ASCO: Cetuximab Ups Survival Over Bevacizumab in Colorectal Cancer

Results of this trial show that the addition of cetuximab (Erbitux) increased overall survival in some metastatic colorectal cancer.

Results from the German AIO (Arbeitsgemeinschaft Internistische Onkologie) KRK-0306 study, FIRE-3, show that the addition of cetuximab (Erbitux) to chemotherapy rather than bevacizumab (Avastin) increased overall survival by nearly 4 months in patients with KRAS wild-type metastatic colorectal cancer. At the 2013 American Society of Clinical Oncology (ASCO) Annual Meeting, Volker Heinemann, MD, PhD, from the department of medical oncology and the Comprehensive Cancer Center at Ludwig-Maximilian-University in Munich, and Sebastian Stintzing, MD, a postdoctoral fellow of the German Cancer Aid at the University of Southern California Norris Comprehensive Cancer Center in Los Angeles, discussed the findings and their implications.

“Colon cancer is the second most frequent cancer in the Western world and is recognized as the most frequent cause of cancer death; 40% to 50% of colon cancer patients develop metastatic disease, to which most of these patients will finally succumb,” Dr. Heinemann commented in a press conference held during ASCO. “Much effort is presently undertaken to develop more effective treatment strategies,” he explained, in describing the FIRE-3 trial. “One already successful approach is to combine chemotherapy with targeted agents inhibiting either cell proliferation or angiogenesis. So far, it is unclear which targeted agent, cetuximab or bevacizumab, should be used preferentially in first-line treatment. This is specifically important, since both drugs markedly differ, not only with regard to their mechanism of action but also with regard to their profile of side effects.”

FIRE-3 Study Design

FIRE-3 is the first head-to-head trial directly comparing FOLFIRI plus cetuximab vs FOLFIRI plus bevacizumab in patients with KRAS wild-type metastatic colorectal cancer, to determine which targeted therapy is optimal as part of the first-line chemotherapy regimen. This randomized phase III study involved patients from 150 cancer centers in Germany and Austria. It compared FOLFIRI plus the anti-EGFR (epidermal growth factor receptor) agent cetuximab vs FOLFIRI plus the anti-VEGF (vascular endothelial growth factor) agent bevacizumab in first-line treatment of metastatic colorectal cancer. A total of 752 patients were enrolled, and of this group, 592 patients with confirmed KRAS wild-type tumors were included in the study and defined the intent to treat population.

The 592 patients were randomized 1:1 to FOLFIRI (the Tournigand regimen) every 2 weeks with either cetuximab (400 mg/m² on day 1, followed by 250 mg/m² weekly) or with bevacizumab (at 5 mg/kg every 2 weeks). There were 297 patients randomized to the FOLFIRI plus cetuximab arm and 295 randomized to FOLFIRI plus bevacizumab. While the median number of months of therapy was comparable between the two arms (4.8 months in the FOLFIRI plus cetuximab arm vs 5.3 months in the FOLFIRI plus bevacizumab arm; P = .112), the difference between the median number of cycles was higher in patients who received bevacizumab with their chemotherapy (a median of 12 cycles with bevacizumab vs 10 cycles with cetuximab, respectively), and this was statistically significant (P = .014).

A total of 526 patients were accessible for response evaluation, since they had undergone at least 3 cycles of treatment and had at least one post-baseline CT scan.

Until September 2008, patients had initially been recruited into FIRE-3 regardless of KRAS mutation status, and “standard inclusion and exclusion criteria were applied,” Dr. Heinemann said. When it became known that cetuximab was active only in patients with KRAS wild-type tumors, an amendment to the study was performed in October 2008 that confined the population to KRAS wild-type patients. “This population represents a majority, approximately 60% of metastatic colorectal cancer patients,” he said. Baseline characteristics were comparable between the study arms. Key inclusion criteria were age 18 years or older with a confirmed diagnosis of metastatic colorectal cancer, ECOG PS (Eastern Cooperative Oncology Group performance status) of 0–2, and allowance for prior adjuvant therapy as long as it was completed more than 6 months before inclusion in the study. Median follow-up in the trial was 33 months and 39 months in the FOLFIRI plus cetuximab and FOLFIRI plus bevacizumab arms, respectively.

The majority of patients in the trial, 98%, had an ECOG PS of 0 or 1. Median age was 64 to 65 years, and the majority in each group was male (about 72% in the FOLFIRI plus cetuximab group and about 66% in the FOLFIRI plus bevacizumab group). Patients in the two trial arms were evenly matched with respect to primary tumor; about 60% had colon cancer and close to 40% had cancer of the rectum (and 3% to 4% of patients had both). About 40% in each group had one site of metastatic disease and about 60% in each group had two or more metastatic sites. About 31% in each group had liver metastasis only. The groups were also well-matched for the percentage in each study arm whose prior treatment included surgery (83% to 85%), adjuvant chemotherapy (about 20% each), and radiotherapy (about 13% each).

The Study Outcomes

Tumor response rates

The trial’s primary endpoint was overall response rate (ORR, as measured by modified RECIST 1.0 criteria). In the intent to treat population of 592 patients, there was a trend in favor of FOLFIRI with cetuximab (62% ORR vs 58% ORR with FOLFIRI plus bevacizumab), but it was not statistically significant (P = .183; odds ratio, 1.249). “However, in [the 526] assessable [for efficacy] patients, where the true drug effect can be evaluated, a significant difference of 9% was observed in favor of the cetuximab arm, which we believe also relates to the survival benefit,” Dr. Heinemann commented. In the assessable for response population, the ORR was about 72% in the cetuximab arm vs about 63% in the bevacizumab arm (P = .017).

In the evaluation of response by RECIST criteria, complete response was higher with cetuximab (at 4.4% vs 1.4% with bevacizumab, a statistically significant difference). Partial response rates were comparable (about 57% in both groups), as was the incidence of progressive disease (5% to 7%, not statistically significant). The incidence of stable disease, however, was higher in the group randomized to chemotherapy with bevacizumab (28.8% vs 17.5% among patients randomized to chemotherapy plus cetuximab, a statistically significant difference).

Survival outcomes

Regarding the secondary endpoint, overall survival, “the key finding of the trial is that a significantly longer survival was obtained in patients receiving FOLFIRI plus cetuximab in comparison to FOLFIRI plus bevacizumab,” Dr. Heinemann said. “Median survival in the cetuximab arm was 3.7 months longer than in the comparator arm [P = .017]. This was accompanied by a hazard ratio of 0.77, which means that patients in the cetuximab arm had about a 23% lower risk of death during the observation period.” The survival curves in the two study arms began to separate about 18 months after the start of treatment. An exploratory subgroup analysis for overall survival showed FOLFIRI plus cetuximab was favored regardless of gender, age, location of tumor (colon vs rectum), number of metastatic sites, presence of liver-limited disease, presence of synchronous metastases, and leukocyte levels (above or below 8/nL).

Median progression-free survival was nearly identical in both study arms, at 10 months in the FOLFIRI plus cetuximab arm and 10.3 months in the FOLFIRI plus bevacizumab arm (P = .547).

Assessment of subsequent second-line therapy received by patients on the trial showed comparability between the two arms of the trial. Close to two-thirds of patients on both arms received “any second-line therapy,” 48.2% in the FOLFIRI plus cetuximab arm received second-line bevacizumab, and 42.9% of patients in the FOLFIRI plus bevacizumab arm received second-line anti-EGFR therapy.

Adverse Events

The FIRE-3 investigators reported that toxicity profiles were as expected and manageable for both study regimens. There was no statistically significant difference in any grade or grade 3 or higher hematologic toxicities between the FOLFIRI plus bevacizumab vs FOLFIRI plus cetuximab arms, and this finding is as might be expected given that the chemotherapy regimen in both arms was identical. Regarding non-hematologic toxicities, “any grade” nausea and vomiting were higher in the bevacizumab arm (about 62% and 33%, respectively) than in the cetuximab arm (about 48% and 25%, respectively), and the incidence of “any grade” hand-foot syndrome was higher in the cetuximab arm (about 27% vs 14%, respectively). Skin toxicities (eg, acneiform rash [incidence of about 17% grade ≥ 3], desquamation, and paronychia) in the cetuximab arm were within the expected range. Unsurprisingly, known toxicities of bevacizumab were more frequent in the FOLFIRI plus bevacizumab arm than in the FOLFIRI plus cetuximab arm: incidence of “any grade” hypertension was about 38% vs about 21%, respectively (though grade ≥ 3 hypertension was comparable at a little over 6% for both groups); incidence of “any grade” bleeding or hemorrhage was about 28% vs 21%, respectively; and incidence of “any grade” abscesses or fistulas was 5.4% vs 1.4%, respectively (all statistically significant differences for these “any grade” adverse events). Notably, the investigators said, the incidence of thrombosis and thrombotic events were comparable between the two treatment arms (6.1% grade ≥ 3 thrombosis for both arms; 5.8% vs 5.1% incidence of thrombotic events, respectively).

Conclusion and Future Assessments

In conclusion, Dr. Heinemann noted that “first-line treatment with FOLFIRI plus cetuximab resulted in a clinically meaningful difference in median [overall survival] of 3.7 months [HR = 0.77] when compared to FOLFIRI plus bevacizumab. Toxicity profiles were manageable and as expected for both combinations. Based on our findings, we believe that substantial gains in survival can be achieved when doctors offer cetuximab as first-line treatment to their patients with KRAS wild-type metastatic colorectal cancer.”

In the oral abstract session during ASCO, Dr. Heinemann’s mentee, Dr. Sebastian Stintzing, presented and discussed the FIRE-3 results. “How can we explain that there’s no difference in progression-free survival but a significant difference in overall survival, even though further-line treatments, to our current knowledge, did not differ significantly? There has been data presented at ASCO this year of Mansmann et al [ASCO 2013 abstract 3630 and ASCO GI abstract 427], analyzing the CRYSTAL trial, that indicates that tumor size reduction is more predictive for overall survival than [progression-free survival], and…greater tumor volume reduction can translate into a significant longer overall survival. We are now in the process of a central independent review of the FIRE-3 CT scans, to analyze tumor volume changes, which is also a secondary endpoint of the FIRE-3 study.”

Data analyzing the possible impact of the second- and further-line therapies with cetuximab or bevacizumab on longer-term outcomes will be presented at the European Society for Medical Oncology (ESMO) 15th World Congress on Gastrointestinal Cancer in Barcelona this summer.

Data on time to treatment failure for patients on both study arms will be calculated and presented at a meeting later this year.

While KRAS wild-type mutations refer to codons 12 and 13, Jean-Yves Douillard, MD, PhD, professor of medical oncology, University of Nantes and Integrated Centers of Oncology ICO Rene Gauducheau Cancer Nantes in France, noted in the Q&A period of the presentation, “we know that other mutations that may account for maybe 15% to 20% [of colorectal cancer cases] may exist, and I hope you have material to look deeper into Ras-NRAS and KRAS-because this will increase the magnitude of the difference between the two arms.” Dr. Stintzing replied, “that’s true, thank you for mentioning this. We have about 90% of tumor samples of this trial [and] we are right now in the process of doing the NRAS, KRAS, and other Ras [family genes] mutational analysis, and we think the data should be done by this year and should be presented this year.”

Asked about the secondary resection rate on the trial, he said that secondary resection with curative intent was comparable between the study arms, and was performed for about 11% to 12% of patients in both study arms.

In response to a question about an assessment of maintenance therapy for patients in both arms of the trial, Dr. Stintzing replied, “until now, we do not have any standard for maintenance therapy. We would at least treat the patients until the smallest tumor volume is achieved, so, until the overall response is achieved. Asking what are we doing if the patients actually got the deepness of response, the tumor nadir-are we going to go on with our treatment or are we going to do any kind of maintenance therapy-is an interesting question that has to be addressed in future trials.”