We would have expected patients with high ERCC1 expression to do worse with FOLFOX. Should we be surprised? ERCC1 joins a long list of failed predictive tests for cytotoxic therapy, which had interesting initial results and then failure in larger studies.
Wells Messersmith, MD
Wells Messersmith, MD, Professor and Head of Medical Oncology and Director of the Gastrointestinal Cancer Program, University of Colorado, Denver, discussed the two studies.
Closer Look at STEAM
Dr. Messersmith said the FOLFOXIRI (fluorouracil [5-FU], leucovorin, oxaliplatin, irinotecan) vs FOLFOX (5-FU, leucovorin, oxaliplatin) comparison in the STEAM trial “basically confirmed” the results of the TRIBE study,1 showing higher response rates and longer progression-free survival with FOLFOXIRI (although the comparator arm was FOLFIRI [(5-FU, leucovorin, irinotecan)], not FOLFOX). In TRIBE, the overall survival endpoint “just missed statistical significance” for benefit with FOLFOXIRI/bevacizumab (31.0 months vs 25.8 months [hazard ratio = 0.79; P = .054]).1
The study demonstrated an improvement in median progression-free survival of about 2 months and response rates that were 15% to 20% higher than seen with FOLFIRI, “although the study did miss the target for statistical significance,” he noted. The liver resection rate was also improved.
“This comes at a price of somewhat higher toxicity,” he noted, adding that bevacizumab-related toxicities were independent of the backbone regimen. FOLFOXIRI is a “more difficult regimen” for patients who are working full time, whereas some other regimens allow them to continue to work without a problem, he said.
With this in mind, Dr. Messersmith said he thinks FOLFOXIRI/bevacizumab is a good first-line option for patients with a good performance status and for those in whom a response is valuable, for example, with borderline unresectable or symptomatic disease.
The study provided “insufficient evidence” for recommending sequential treatment or a strategy that alternates FOLFOX with FOLFIRI, he added. “Data are still maturing…. We await the overall survival data.”
Closer Look at MAVERICC
Commenting on MAVERICC, Dr. Messersmith noted, “FOLFIRI/bevacizumab nearly beat out FOLFOX/bevacizumab.” He indicated that the progression-free survival in MAVERICC was similar to that of other studies (10–12 months), but overall survival was worse than that of other studies (24–28 months).
The strongest trend in progression-free survival favoring FOLFIRI/bevacizumab was seen in patients with left-sided tumors (hazard ratio = 0.71), he pointed out. “It will be interesting to see if ‘sidedness’ is relevant, independent of other markers, in other contemporary trials,” he said.
Discussing the lack of ERCC1 as an outcomes predictor, Dr. Messersmith noted that this “surprising” result showed that ERCC1 was “irrelevant.” He added: “We would have expected patients with high ERCC1 expression to do worse with FOLFOX. Should we be surprised? ERCC1 joins a long list of failed predictive tests for cytotoxic therapy, which had interesting initial results and then failure in larger studies,” he said.
In addition to studies of ERCC1 and ERCC2, this “strong pattern” has been seen with thymidylate synthase and methylenetetrahydrofolate reductase for 5-FU efficacy, dihydropyrimidine dehydrogenase deficiency for 5-FU toxicity, and UDP-glucuronosyltransferase 1A1 for irinotecan toxicity.
“As opposed to RAS for EGFR [epidermal growth factor receptor]-targeting monoclonal antibodies and microsatellite instability for immunotherapy, tumor (and germline) predictors for cytotoxics have failed in colorectal cancer,” he observed. ■
Disclosure: Dr. Messerman reported no potential conflicts of interest.
1. Loupakis F, Cremolini C, Masi G, et al: Initial therapy with FOLFOXIRI and bevacizumab for metastatic colorectal cancer. N Engl J Med 371:1609-1618, 2014.
Two trials reported at the 2016 Gastrointestinal Cancers Symposium evaluated bevacizumab (Avastin)-containing regimens in the first-line treatment of metastatic colorectal cancer and supported some, but not all, previous findings.
The STEAM trial found some numerical differences but no...