Quality-Adjusted Survival With nab-Paclitaxel Versus Standard Paclitaxel in Metastatic Breast Cancer: A Q-TWiST Analysis.

Background: In this analysis we compared quality-adjusted survival outcomes between nab-paclitaxel (nab-P) and standard paclitaxel (Pac) using data from the nab-P phase III registration trial in metastatic breast cancer.

Patients And Methods: Quality-adjusted overall survival was estimated using the quality-adjusted time without symptoms or toxicity (Q-TWiST) approach. Overall survival was partitioned into time without progression/Grade ≥ 3 adverse events (AEs) toxicity (TWiST), time with Grade ≥ 3 AE toxicity (TOX), and time after relapse (REL). Q-TWiST was calculated by multiplying mean time in each health state by its assigned utility (base-case utility values: time without symptoms of disease progression or toxicity of Grade ≥ 3 adverse events [TWiST] = 1.0, TOX = 0.5, and REL = 0.5). In threshold analyses, TOX and REL varied from 0.0 to 1.0 whereas TWiST was maintained at 1.0. Comparisons were made for the intent-to-treat population and the subset of patients initiating the study drugs as second or subsequent lines (2L+) of chemotherapy (per approved nab-P indication; 2L+ subpopulation). A ≥ 15% relative Q-TWiST gain (vs. mean Pac overall survival) was considered clearly clinically important.

Results: In the intent-to-treat population, nab-P (n = 229) versus Pac (n = 225) resulted in nonsignificant gains of 1.4 months of mean Q-TWiST (11.6 vs. 10.2 months; 95% confidence interval [CI], -0.03 to 2.8). In the 2L+ subpopulation, nab-P (n = 132) versus Pac (n = 136) resulted in a statistically significant gain of 2.2 months of mean Q-TWiST (10.5 vs. 8.4 months; 95% CI, 0.6-3.8), with a 17.1% relative Q-TWiST gain (threshold analysis range, 14.0%-19.5%, both figures significant).

Conclusion: In its approved indication for metastatic breast cancer, nab-P showed a statistically significant and clearly clinically important improvement in quality-adjusted survival time versus Pac in the 2L+ subpopulation.