It’s a tie: CALGB 80405 does not show efficacy difference for first-line Erbitux or Avastin in mCRC patients
| Jun 2, 2014
By Arnold DuBell and Stephanie Hawthorne
Both Avastin® (bevacizumab, Genentech/Roche/Chugai) and Erbitux® (cetuximab, Bristol -Myers Squibb/Eli Lilly/Merck KGaA) are approved for the treatment of first-line metastatic colorectal cancer (mCRC) and have demonstrated progression-free survival (PFS) and overall survival (OS) benefits when added to standard chemotherapy in these patients. Avastin is approved for use in all first-line patients, while Erbitux is approved for use only in the subgroup of patients with wildtype KRAS. As with all therapies that exist in the same indication, physicians are left wondering how best to incorporate these agents into their practices. As multiple Phase III trials showed combining these two agents are not effective at best (and harmful in the extreme), two trials set about to understand whether one regimen is superior as an initial treatment option.
At the 2013 ASCO annual meeting, the data from one of these trials, FIRE-3, were presented. FIRE-3 was a German trial designed to evaluate FOLFIRI plus either Erbitux or Avastin (at 5 mg/kg every two weeks) as first-line therapy; it was amended in 2008 to include only KRAS wildtype patients. The data suggested a non-significant benefit compared to Avastin for overall response rate, which was the primary endpoint (62% versus 58%, p=0.183). Moreover, a PFS benefit was not observed (HR 1.06, p=1.06); however, Erbitux was associated with a significant 3.7-month OS benefit compared to Avastin (HR 0.77, p=0.017). The Kaplan-Meier OS curves were especially intriguing, as the curves for each arm began to separate between 18 and 24 months, well after the 10-month PFS median, and the curves continued to separate past the median OS, where the superiority of Erbitux becomes far more pronounced and is evidently durable.
The conflicting data found in FIRE-3 begged for a repeat, and CALGB 80405, which was presented in the ASCO 2014 plenary session, has been highly anticipated for this reason. CALGB 80405 evaluated the addition of Erbitux or Avastin to physician’s choice of first-line chemotherapy (FOLFIRI or FOLFOX). When initiated in 2004, the trial enrolled all patients, as well as recruiting patients onto a third arm of Avastin plus Erbitux and chemotherapy. The trial was amended in 2009 to include only KRAS wildtype patients and halt recruitment to the arm with both Erbitux and Avastin. Ultimately, the trial enrolled 1,137 patients, and the primary endpoint was OS.
CALGB 80405 failed to show that either Erbitux or Avastin was associated with improved OS, as the Kaplan-Meier curves were essentially superimposable (median OS: 29.9 months for Erbitux versus 29.0 months for Avastin, HR 0.925, p=0.34). The curves for PFS were also essentially superimposable (median PFS: 10.8 months versus 10.4 months, HR 1.04, p=0.55). Subgroup analyses of patients treated with FOLFOX, which accounted for 73% of the patients enrolled in the trial, also showed no benefit for either drug (HR 0.9, p=0.09). Other data, including response rates, still need to be collected.
The “take-home” message is that efficacy differences are, for now, unlikely to drive the decision when choosing between these two agents for first-line mCRC therapy. Perhaps differences will emerge when evaluating secondary endpoints such as complete response rates, broader RAS mutant populations, or ability to resect patients to a disease-free state. For now, though, the choice of first-line therapy for mCRC patients is likely to be driven largely by the patient’s preference regarding potential side effects. The toxicity profiles were as expected for these agents, with the incidence of Grade 3-4 rash (7% versus 0%) and diarrhea (11% versus 8%) increased in the Erbitux arm, and the incidence of Grade 3-4 hypertension (1% versus 7%) and gastrointestinal events (0.5% versus 2%) increased in the Avastin arm. Updated DSQL skin satisfaction patient-reported survey data confirmed that skin rash was a significant concern for these patients (p<0.0001); however, this concern did not get reflected into results for the EORTC Global quality of life survey data (p=0.0546). The presenter, Dr. Venook, stated his belief that this may be because patients who get a rash feel “that the drug is working.”
In the absence of a confirmed efficacy benefit, cost could be another differentiator. In the post-plenary session, Dr. Saltz noted that the cost of Avastin per month is about half the cost of Erbitux per month (he referred to prices of $5,000 per month for Avastin and $10,000 per month for Erbitux).
Dr. Venook did not make an attempt to explain the discrepancy between the CALGB study and FIRE-3. The discussant (Dr. Tabernero) also did not do attempt to explain this. However, he did present updated FIRE-3 data, that indicated that use of an expanded RAS population (e.g., all RAS proteins are wildtype, not just the KRAS exon 2 as originally indicated) resulted in an improvement in the hazard ratio for both PFS (from 1.06 to 0.93) and OS (from 0.77 to 0.7). Dr. Tabernero suggested that this might result in CALGB 80405 finding a patient population in which Erbitux will show an OS advantage. Kantar Health is more skeptical of this claim, since it would be unlikely to shift the hazard ratios enough to be significant.
In a sense, CALGB 80405 provided a more definitive result than FIRE-3. The results lend themselves to a continuation of the current treatment plan. In the United States, Avastin is already outpacing Erbitux, being used in the majority of first-line KRASwildtype patients, while in Europe Erbitux has a slight lead over Avastin. Given the clear results of CALGB 80405 and the apparently conflicted results of FIRE-3, global treatment patterns are unlikely to change: Tie goes to the runner!
 Stintzing, LBA3506, ASCO 2013
 Venook, LBA3, ASCO 2014
 Kantar Health, CancerMPact® Treatment Architecture United States and Western Europe, accessed June 1, 2014.