Muddled Data Dents Medarex Shares

A study that fails its endpoint is never a good thing. But then, the statement, as written, seems designed to obfuscate. In English, Medarex is confessing that the observed response rate in the trial wasn't robust enough to rule out -- by statistical analysis -- a "true" response rate of less than 10%.

If that still has your head scratching, let me be even more clear: I'm betting that less than 10% of melanoma patients in this study reported meaningful reduction in the size of their tumors. Maybe the response rate edged up into the low-double digits (11% to 12%), but from a statistical perspective, that wasn't good enough to consider the study a success.

Going back to what I wrote in September, I find it hard to envision a scenario by which the FDA approves a melanoma drug that can't muster more than a 10% response rate in its key clinical trial. That's especially true in a situation where the study is conducted without the benefit of a comparator arm, as was the case here.

Medarex isn't providing any specific data on ipilimumab response rates. Instead, the company said response rates across all three studies ranged from mid-single digits to mid-teens.

In the past, Medarex executives insisted that the FDA has never placed a minimum response requirement on ipilimumab trials. Company supporters have pointed to this to support their belief that ipilumumab would win FDA approval even with less-than-robust efficacy -- especially since the melanoma patients in the studies had advanced disease that failed to respond to other treatments.