FLEX: Erbitux Boosts Survival Five Weeks

Overall survival in the Erbitux-chemo patients was 11.3 months compared to 10.1 months for the chemotherapy-only patients, a difference of five weeks. The result was statistically significant with a p value of 0.044, which means there was a 4.4% chance the result of the study were random chance.

The generally accepted threshold for statistical significance in a clinical trial is a p value of 0.05, or 5%, so the overall survival benefit for Erbitux isn't very robust from a statistical standpoint.

Measured another way, the FLEX study's hazard ratio was 0.87, which means that patients treated with Erbitux had a 13% reduction in the relative risk of death compared to patients treated with chemo alone. That's a relatively modest benefit.

A simple way to think about this would be to compare the FLEX survival benefit statistics to a grade of C-minus or D -- still passing but just barely.

ImClone and partner Bristol-Myers Squibb (BMY Quote) previously announced plans to seek approval later this year from the U.S. Food and Drug Administration for Erbitux in lung cancer. The weak statistical power of the FLEX study does raise the risk that the FDA requests more data, perhaps even another clinical trial, before approving Erbitux.

ImClone's Rowinsky insists that Erbitux, right now, is an approvable lung cancer drug. "That p value will get us approved. The study was well designed and inclusive and we hit our primary endpoint."

Data on key secondary efficacy endpoints in the Erbitux FLEX study were also mixed. The response rate, or the percentage of patients whose tumors shrank significantly, was higher in the Erbitux-chemo patients (36.3%) compared to the chemo-only patients (29.2%). This was a statistically significant result.