OBJECTIVES: Acalabrutinib (A) is a next-generation Bruton tyrosine kinase inhibitor (BTKi) approved for chronic lymphocytic leukemia (CLL) treatment. ELEVATE-TN trial interim data (~47 months median follow-up [FU]) did not show survival benefit for A±obinutuzumab (O) versus chlorambucil+obinutuzumab (C+O) despite significantly superior progression-free survival (PFS). One explanation is C+O-to-A patient crossover (39%), which may have provided added C+O overall survival (OS) benefit. We evaluated OS by adjusting for crossover using a historical C+O cohort.

METHODS: The CLL11 trial demonstrated superior PFS and OS for C+O versus C and C+rituximab in first-line CLL prior to BTKi availability. Therefore, OS data for BTKi-naïve C+O patients were digitized and compared with A±O from ELEVATE-TN. Individual patient datasets were used to plot Kaplan-Meier (KM) curves and calculate hazard ratios (HR) using Cox regression in R software. To validate our hypothesis (impact of crossovers on relative survival benefit), we compared 28-month FU data, where we observed lower OS HRs (non-significant). CLL11 and ELEVATE-TN trials enrolled patients with similar baseline characteristics; however, this analysis did not account for within-trial heterogeneity.

RESULTS: There was a statistically significant reduction in risk of death with all 3 ELEVATE-TN treatments versus CLL-11 C+O: A+O versus C+O_CLL11 (HR=0.23 [95% CI=0.12-0.42, p=0.001]); A versus C+O_CLL11 (HR=0.43 [95% CI=0.27-0.70, p<0.05]); C+O versus C+O_CLL11 (HR=0.44 [95% CI=0.27-0.71, p<0.05]). In an additional analysis with 28-month FU comparing ELEVATE-TN arms with C+O from CLL-11, KM curves for C+O overlapped (HR=0.79 [95% CI=0.42-1.47], p>0.05), indicating no difference. However, HRs for A+O versus C+O_CLL11 (HR=0.39 [95% CI=0.18-0.84]) and for A versus C+O_CLL11 (HR=0.48 [95% CI=0.24-0.99]) were statistically significant.

CONCLUSIONS: The non-statistically significant OS for A±O versus C+O was likely due to treatment crossover. Conducting a treatment-switching analyses was challenging due to immature OS data; however, this analysis supports the hypothesis that crossover to A bolstered ELEVATE-TN C+O OS.