MATCHING-ADJUSTED INDIRECT COMPARISON STUDy

Indirect treatment comparisons (ITCs) can be limited by cross-trial differences. By combining individual patient data with published aggregate data, matching-adjusted indirect comparison (MAIC) studies may reduce observed cross-trial differences and provide decision makers with timely comparative evidence.

Chowdhury et al (2020)¹ conducted an anchored MAIC study using a Bayesian framework to compare the efficacy of ERLEADA plus androgen deprivation therapy (ADT) and enzalutamide plus ADT in patients with non-metastatic castration-resistant prostate cancer (nmCRPC). The anchored MAIC utilized individual patient-level data from the phase 3 SPARTAN study to match aggregated patient characteristics reported in PROSPER to control for potential treatment effect modification. In this approach, individual baseline data from SPARTAN patients were assigned inverse probability weights to ensure that the weighted mean or median baseline characteristics in SPARTAN closely matched the reported mean or median baseline characteristics in PROSPER.

Efficacy outcomes assessed were metastasis-free survival (MFS) and overall survival (OS). Since MFS was defined differently in the two studies, the MAIC used the PROSPER study definition of MFS, calculated as the time from randomization to radiographic progression or death within 112 days of treatment discontinuation. Following matching adjustment, ITC was performed to compare the MFS and OS hazard ratios (HRs) based on the reweighted SPARTAN population with reported HRs from PROSPER.

Prior to matching, the SPARTAN and PROSPER patient populations showed differences in the following baseline variables: median prostate-specific antigen (PSA) doubling time (PSADT; 4.4 vs. 3.7 months), percentage of patients with PSADT <6 months (70% vs. 77%), and median serum PSA at baseline (7.8 vs. 10.8 ng/mL). After matching, baseline characteristics were balanced (Table: Baseline Characteristics Before and After Matching).

Using the MFS definition from PROSPER, the MAIC results suggested a 73.6% probability that ERLEADA plus ADT has a more favorable MFS benefit compared to enzalutamide plus ADT; however, when using the MFS definition from SPARTAN, the probability decreased to 59.6% (Table: SPARTAN HRs [MFS and OS] Before [Original] and After [MAIC-Weighted] Matching for PROSPER Characteristics and MAIC Results). The MAIC results for OS suggested an 83.5% probability that ERLEADA has a greater survival benefit vs enzalutamide.

Chowdhury et al (2020)² conducted another anchored MAIC study using a similar Bayesian framework to compare the adverse events (AEs) and health-related quality of life (HRQoL) in patients with nmCRPC treated with ERLEADA plus ADT vs enzalutamide plus ADT. Following matching adjustment, ITC was performed for tolerability and HRQoL (assessed by Functional Assessment of Cancer Therapy-Prostate [FACT-P]) outcomes.

Baseline characteristics of patients enrolled in SPARTAN before and after matching were reported in the table above. In the ERLEADA plus ADT group of the SPARTAN study, 96.5% of patients had AEs before matching; this proportion was 96.3% after matching. In the placebo plus ADT group of SPARTAN, 93.2% and 94.0% had AEs before and after matching, respectively. Among patients in the ERLEADA group in SPARTAN, 24.8% and 23.7% had serious AEs before and after matching, respectively. In the ADT arm, these proportions were 23.1% and 22.3%, respectively.

ERLEADA had a higher probability of better tolerability vs enzalutamide as evidenced by reduced occurrence of fatigue, hypertension, decreased appetite, fall, headache, and nausea (Table: Adverse Events for the MAIC of ERLEADA vs Enzalutamide). Conversely, the probabilities of occurrence of asthenia, diarrhea, and arthralgia were higher with ERLEADA vs enzalutamide. The probability that patients treated with ERLEADA would have overall lower rates of any AE and serious AEs vs enzalutamide was 66.9% and 90.9%, respectively.

The MAIC results for HRQoL suggested a 73.1% probability of a more favorable improvement in FACT-P total score with ERLEADA vs enzalutamide (LS mean difference [95% CrI] 1.50 [-3.27 to 6.27]). Similar results were observed for all other scores, subscales, and composite scores assessed.