Bevacizumab may improve survival after progression in metastatic colorectal cancer

Results suggested that treatment with bevacizumab beyond initial disease progression may improve OS for patients with metastatic colorectal cancer, but the results were only hypothesis-generating and should not be taken as definitive.

Researchers conducted the Bevacizumab Regimens: Investigation of Treatment Effects and Safety (BRiTE) study, published in a recent issue of Journal of Clinical Oncology.

The study involved 1,445 patients with disease progression treated at 248 centers in 49 states. The researchers segregated patients into three groups: those who did not receive post-progression treatment (n=253), those assigned to post-progression treatment without the VEGF-inhibitor bevacizumab (n=531) and those assigned to post-progression treatment with bevacizumab (n=642).

Median OS was 25.1 months and median PFS was 10 months in the entire study population. Patients assigned bevacizumab (Avastin, Genentech) had a median OS of 31.8 months compared with 19.9 months in the no bevacizumab group and 12.6 months in the no treatment group.

The researchers added that after adjusting for various prognostic factors, bevacizumab also had a strong effect on survival beyond first progression (HR=0.49; 95% CI, 0.41-0.58) compared with patients in the no bevacizumab group.

Although it appears that bevacizumab after disease progression improved survival, Daniel G. Haller, MD, a professor of hematology/oncology at Abramson Cancer Center in Philadelphia and the journal’s editor in chief, said in an accompanying editorial written with Lee M. Ellis, MD, that the data did not “permit any conclusion about treatment with bevacizumab beyond progression.”

“What Dr. Grothey and colleagues did, obviously with the help of the pharmaceutical firm that sponsored this study, was look at three practices and then associate survival with these practices,” Haller told HemOnc Today. “The study doesn’t necessarily independently or prospectively test the hypothesis; it merely generates it.

“Those people who progressed on chemotherapy plus bevacizumab and received no further therapy did the worst. That may be intuitive, but you don’t know if they did worse because they didn’t receive therapy or they didn’t get treated because they were doing badly.”

Haller said that the journal rarely publishes the results of hypothesis-generating studies but decided to do so in this case because the hypothesis is biologically plausible and because the hypothesis would be tested in a randomized controlled trial. He added that the journal had also published an editorial by Mark N. Levine, MD, and Jim A. Julian, MD, both of Juravinski Cancer Centre in Ontario, Canada, about the benefits and drawbacks of registrational studies.

Levine and Julian noted that, even though BRiTE did include a broad range of patients, their characteristics at progression were unknown and characteristics of the patient groups were not compared. Because of what they called a “strong likelihood of bias,” they said results from BRiTE should be considered “exploratory in nature.”

“Far too often people take registrational trials and either because of enthusiasm, optimism or marketing want to take what is hypothesis-generating data and turn them into level-one evidence, which they clearly are not,” Haller said. – by Jason Harris

J Clin Oncol. 2008;26:5326-5334.
J Clin Oncol. 2008;26:5313-5315.
J Clin Oncol. 2008;26:5316-5319.