall HR [95% CI], 0.55 [0.31–0.96]; P=0.03; Fig. 2A). Remission duration by MRD status is shown in Fig. S2. Significantly more patients proceeded to stem cell transplant directly after treatment with inotuzumab ozogamicin versus standard therapy (41% [n=45/109] vs 11% [n=12/109]; P<0.001); more patients received myeloablative conditioning (inotuzumab ozogamicin, 71% [n=32/45] vs standard therapy, 75% [n=9/12]) versus reduced intensity conditioning (29% [n=13/45] vs 17% [n=2/12]). More patients achieving CR/CRi (per investigator’s assessment) proceeded directly to stem cell transplant after inotuzumab ozogamicin versus standard therapy (48% [n=41/85] vs 32% [n=10/31]; P=0.12); the remission duration for these patients was 5.5 (95% CI, 4.9–8.0) months for inotuzumab ozogamicin versus 5.7 (0.8–not reached) months for standard therapy.

In the ITT survival analysis (inotuzumab ozogamicin, n=164; standard, n=162), progression-free survival was significantly longer with inotuzumab ozogamicin vs standard care (HR, 0.45 [97.5% CI, 0.34–0.61]; P<0.001; Fig. 2B), with median progression-free survival 5.0 [95% CI, 3.7–5.6] vs 1.8 [95% CI, 1.5–2.2] months. The overall survival hazard ratio was 0.77 (97.5% CI, 0.58–1.03; P=0.04; Fig. 2C) Median overall survival for inotuzumab ozogamicin was 7.7 [95% CI, 6.0–9.2] vs 6.7 [4.9–8.3] months for standard care; 2-year overall survival rate, 23% [95% CI, 16%–30%] vs 10% [5%–16%]; Fig. 2C). The primary objective to demonstrate significantly improved final overall survival with inotuzumab ozogamicin vs standard care was not met at prespecified boundary of P=0.0208. However, overall survival data appeared to depart from the proportional hazards assumption; therefore, an exploratory posthoc restricted mean survival time analysis was applied³² (truncation time, 37.7 months) to more precisely define the clinical benefit of inotuzumab ozogamicin. This analysis demonstrated a longer mean overall survival with inotuzumab ozogamicin vs standard care (mean [SE] 13.9 [1.10] vs 9.9 [0.85] months; P=0.005).

Safety

Among the safety population (data cutoff: October 2, 2014), the most common all-cause and treatment-related hematologic treatment emergent adverse events in both arms were cytopenias (Table S1). The incidence of grade ≥3 thrombocytopenia was lower with inotuzumab ozogamicin versus standard therapy (37% vs 59%); fewer patients received platelet transfusions with inotuzumab ozogamicin versus standard therapy (64% vs 95%; median [range] days of transfusion, 5 [1–51] vs 7 [1–26]). Grade ≥3 febrile neutropenia occurred in 24% of patients treated with inotuzumab ozogamicin and 49% of patients treated with standard therapy. Common nonhematologic treatment emergent adverse events (all grade; grade ≥3) with inotuzumab ozogamicin included nausea (32%; 2%), headache (28%; 1%), and pyrexia (27%; 4%); most common nonhematologic treatment emergent adverse events (all grade; grade ≥3) with standard therapy were nausea (47%; 0%), pyrexia (43%; 5%), and diarrhea (40%; 1%). Overall, the number of patients experiencing serious adverse events was similar with inotuzumab ozogamicin and standard therapy (48% vs 46%) during a median of 3 versus 1 treatment cycle, respectively; febrile neutropenia was the most frequently reported serious adverse event in both arms (inotuzumab ozogamicin, 12%; standard, 18%; Table 3).

Liver toxicities were more common with inotuzumab ozogamicin versus standard therapy (Table S1) for the total duration of treatment cycles initiated (inotuzumab ozogamicin, 369 cycles; standard care, 152 cycles); the most frequent any grade liver-related treatment emergent adverse events were increased aspartate aminotransferase (inotuzumab ozogamicin, 20%; standard, 10%), hyperbilirubinemia (15%; 10%), and increased alanine aminotransferase (14%; 11%). All cases of veno-occlusive disease were reported for up to 2 years from randomization. Veno-occlusive disease occurred more frequently with inotuzumab ozogamicin compared with standard therapy (11% [n=15] vs 1% [n=1]). In the inotuzumab ozogamicin arm, 5 patients developed veno-occlusive disease during or shortly after study treatment (2 of these 5 patients received prestudy stem cell transplant). Of 48 patients in the inotuzumab ozogamicin arm who proceeded to transplant, 10 developed veno-occlusive disease after transplant (3 with prestudy transplant), 7 of whom had received defibrotide (2 cases resolved, 4 ongoing, and 1 fatal). Of 20 patients in the standard therapy arm who proceeded to transplant, 1 developed veno-occlusive disease after transplant; no cases of veno-occlusive disease occurred during standard therapy. The median (range) time to develop veno-occlusive disease after transplant in the inotuzumab ozogamicin arm was 16 (3–39) days. In a multivariate analysis of baseline factors associated with post-transplant veno-occlusive disease, alkylator conditioning regimen (dual [n=8] vs single [n=33]) was the only significant covariate (P=0.04; Table S2).

A total of 17 and 11 grade 5 treatment emergent adverse events occurred with inotuzumab ozogamicin and standard therapy, respectively, including 4 and 2 fatal adverse events deemed to be treatment-related. Two treatment-related deaths due to veno-occlusive disease occurred with inotuzumab ozogamicin (both after poststudy transplant).