BackgroundIbrutinib (Ibr), an oral, first-in-class covalent Bruton's tyrosine kinase inhibitor, showed in the Phase 3 RESONATE trial significantly improved progression-free survival (PFS, hazard ratio [HR] =0.22, p<0.001) and overall survival (OS, HR=0.39,p=0.001) compared with ofatumumab (ofa) in patients with previously treated CLL who were not eligible for chemoimmunotherapy (Byrd et al, NEJM 2013). Long-term follow-up data from a single arm Phase 2 study have also demonstrated that patients treated with ibrutinib have long durable responses with a PFS at 2.5 years of 69% (Byrd et al, Blood 2015). While ofatumumab is a licensed comparator and included in treatment guidelines, some Health Technology Assessment (HTA) bodies require comparisons with a wider range of treatments. In the absence of direct head-to-head comparison of single-agent ibrutinib with other frequently used treatments in this patient population, additional comparative evidence against standard of care as observed in clinical practice can provide useful insights on the relative efficacy of ibrutinib. Naïve (unadjusted) comparisons of outcomes from different sources are prone to bias due to confounding, as treatment assignments were not randomly assigned, and populations can vary in important prognostic factors.The objective of this analysis was to compare the relative efficacy of Ibr versus physician's choice in R/R CLL-patients based on patient-level data from RESONATE pooled with an observational cohort, adjusting for confounders using multivariate statistical modelling.MethodsPatient-level data from the Phase 3 RESONATE trial (Ibr: n=195; ofa: n=196) were pooled with data from a retrospective observational study conducted in the Stockholm area in Sweden. This retrospective study collected efficacy and safety data from a detailed, in-depth retrospective review of individual patient files from 148 consecutively identified patients with R/R CLL initiated on second or later line treatment between 2002 and 2013 at the four CLL-treating centers in Stockholm, Sweden, with complete follow-up. Longitudinal follow-up in subsequent treatment lines was available for patients in 3rd (n=91), 4th (n=51), 5th (n=29), and 6+ (n=15) line, and as such individual patients could contribute information to the analysis for multiple lines of therapy, with baseline defined as the date of initiation of the actual treatment line. A multivariate cox proportional hazards model was developed to compare PFS and OS between treatments, including line of therapy, age, gender, Binet stage, ECOG, and refractory disease as covariates. Adjusted HRs and 95% CIs are presented vs. Ibr.ResultsAcross all treatment lines, fludarabine-cyclophosphamide (FC) (n=64), chlorambucil (n=59), alemtuzumab (n=33), FC+rituximab (FCR) (n=30), bendamustine+rituximab (BR) (n=28), and other rituximab-based combination chemotherapy (n=28) were the most frequently used treatments. Line of therapy, age and gender, Binet stage, ECOG performance status, and refractory disease were all independent risk factors for worse outcome on both PFS and OS. The adjusted HR for PFS and OS pooled observational data versus Ibr were 6.80 [4.72;9.80] (p<0.0001) and 2.90 [1.80;4.69] (p<0.0001). HR's for PFS/OS versus most frequent treatment regimens ranged between 2.50/1.82 (FCR) and 14.00/5.34 (anti-CD20 Mab). Baseline adjusted results for the Ofa-arm in RESONATE were comparable for both PFS and OS to outcome data from the consecutive historical cohort, however OS outcomes for Ofa were partly confounded by cross-over to Ibr.ConclusionsComparison of results from the Phase 3 RESONATE study with treatments used as part of previous standard of care in a well-defined cohort of consecutive Swedish patients shows that ibrutinib is superior to physician's choice in patients with relapsed/refractory CLL, suggesting a more than 6 fold improvement in PFS and almost 3 fold improvement in OS. Results were consistent across all different physician chosen treatments and provides further evidence that ibrutinib improves both PFS and OS vs current and prior standard of care regimens. [Display omitted] DisclosuresÖsterborg:Janssen Cilag: Research Funding. Asklid:Janssen Cilag: Research Funding. Diels:Janssen: Employment. Repits:Janssen Cilag: Employment. Söltoft:Janssen Cilag: Employment. Hansson:Jansse Cilag: Research Funding. Jäger:Janssen Cilag: Research Funding.