Phase III Trial (MPACT) of Weekly Nab-Paclitaxel Plus Gemcitabine in Metastatic Pancreatic Cancer: Influence of Prognostic Factors on Survival

Abstract

Background: In MPACT, nab‐paclitaxel (nab‐P) + gemcitabine (G) produced a longer median overall survival (OS; 8.5 vs 6.7 mo; hazard ratio [HR], 0.72; P = 0.000015) and median progression‐free survival (PFS; 5.5 vs 3.7 mo; HR, 0.69; P = 0.000024) vs G alone in patients with metastatic pancreatic cancer (PC). Here, we evaluated the influence of prognostic factors on OS and PFS. Methods: 861 patients with metastatic PC were randomized 1:1 and stratified by region, presence of liver metastases, and Karnofsky performance status to receive nab‐P 125 mg/m2 + G 1000 mg/m2 days 1, 8, and 15 every 4 weeks or G alone 1000 mg/m2 weekly for 7 weeks followed by 1 week of rest (cycle 1) and then days 1, 8, and 15 every 4 weeks (cycle > 2). A step‐wise, multivariate Cox proportional hazards (CPH) model (with significance level for entry of 0.20 and for stay of 0.10) was used to identify predictors of OS and evaluate treatment effect. A separate CPH analysis was performed for PFS. The following factors were tested: treatment arm, age (< 65 years vs > 65 years), sex, Karnofsky performance status (70‐80 vs 90‐100), geographic region (North America used as the reference), PC primary location (head vs other), presence of biliary stent, previous Whipple procedure, presence of liver metastases, presence of pulmonary metastases, peritoneal carcinomatosis, stage at diagnosis (IV vs other), number of metastatic sites, and serum CA19‐9 level. Results: Region of Eastern Europe, age > 65, Karnofsky performance status 70‐80, presence of liver metastases, and higher number of metastatic sites all predicted worse OS (Table). The treatment effect for OS remained significant after correcting for these factors in the multivariate analysis (HR, 0.72; 95% CI, 0.605‐0.849; P = 0.0001). In another multivariate analysis in which baseline CA19‐9 was added to the final model described above, the treatment effect HR was 0.67 (95% CI, 0.573 ‐ 0.794; P < 0.0001). Baseline CA19‐9, which was a significant predictor of OS by univariate analysis, was not significantly predictive after correction for the above factors. The results of the PFS analysis identified region of Australia, age > 65, Karnofsky performance status 70‐80, and presence of liver metastases as significant predictors of worse PFS (Table). The treatment effect for PFS remained significant after correcting for these factors (HR, 0.66; 95% CI, 0.544‐0.796; P < 0.0001, CPH model). Conclusion: The most important predictors of longer OS and PFS were higher Karnofsky performance status, age < 65 years, absence of liver metastases, and region of North America. A lower number of metastatic sites predicted longer OS but not PFS. Assignment to nab‐P + G was an independent significant predictor of improved survival. (Table Presented).