FDA Approval: Blinatumomab for Patients with B-cell Precursor Acute Lymphoblastic Leukemia in Morphologic Remission with Minimal Residual Disease

For adults with newly diagnosed acute lymphoblastic leukemia (ALL), rates of complete remission (CR) range between 78% and 93% (1–5). Some patients who achieve a first morphologic CR may survive long-term, but approximately one-third of patients with standard-risk ALL and two-thirds of patients with high-risk disease will still experience relapse. A second remission may be achieved following reinduction chemotherapy, but such treatment infrequently results in long-term survival (6). Given the high rate of relapse for patients with ALL who achieve morphologic CR, it is assumed that such patients have minimal residual disease (MRD) below the level of detection by light microscopy. MRD can be detected using high-sensitivity assays, including flow cytometry or molecular testing. Using these tests, MRD detected after intensive chemotherapy has been reported to be one of the most significant risk factors for relapse-free survival (RFS) and/or overall survival (OS) in patients with ALL in remission (7, 8).

There are little data from randomized trials that address treatment for detectable MRD in patients with ALL. UKALL 2003 included a randomization to test an augmented postremission regimen in first-line therapy for patients with clinical standard or intermediate-risk Ph-negative ALL who had MRD > 0.01% at the end of induction (9). With a median follow-up of 70 months, there was a significant improvement in event-free survival [EFS; HR 0.61; 95% confidence interval (CI): 0.39–0.98, P = 0.04; 5-year EFS 90% vs. 83%], but there was no significant difference in 5-year OS (93% vs. 89%). There are no randomized studies of chemotherapy that assess the long-term outcomes in adult or pediatric patients with B-cell precursor (BCP) ALL who were treated for MRD detected after consolidation. Blinatumomab has approval for treatment of relapsed or refractory (R/R) BCP ALL. Herein, we summarize the FDA's review of blinatumomab for the treatment of patients with BCP ALL in first or second complete remission (CR1 or CR2) with MRD ≥ 0.1%.

MT103-203 was an open-label, multicenter, single-arm study (BLAST; NCT01207388) that included patients with BCP ALL who were ≥ 18 years of age, had received at least three chemotherapy blocks of standard ALL therapy (e.g., induction, intensification, and consolidation), were in morphologic CR, and had marrow MRD at a level of ≥ 0.1% using an assay with a minimum sensitivity of 0.01%. Blinatumomab was administered as a constant infusion of 15 μg/m²/day (equivalent to the recommended dosage of 28 μg/day for patients > 45 kg) intravenously for all treatment cycles. Patients received up to 4 cycles of treatment.

The primary endpoint of the trial was undetectable MRD in marrow in an assay with sensitivity ≤ 0.01% (for the purposes of this analysis, we term this complete MRD response) after one cycle of blinatumomab. A sample size of 100 subjects had 90% power to detect a 61% response rate and exclude a 44% response rate assuming a true rate of 61% with a one-sided type I error of 2.5%. The FDA also assessed hematologic RFS, defined as time from the first dose of blinatumomab to hematologic or extramedullary relapse, secondary leukemia, or death due to any cause.

There were 116 patients enrolled and treated on MT103-203. The FDA excluded 29 patients from the analyses who were not in CR with hematologic recovery, who had screening MRD less than 0.1% or unquantifiable MRD, or who received other systemic treatments for leukemia within 2 weeks prior to blinatumomab or prior to the planned posttreatment MRD assessment. Because there were no data in the submission or the literature suggesting that MRD was prognostic for patients in third complete remission (CR3), the single subject in CR3 who was enrolled was also excluded. The final efficacy analysis population included 86 patients. The demographics and disease characteristics of the efficacy analysis population are shown in Table 1.

Among the 86 patients in the efficacy analysis population, a complete MRD response was achieved by 81% (95% CI: 72%–89%). As shown in Table 2, the complete MRD response rates were consistent across patients in CR1 and CR2. For patients who did not achieve a complete MRD response after the first cycle, none responded with additional cycles of blinatumomab. Duration of MRD response was not assessed.

In the supportive exploratory analyses, the estimated median hematologic RFS was 22.3 months for all subjects, 35.2 months for those in CR1, and 12.3 months for those in CR2 (Fig. 1). For the 12 patients in CR1 who did not undergo allogeneic hematopoietic stem cell transplantation (HSCT) after treatment with blinatumomab, the median hematologic RFS was not reached within the study follow-up time.

For 80 of the 86 patients, the MRD assay used for assessment of the primary endpoint had a sensitivity of ≤ 0.005%. With this higher sensitivity assay, undetectable MRD was achieved by 65 of the 80 patients (81%: 95% CI: 71%–89%), and the estimated median hematologic RFS was 24.2 months [95% CI: 17.9–nonestimable (NE)].

Safety was assessed in 137 patients with ALL in CR with complete or incomplete hematologic recovery and MRD ≥ 0.01%, who were treated with blinatumomab. Treatment-related mortality was 2% in this MRD-positive population. Adverse events of special interest are shown in Table 3. For context, key safety data for 706 patients with R/R BCP ALL treated with blinatumomab in six studies are presented (10, 11). The rates of cytokine release syndrome and sepsis are slightly lower in the MRD-positive ALL population than in the R/R ALL population, but the two populations had similar incidences of neurologic toxicities. Common (≥ 10%) and grade ≥ 3 adverse reactions are summarized in Supplementary Tables S2 and S3.

The approval of blinatumomab for the treatment of patients with BCP ALL in CR1 or CR2 with MRD ≥ 0.1% represents the first approval of a drug for treatment of patients in CR with MRD in any hematologic malignancy. In this setting, the FDA considered MRD to be a general measure of tumor burden, and no companion diagnostic device was required for this approval. Nonetheless, treatment of patients with ALL who have only detectable MRD would be a paradigm shift in acute leukemia therapy and some unique challenges presented during the review process.

The first issue was the acceptability of the MRD level used to identify the intended population. MT103-203 had an eligibility cut-off of MRD ≥ 0.1%. The FDA sought to determine the uniformity of the natural history of patients with MRD above this cut-off point. Others have addressed this question using measurement of MRD at the end of induction (Supplementary Table S4), but data were sparse for MRD detected after consolidation. The FDA performed a log-group analysis of hematologic RFS using patient-level historical data provided by the applicant (Supplementary Fig. S1). The analysis showed that patients in first remission with MRD ≥ 0.1% had a range of estimated median hematologic RFS of only 2.0 to 10.6 months by MRD log-group. Therefore, the FDA concluded that patients with BCP ALL in CR with MRD ≥ 0.1% after 3 blocks of chemotherapy have a uniformly poor prognosis.

The second issue arose in the assessment of reduction of MRD to < 0.01% as a clinical benefit of treatment with blinatumomab. Currently, there are no meta-analysis results demonstrating both trial-level and patient-level surrogacy for any level of MRD with RFS or OS. Most studies that proposed to establish “MRD negativity” as a clinical endpoint were based on responder analyses (i.e., MRD responder vs. MRD nonresponder), which has inherent deficiencies. The meta-analysis by Berry and colleagues 2017 (12) and other studies in the literature (Supplementary Table S5) demonstrated only that MRD has prognostic value in ALL. Thus, there remain insufficient data to validate MRD < 0.01% as a surrogate for clinical benefit in this population.

To estimate the impact of treatment with blinatumomab on long-term clinical outcome, the applicant submitted a propensity score analysis that assessed the effect of blinatumomab on RFS and OS by comparing subjects in first remission (with or without full hematologic recovery) from MT103-203 and Study 20120148. Study 20120148 was a retrospective study of outcomes for patients with BCP ALL and MRD ≥ 0.01% after 3 blocks of intensive chemotherapy. The applicant concluded from their analysis that the MRD-positive subjects treated with blinatumomab had a significantly greater hematologic RFS (but not OS) than those without blinatumomab. However, the FDA observed that conclusions were limited because of confounding by inappropriate data matching, lack of matching for covariates that would affect the RFS endpoint, inclusion of patients with incomplete hematologic recovery (CRi), lack of patients in CR2, lack of comparability between groups in the duration of follow-up, and unequal use of HSCT (13). Consequently, the FDA could not confirm the estimate of the benefit of blinatumomab compared with the historical control using the propensity score analysis.

Given the inability of the propensity score analysis to establish the benefit of blinatumomab, only robust results in a single-arm trial could be considered evidence of effectiveness. For the patients from MT103-203 in CR1, the estimated median hematologic RFS of 35.2 months was considerably longer than the expected RFS from the Study 20120148 and the literature, and this finding held true even in patients who did not go on to HSCT. For patients in CR2, the strongest published data demonstrated an estimated median event-free survival of 7 months with MRD > 0.01% after first salvage therapy (14). In MT103-203, patients in CR2 had an estimated median RFS of 12.3 months, which was consistent with a potential benefit for this subpopulation. In MT103-203, although response criteria for the primary endpoint required a minimum assay sensitivity of only 0.01%, 93% of patients had MRD evaluated in a more sensitive assay with sensitivity of at least 0.005%. In this subset, undetectable MRD was achieved by 81% of patients, and the estimated median hematologic RFS was 24.2 months. Thus, the depth of the MRD response after treatment with blinatumomab was also striking. Furthermore, blinatumomab was approved previously for the treatment of R/R ALL based on an overall survival benefit in a randomized trial. Taken together, these data represented considerable evidence of effectiveness.

Although it is tempting to draw conclusions regarding the long-term benefit of blinatumomab based solely on responder analyses of MT103-203, the crucial distinction must be made between the longer RFS and OS observed in patients who achieved complete MRD response compared with nonresponders (15) and any predicted treatment effect of blinatumomab. Without comparison in a randomized setting, the long-term benefit of blinatumomab in this population cannot be established. Therefore, two postmarketing requirements (PMR) were issued with this accelerated approval to confirm the clinical benefit of treatment with blinatumomab for residual disease. These PMRs refer specifically to trials E1910 (NCT02003222) and AALL1331 (NCT02101853).

Blinatumomab has a boxed warning for CRS and neurotoxicity and is available only under a risk evaluation and mitigation strategy (REMS). Safety analyses showed that these risks also exist for patients in morphologic CR with residual disease. Seven percent of these patients experienced a CRS event. Notably, the step-dose of blinatumomab approved previously was not used in this population and may have contributed to this risk. The incidence of neurologic toxicities was similar in this population compared with the R/R ALL population. Overall, the concerns addressed by the REMS are also pertinent to the patients in morphologic CR treated with blinatumomab.

Patients with BCP ALL who are in morphologic CR but have MRD ≥ 0.1% have an increased risk of early relapse and poor long-term outcomes. In MT103-203, there was a high rate of complete MRD response after a single cycle of blinatumomab, and the depth of this response was also notable. In addition, the estimated median hematologic RFS was considerably longer than expected on the basis of outcomes from patient-level historical control data and those reported in the literature, and, remarkably, patients who did not go on to undergo stem cell transplantation also appeared to derive benefit. It is essential to note that for the proposed indication, neither complete MRD response nor hematologic RFS alone in a single-arm trial would be sufficient to support approval. However, the totality of this data in the context of a drug with a known survival benefit in patients with more advanced disease provided substantial evidence of effectiveness and formed the basis for the accelerated approval. Given the uncertainty that exists due to the lack of a comparator arm and the novel use of molecular response as an endpoint, confirmation of clinical benefit in randomized trials is required.

No potential conflicts of interest were disclosed.

The Editor handling the peer review and decision-making process for this article has no relevant employment associations to disclose.

Conception and design: E.Y. Jen, D. Przepiorka, A.T. Farrell, R. Pazdur

Development of methodology: Q. Xu, D. Przepiorka, R. Pazdur

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Pazdur

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.Y. Jen, Q. Xu, A. Schetter, D. Przepiorka, Y.L. Shen, R. Sridhara, A. Deisseroth, R. Pazdur

Writing, review, and/or revision of the manuscript: E.Y. Jen, A. Schetter, D. Przepiorka, Y.L. Shen, R. Sridhara, A. Deisseroth, R. Philip, A.T. Farrell, R. Pazdur

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D. Przepiorka, R. Pazdur

Study supervision: R. Pazdur

Other (consultation and review): D. Roscoe

The authors thank Dr. Kristopher Kolibab for expert project management.