Comparison Across Trials with External Control for Label Expansion

In Section 4.2, we reviewed how to use an external control arm in clinical development for an NDA and BLA. In the drug development process, after the first NDA or BLA, a supplemental new drug application (sNDA) or supplemental Biologic License Application (sBLA) may also be applied to expand labels to more indications, patients, or regimens. For oncology drug development, since cancer can be divided into different stages and subtypes, label expansion is a very common way to benefit a broader patient population with the approved drugs. In addition, cancer treatments often adopt various combination treatment regimens, while the initial NDA or BLA may only cover one regimen. Moreover, in clinical practice, some regimens may be used off-label. These clinical practice regimens provide large amounts of RWD and historical data. Therefore, how to adopt these data as external controls for label expansion is also critical in drug development. In this section, two unique examples with different emphasis will be used to illustrate how to use external controls for label expansion. Both examples use the propensity score (PS) method to account for confounding factors due to lack of direct randomization. In this section, only the practical use of the PS method will be discussed. The theoretical details of the PS method will be discussed in later sections (Section 4.4.3 and Section 7). In Section 4.3.1, we will emphasize how to use PS matching (PSM) to construct an external control arm. In Section 4.3.2, we will illustrate how to use inverse probability treatment weighting.

Velcade® Label Expansion for Patients with Relapsed and/or Progressive Multiple Myeloma

Multiple myeloma (MM) is a cancer that forms in a type of white blood cell called a plasma cell. The treatment of MM has changed during the last 10-15 years. With the development of the proteasome inhibitor Velcade® and the immunomodulatory agent Revlimid, Velcade®-dexamethasone (VD)-based and Revlimid-dexamethasone (RD)-based therapies have become the backbone of combination therapy (Rajkumar and Kumar 2016). When Velcade® was approved as a single-agent treatment for relapsed MM, the efficacy and safety profile of Velcade® was characterized in the phase III APEX study and in the phase III DOXIL-MMY-3001 study (Kane et al. 2006). As the other research progressed, results from several clinical trials suggested that adding dexamethasone to bortezomib can improve response rates in patients. The VD doublet regimen was widely used in routine clinical practice for relapsed MM patients in 2013. However, a direct comparison of VD versus Velcade® monotherapy was missing. Therefore, a meaningful cross-study comparison using different adequate arms to compare the VD doublet regimen versus Velcade® monotherapy was needed. The label expansion of Velcade® plus dexamethasone in patients with relapsed and/or progressive MM who have received at least one prior therapy was approved in the European Union (EU) in 2013 (European Medicines Agency [EMA] 2013, Dimopoulos et al. 2015). This approval was based on an integrated analysis that comprised three clinical trials: MMY-2045, APEX, and DOXIL-MMY-3001. The study information is summarized in Table 4.12.

Outcomes were compared using subjects in the VD group from Study MMY-2045 to propensity-score-matched control subjects from the Velcade monotherapy arms in Study APEX and Study MMY-3001.

TABLE 4.12

Summary of Related Study Information for Velcade® Label Expansion Studies


Study Design

Treatment Groups

No. of Subjects


Phase 3 randomized study Primary' endpoint: time to progression

Group A: Velcade Group B: Velcade + CAELYX/ DOXIL

Group A: 322 Group B: 324


Randomized, open-label study in patients with RRMM

Group A: (Velcade) Group B: High-dose dexamethasone

Group A: 333 Group B: 336


Phase 2

Part 1 (nonrandomized treatment - all subjects):

Velcade + Dex

Part 2: (subjects with SD were randomized to Group В, C, or D below; subjects with >PR continued as in group A)

Group B: Velcade + Dex Group C: Velcade + Dex + cyclophosphamide Group D: Velcade + Dex + lenalidomide

Group A: 144 Group B: 7 Group C: 8 Group D: 4

The PSM in this analysis included eight identified variables that were related to clinical outcome: age; Eastern Cooperative Oncology Group (ECOG) score; type of myeloma; percentage of plasma cells; prior dexamethasone; and hemoglobin, creatinine clearance, and albumin. A total of 127 patients were identified in each arm using PSM.

The clinical outcomes were compared using ORR, progression-free survival (PFS), time to progression (TTP), and OS. The ORR in the VD group is significantly higher than that in the Velcade® monotherapy group. The odds ratio with 95% Cl was 3.769 (2.045,6.947). PFS also provided evidence to show clinical benefit for VD. The median PFS was 10.7 months for the VD arm and was 6.2 months for the Velcade® monotherapy group. The hazard ratio for VD versus Velcade® monotherapy was 0.511, with 95% Cl (0.309,0.845), and p-value 0.008.

In this example, the PSM method was performed to match the patients in the external control group. Instead of using several variables in the exact key variable matching procedure, PSM only uses PS as a summary of important variables in the matching procedure. Other PS methods will be introduced in the later sections.

< Prev   CONTENTS   Source   Next >