UCSF

Objective

The objective of this study was to evaluate whether building upon multidrug chemotherapy regimens represents a viable strategy in pancreatic cancer clinical trial design.

Methods

We performed a pooled analysis of all single-arm phase II studies in which a specific targeted agent (the anti-vascular endothelial growth factor monoclonal antibody bevacizumab) was added to gemcitabine-based cytotoxic doublets. The primary end point was overall survival (OS). Secondary end points included objective response rate, CA-19-9 biomarker response rate, and adverse event frequencies. Kaplan-Meier methods estimated time-to-event end points, whereas the Cox proportional hazard model estimated univariate hazard ratios of death.

Results

For the 300 patients included in the pooled analysis, median OS was 9.1 months (95% confidence interval, 8.3-10.2). Differences in OS were observed according to patients' baseline performance status (median OS, 10.4 vs 8.6 months for Eastern Cooperative Oncology Group 0 vs 1, respectively). Moreover, bevacizumab-related adverse events were not observed at increased frequency with gemcitabine-based doublets compared with historic data.

Conclusions

Recognizing the limitations of cross-study comparisons, these results compare favorably to those from Cancer and Leukemia Group B 80303, a phase III trial testing bevacizumab in combination with gemcitabine alone. This is the largest data set available to demonstrate the feasibility of building upon more intensive chemotherapy backbones in clinical trials of novel targeted agents in pancreatic cancer.