Send to

Choose Destination
J Thorac Oncol. 2014 Sep;9(9):1411-7. doi: 10.1097/JTO.0000000000000274.

Biomarker analyses from a randomized, placebo-controlled, phase IIIb trial comparing bevacizumab with or without erlotinib as maintenance therapy for the treatment of advanced non-small-cell lung cancer (ATLAS).

Author information

*University of California Los Angeles, Los Angeles, CA; †Kaiser Permanente Northern CA, Antioch, CA; ‡Sarah Cannon Research Institute, Nashville, TN; §Northwest Medical Specialties, Tacoma, WA; ‖Sibley Memorial Hospital, Washington, DC; ¶Integrated Community Oncology Network, Jacksonville, FL; #The Mark H. Zangmeister Center, Columbus, OH; **Florida Cancer Specialists, Bonita Springs, FL; ††Arch Medical Services Inc., The Center for Cancer Care and Research, Saint Louis, MO; ‡‡National Taiwan University, Taipei, Taiwan; §§F. Hoffmann-La Roche Ltd., Basel, Switzerland; ‖‖Genentech Inc., South San Francisco, CA; ¶¶Weill Cornell Medical College and Thoracic Oncology Service, New York, NY; and ##Dana-Farber Cancer Institute, Boston, MA.



ATLAS compared bevacizumab plus erlotinib (B+E) with bevacizumab plus placebo (B+P) as maintenance therapy after first-line bevacizumab plus chemotherapy (B+C) for advanced non-small-cell lung cancer (NSCLC). Prespecified biomarkers were prospectively evaluated.


Tumor samples were analyzed for: epidermal growth factor receptor (EGFR) expression (immunohistochemistry [IHC]); EGFR gene copy number (fluorescence in-situ hybridization [FISH]); EGFR mutations (exon 19 deletions/L858R mutations); and KRAS mutations (exons 2/3). Progression-free survival (PFS) and overall survival (OS) were estimated.


Of 743 patients randomized to receive maintenance treatment (after four cycles of B+C without progression), 190 (B+E) and 177 (B+P) were evaluable for biomarker status. Median PFS (from randomization) was 4.4 months (B+E) versus 3.7 months (B+P; hazard ratio [HR], 0.75; 95% confidence interval [CI], 0.57-0.99), which was numerically similar to the intent-to-treat PFS. PFS benefit of B+E was observed across most biomarker subgroups. EGFR IHC, EGFR FISH, and EGFR/KRAS mutation status were not predictive of outcome. B+E-treated patients with EGFR mutation-positive NSCLC had longer PFS compared with B+P-treated patients (HR, 0.44; 95% CI, 0.22-0.86; p = 0.0139). Patients with KRAS wild-type disease had significant PFS improvements with B+E, compared with B+P (HR, 0.66; 95% CI, 0.485-0.914; p = 0.0105). No OS benefit of B+E was observed.


Patients with KRAS wild-type or EGFR mutation-positive NSCLC derived PFS benefits from B+E. However, EGFR IHC, EGFR FISH, and EGFR or KRAS mutation status were not strongly predictive of survival. A larger sample size would be needed to confirm the initial trends observed in this study.

[Indexed for MEDLINE]
Free full text

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center