OBJECTIVES: Surrogate endpoints are becoming increasingly important in health technology assessment, where decisions are based on complex cost-effectiveness models that require numerous input parameters. When not all estimates are available, the model may require simplification. The efficacy of docetaxel plus prednisone as a treatment option for 1L mCRPC was established by the TAX 327 Study; however, radiographic progression-free survival (rPFS) was not recorded. We used advanced surrogate modelling techniques to predict the unreported effect estimate of rPFS from TAX 327.

METHODS: A systematic literature review was conducted to identify randomized controlled trials (RCTs) evaluating asymptomatic or mildly symptomatic patients receiving 1L treatments for mCRPC. We searched MEDLINE®, Embase, Cochrane CENTRAL, and Cochrane Database of Systematic Reviews (inception-October 2022) using Ovid®. The Daniels and Hughes Surrogate Model was used to model jointly available data on treatment effects on overall survival (OS) and rPFS to predict the unreported effect on rPFS in TAX 327.

RESULTS: Thirty-eight RCTs with published data met the eligibility criteria and of these, 7 RCTs reported jointly OS and rPFS. One additional RCT recently reported both OS and rPFS and was included in the model. Model inputs for OS from TAX 327 favored docetaxel 75 mg/m² and 30 mg/m² vs mitoxantrone 12 mg/m² (hazard ratio [HR]: 0.76; 95% confidence interval [CI]: 0.62, 0.94 and HR: 0.91; 95% CI: 0.75, 1.11, respectively). Results for rPFS predicted by the model statistically favored docetaxel 75 mg/m² vs mitoxantrone 12 mg/m² (HR: 0.689; 95% CI: 0.553, 0.858; P=0.0009) and docetaxel 30 mg/m² vs mitoxantrone 12 mg/m² (HR: 0.746; 95% CI: 0.599, 0.929; P=0.009).

CONCLUSIONS: The use of advanced surrogate modelling techniques allowed for estimation of relevant parameters to enable implementation in of a cost-effectiveness model requiring rPFS for docetaxel while precluding loss of valuable clinical data.