OBJECTIVES

: CASTOR is an international randomised controlled trial comparing daratumumab in combination with bortezomib (Velcade^®) and dexamethasone (DVd) versus bortezomib plus dexamethasone (Vd) among patients with relapsed or refractory multiple myeloma who had received at least one prior line of therapy. Consequential to the trial’s international design, many patients, particularly in the Vd arm, received subsequent therapy with regimens not available in the UK upon progression. The standard OS analysis generates bias and would underestimate the true OS benefit of DVd in UK. The aim of this analysis is to adjust for this bias and to provide OS estimates generalisable to the UK setting in 2L patients.

METHODS

: A variety of methods were considered, however, due to data complexity and the array of treatment switches only the inverse probability of censoring weights (IPCW) method was viable. Patients receiving a treatment that is not available in the UK were artificially censored at time of initiation. Time-dependant stabilized weights were calculated using repeated logistic regression models, including baseline and time varying covariates with stepwise selection used to determine regression models that best fit the data. Adjusted HRs for OS were calculated using a cox proportional hazards model including baseline characteristics and time-dependant weights. As the availability of subsequent treatment differs between Scotland and England, analyses were conducted separately by country. This methodology followed the National Institute for Health and Care Excellence (NICE) Decision Support Unit (DSU) technical support document.

RESULTS

: Prior to adjustment, the OS hazard ratio (HR) was 0.51 (0.31; 0.84), p=0.0084 (unstratified analysis) in 2L patients. IPCW-adjusted HRs were 0.40 (0.22; 0.73, p= 0.0025) and 0.44 (0.24; 0.79, p= 0.0066) for England and Scotland, respectively.

CONCLUSIONS

: A higher OS benefit of DVd versus Vd is seen following adjustment for bias introduced by subsequent treatment not available in the UK setting.