A Matching-Adjusted Indirect Comparison of Acalabrutinib with and without Obinutuzumab Versus Zanubrutinib in Treatment-Naïve Chronic Lymphocytic Leukemia

Introduction

The second-generation Bruton tyrosine kinase inhibitors acalabrutinib (acala) and zanubrutinib (zanu) have not been compared with each other in a head-to-head randomized controlled trial (RCT). Acala with and without obinutuzumab (obin) was evaluated in the ELEVATE-TN RCT in treatment-naïve patients with chronic lymphocytic leukemia (CLL). Zanu was evaluated in treatment-naïve patients with CLL/small lymphocytic leukemia (SLL) in cohorts 1 and 2 of the SEQUOIA RCT. We used an unanchored matching-adjusted indirect comparison (MAIC) to compare the efficacy and safety of acala ± obin versus zanu in patients with treatment-naïve CLL/SLL without del(17p). Datasets were derived from comprehensive individual patient data (IPD) from ELEVATE-TN and published aggregate data from cohort 1 in SEQUOIA.

Methods

In this unanchored MAIC, we weighted IPD for the acala + obin and acala monotherapy arms from ELEVATE-TN to match zanu baseline data from SEQUOIA. We excluded patients with del(17p) in ELEVATE-TN to match cohort 1 in SEQUOIA, which excluded patients with del(17p).

Patients were matched based on variables considered prognostic and/or predictive of investigator-assessed progression-free survival (INV-PFS) in an exploratory multivariate Cox-regression analysis of ELEVATE-TN. These were age, ECOG status, Binet stage, bulky disease, B2 microglobulin, cytopenia, del(11q), trisomy 12, IGHV status,and TP53 mutation.

The efficacy analysis assessed INV-PFS in randomized patients (acala + obin, n = 162; acala monotherapy, n = 163; zanu, n = 241) using the Oct 2021 data cut-off (DCO) for ELEVATE-TN and the Oct 2022 DCO for SEQUOIA (median follow-up 58 versus 44 months). Pseudo-IPD for INV-PFS for zanu were obtained from Kaplan-Meier curves.

The safety analysis assessed the incidence of adverse events (AEs) and reported the odds ratios (ORs) of AEs in treated patients (acala + obin, n = 162; acala monotherapy, n = 162; zanu, n = 240). To compare the incidence of AEs, the ELEVATE-TN Sep 2020 DCO was used to match the median follow-up from the SEQUOIA Oct 2022 DCO (47 versus 44 months).

Results

In the efficacy analysis, the acala + obin and acala monotherapy effective sample sizes (ESSs) post-matching were 124 (76%) and 105 (64%), respectively. Matching had little impact on acala + obin and acala monotherapy INV-PFS (Figure 1). Post-matching, 36-month INV-PFS was higher with acala + obin (95%; 95% CI: 90-97) than with zanu (84%; 95% CI: 79-88). The MAIC-weighted Cox hazard ratio (HR) found INV-PFS to be superior with acala + obin versus zanu (HR: 0.41; 95% CI: 0.23-0.74; Figure 1). No evidence of a difference between acala monotherapy post-matching and zanu was found when looking at 36-month INV-PFS (86%; 95% CI: 78-91 versus 84%; 95% CI: 79-88; HR: 0.91; 95% CI: 0.53-1.56).

In the safety analysis, the acala + obin and acala monotherapy ESSs post-matching were 123 (76%) and 103 (64%), respectively. There were no significant differences in the odds of having most types of AE with acala + obin versus zanu, except for higher odds of having any grade neutropenia (OR: 2.19; 95% CI: 1.33-3.60) and arthralgia (OR: 2.33; 95% CI: 1.37-3.96). Acala monotherapy showed lower odds of having any grade hypertension (OR: 0.44, 95% CI: 0.20-0.99; Figure 2) and no significant differences in the odds of having other types of AEs versus zanu.

Conclusions

In this MAIC of ELEVATE-TN and SEQUOIA evaluating patients with treatment-naïve CLL/SLL without del(17p), acala + obin had a longer INV-PFS versus zanu, while there was no evidence of a difference between acala monotherapy and zanu. Acala + obin and acala monotherapy generally had similar safety profiles to zanu. However, acala monotherapy was associated with lower odds of any grade hypertension versus zanu, while acala + obin was associated with higher odds of any grade neutropenia and arthralgia versus zanu.

Although the results of MAIC analyses can only be hypothesis-generating, these results address an important data gap by systematically comparing these two commonly used regimens where randomized, prospective data are not available.