SAN FRANCISCO—Overall survival was no better for advanced colorectal cancer patients with KRAS-normal tumors treated with brivanib alaninate plus cetuximab than for those treated with cetuximab alone in the randomized Phase III National Cancer Institute of Canada and Australasian Gastro-Intestinal Trials Group CO.20 study (Abstract 386) reported here at the Gastrointestinal Cancers Symposium.
Although the combination appeared to improve response rates and progression-free survival, suggesting that brivanib may have activity in this disease setting, the unexpectedly large toxicity profile may have compromised efficacy.
With a median follow-up of 19 months, the 376 patients in the combination arm had a median overall survival of 8.8 months, compared with 8.1 months for the 374 patients in the cetuximab-plus-placebo group. The difference was not statistically significant and did not meet the primary endpoint of an improvement of 3.2 months with the combination, noted Lillian Siu, MD, Director of the Phase I Program and Co-Director of the Robert and Maggie Bras and Family Drug Development Program at Princess Margaret Hospital and Professor of Medicine at the University of Toronto, who presented the study results.
Median progression-free survival for patients treated with brivanib-cetuximab was 5.0 months compared with 3.4 months for patients in the cetuximab-placebo arm, which was a statistically significant difference. Similarly the objective response rates were significantly different in the two arms, at 13.6% for the combination and 7.2% for the cetuximab-placebo.
The rate of Grade 3 or higher adverse events was significantly higher in the brivanib-cetuximab patients than in the control arm (78% vs. 53% for non-hematologic toxicities). The most common severe toxicity was fatigue (25% in the brivanib-cetuximab arm vs. 11% in the cetuximab-placebo arm), followed by hypertension (11% vs. 1%), rash (10% vs. 5%), abdominal pain (10% vs. 5%), diarrhea (7% vs. 3%), dehydration (7% vs. 2%), and anorexia (5% vs. 1%).
Biochemical abnormalities were also significantly more common in the experimental arm compared with the control arm, including AST elevation (17% vs. 6%), ALT elevation (21% vs. 4%), hyponatremia (13% vs. 7%), and TSH elevation (24% vs. 4%).
The high rate of adverse effects led to lower treatment intensities for both agents in the experimental arm, Siu reported. For example, just 57% of patients in the experimental arm received at least 90% of the planned dose intensity for cetuximab compared with 83% in the control arm. Forty-eight percent of patients in the experimental arm had 90% or more of the planned intensity of brivanib, compared with 87% of the patients receiving at least 90% of the planned intensity of the matched placebo.
When asked after her presentation why patients in the combination arm had such a low rate of cetuximab intensity in the experimental arm, Siu explained that if physicians could not distinguish immediately which agent was causing the problem, then “by protocol, we would have them reduce both drugs in those cases.”
Discussant: Toxicity Clinically Significant
During his discussion of the abstract, Herbert Hurwitz, MD, Associate Professor of Oncology at Duke University School of Medicine, noted that the adverse event rates in the trial were clinically significant. They led to dose reductions for both the experimental and standard agent and trial discontinuations — “thereby compromising the value of not only the experimental drug, but the standard drug as well,” he said.
Rushing to Phase III
Hurwitz also pointed out that the large phase III trial was launched based on “modest” phase Ib data, with no phase II study in between. While that unconventional approach was used successfully for the phase III trial of regorafenib, which was reported in the same session at the meeting, the brivanib-cetuximab trial illustrates the risks associated with rushing to phase III with limited early data.
Investigators had some experience with brivanib in other therapeutic settings and substantial experience in the clinic with cetuximab, yet the toxicities associated with the combination in this setting were substantial. If a phase II trial had been done, the investigators might have had a better understanding of those problems before launching the large, registration-style trial.
Finally, he said, the value of bFGF as a target in colorectal cancer might need to be reexamined after this trial. Brivanib is a dual-targeted inhibitor that blocks the FGF and VEGF pathways. However, the limited efficacy seen in this trial does not support evidence from pilot studies and preclinical data that suggested that bFGF is a valuable therapeutic target in this disease.
The study was supported by the National Cancer Institute of Canada. Siu and other study coauthors received research funding from Bristol-Myers Squibb, which makes brivanib. Additionally, one coauthor reported receiving honoraria or other funds from Bristol-Myers Squibb and another reports employment with the company. Hurwitz reports serving as a consultant or advisory role with Bristol-Myers Squibb.