Vatalanib in advanced colorectal cancer: two studies with identical results.

Because of biologic variability, limited impact of treatment, chance, and sometimes bias, the results of clinical trials in oncology are often divergent and occasionally fully contradictory. The two studies in advanced colorectal cancer on the antiangiogenic drug vatalanib (PTK/ZK) in firstand second-line treatment, called Colorectal Oral Novel Therapy for the Inhibition of Angiogenesis and Retarding of Metastases (CONFIRM) 1 and 2, respectively, are a striking exception because their results are truly identical. The studies showed the same hazard ratio (HR) for progressionfree survival (PFS; 0.88 and 0.83) and for overall survival (OS; 1.08 and 1.00), the same increased efficacy in the subgroup with high lactate dehydrogenase (HR for PFS 0.67 and 0.63) and the same lack of effect on the rate of objective responses (data not shown). The exposure to PTK/ZK or placebo and to the component drugs of infusional fluorouracil, leucovorin, and oxaliplatin (FOLFOX) were similarly reduced (10% to 20%) compared with the controls. The increased incidence of serious adverse events (SAEs) in the experimental arms over the control arms was identical (9% and 10%) as was the increased treatment discontinuation proportion due to SAEs (14% and 17%). Even the time to deterioration of performance status (HR 1.55 and 2.23), and the time to more than 5% weight loss (HR 1.91 and 1.83) were similarly decreased in the PTK/ZK arms. From every viewpoint the results are superimposable, to an extent that is not usually observed even when the same trial is replicated in the same target population. The implications of these unusual similarities are two-fold. First, the results of either study represent strong, independent confirmation by each other; as a consequence, bias and chance can be reasonably ruled out as possible explanations for these findings. Indeed, the results can be considered a reliable picture of the effects of PTK/ZK plus FOLFOX in advanced colorectal cancer. Second, notwithstanding the differences between the studies, their results can be discussed together. Taken together, the message is clear: PTK/ZK, when used in combination with FOLFOX in metastatic colorectal cancer, has but a marginal effect on PFS, and no detectable effect on OS. Considering the non-negligible added toxicity, these results do not support the clinical use of this drug. Even if one ignores OS and accepts the effect on PFS as the main criterion on which to base clinical decisions in advanced colorectal cancer, the size of the effect observed in both studies is too small to justify the use of PTK/ZK instead of standard regimens. In addition, just because of the strong internal validity of these trials, further studies on this drug do not seem warranted in this disease, at least in unselected populations of patients. Along this line, the other consistent message is that the effect of PTK/ZK appears to be stronger in the subgroup of patients with high lactate dehydrogenase at diagnosis (HR for PFS 0.67 and 0.63 in CONFIRM 1 and CONFIRM 2, respectively). However, the size of the benefit observed even in this subgroup of patients is not outstanding from a clinical perspective: the increase in median PFS was less than 2 months in both studies, tempering the enthusiasm to pursue this further. One wonders why the results have been so limited in the face of a strong theoretical rationale: complete blockade of the vascular endothelial growth factor receptor (VEGFR) pathways through inhibition of the three receptors should affect both angiogenesis and lymphangiogenesis. A number of reasons may be considered. First, the rationale itself. The relevance of triple VEGFR inhibition in colorectal cancer is unknown, as evidenced by the absence of strong data on experimental colon tumor systems. In addition, no formal phase II or randomized phase II trial suggesting substantial activity in colorectal cancer was done before these large phase III studies. Empiric evidence indicates that demonstration of activity in phase II trials is a prerequisite for success. Second, pharmacokinetic reasons. With a half-life of 4 to 6 hours, the once-daily schedule may have been inappropriate. Easy to speak after the facts: two observations, in fact, mitigate what now sounds like a mistake. First, an attempt was made to pursue the theoretically more appropriate twice-daily schedule, but the toxicity was higher. Second, with the once-daily regimen, chemotherapy had to be reduced by 15% to 20% in the CONFIRM studies. Thus, if the twice-daily schedule of PTK ZK was to be used, either its dose should have been cut drastically below that of biologic activity, or an unacceptably low chemotherapy dose intensity should have been used. Third, companion chemotherapy. It is possible that FOLFOX is not the ideal partner chemotherapy for this class of agents. The recently reported Horizon II study in first-line advanced colorectal cancer compared FOLFOX with or without cediranib, a multikinase VEGFR inhibitor, similar to PTK/ZK. The results mirror those of CONFIRM 1. Incidentally, some evidence is also emerging JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L S