Comparing Enzalutamide and Darolutamide for Prostate Cancer

In an interview with Targeted Oncology, Andrew Armstrong, MD, MSc, medical oncologist at Duke Cancer Institute, discusses a statistical methodology comparing enzalutamide (Xtandi) and darolutamide (Nubeqa) in metastatic hormone-sensitive prostate cancer (mHSPC).

The analysis used a matching-adjusted indirect comparison (MAIC), and it drew from 2 specific clinical trial datasets: the ARCHES study and the ARANOTE study. The ARCHES study, which investigated enzalutamide, was particularly valuable because it provided access to individual patient-level data. This is a critical prerequisite for performing a MAIC, as it allows for the precise statistical weighting and adjustment of patient characteristics to create a comparable population. The ARANOTE study, on the other hand, which examined the efficacy of darolutamide, provided only trial-level data.

Both trials shared several important design features. They both enrolled men with metastatic hormone-sensitive prostate cancer who had not been previously treated with an androgen receptor inhibitor. They were both global, multi-center studies. Furthermore, they used the same rigorous imaging frequency to monitor patients' disease progression, measured prostate-specific antigen (PSA) outcomes, and shared the identical primary end point of radiographic progression-free survival (rPFS).

Despite these similarities, there were crucial differences between the two study populations that required careful adjustment. For example, the ARANOTE study did not include any patients from the United States, with its patient cohort largely drawn from South America, Eastern Europe, and Asia. There were also notable imbalances in baseline characteristics, such as patient performance status and the proportion of individuals who had received prior docetaxel chemotherapy. The ARCHES trial included about 17-18% of patients with prior docetaxel exposure, while the ARANOTE study had a 0% rate. These differences are significant confounding factors that could skew results if not properly accounted for.

The MAIC analysis meticulously considered all these potential confounding factors, including disease volume, patterns of disease spread, functional status, and geographic location. The results of this rigorous adjusted comparison were quite surprising. The analysis indicated that enzalutamide showed superior efficacy for the primary endpoint, delaying radiographic progression-free survival by nearly 50% compared to darolutamide. This finding was statistically significant even after the careful adjustment for the differences in the patient populations.

It is crucial to acknowledge the limitations of this MAIC. A significant drawback is that the analysis did not include overall survival (OS) data, which is widely considered the gold standard for evaluating cancer treatments. This was primarily because the ARANOTE study's OS data is not yet mature. It's also possible that a definitive OS endpoint will never be reached due to factors such as patient crossover to other life-prolonging therapies and the general availability of these agents in clinical practice, which complicates long-term follow-up. Nevertheless, for the primary end point of rPFS, the MAIC provides the most compelling available evidence to date, suggesting that in a statistically matched patient population, enzalutamide may offer a more significant delay in disease progression. This valuable insight can help guide clinicians in making informed treatment decisions in the absence of a direct head-to-head trial.