OBJECTIVES: Fulvestrant (FUL) and everolimus plus exemestane (EVE+EXE) are common therapies for advanced breast cancer resistant to prior endocrine therapy. PALOMA-3 found the new regimen palbociclib plus fulvestrant (PAL+FUL) was associated with clinically meaningful improvements in overall survival (OS) versus FUL, despite narrowly missing statistical significance. As no direct data are available for PAL+FUL versus EVE+EXE, this study’s objective was to indirectly estimate comparative OS between PAL+FUL and EVE+EXE.

METHODS: A systematic literature review of clinical evidence was conducted. Two separate network meta-analyses (NMAs) were formed using a network of studies linking OS evidence from PAL+FUL to EVE+EXE: fractional polynomial (FP) methodology was used to capture a multi-dimensional treatment effect, and a proportional hazards (PH) NMA was carried out assuming constant treatment effects over time. The PH assumption was tested using cumulative log-log and Schoenfeld residual plots.

RESULTS: A linear network including five studies was used. PH tests showed the assumption did not hold for OS in most studies in the network. In the PH-based NMA, PAL+FUL showed OS benefit over EVE+EXE with a median hazard ratio (HR) of 0.74 (95% credible interval [CrI]: 0.51, 1.09), not statistically significant. In the FP NMA, the median HRs decreased over time from 0.94 at 6 months (95% CrI: 0.48, 1.89), to 0.90 at year 1 (95% CrI: 0.53, 1.54), and 0.76 at year 5 (95% CrI: 0.49, 1.20). The results were aligned with clinical trial data.

CONCLUSIONS: PALOMA-3 was the first trial to report OS for a CDK 4/6 inhibitor in this population. Using these data, the FP found PAL+FUL was associated with greater OS than EVE+EXE, findings backed by the PH-based NMA. While the results were not statistically significant, this could be due to the clinical trial results, studies not powered to detect differences in OS, and between trial variance.