Data for Genasense did not show adequate effectiveness against malignant melanoma.

The US FDA advisory committee's rejection of Genta/Aventis's antisense treatment oblimersen sodium (Genasense) for malignant melanoma provides a lesson on the agency's expectations for survival data for regulatory approval.

The agency has stated that for some forms of cancer, the use of progression-free survival (PFS) is an appropriate endpoint. In these conditions, PFS can indicate attractive therapeutic options, and can provide an informative clinical marker in patients who are not expected to live for long.

But the advisory committee voted 13 to 3 that the PFS data that Genta provided for oblimersen could not be considered substantial evidence of effectiveness, because PFS was submitted as the secondary endpoint after the primary endpoint of overall survival failed.

Genta's submission data were based largely on an open-label 771-patient trial (GM-301) of oblimersen combined with dacarbazine. There was no statistically significant effect on overall survival. Patients on oblimersen/dacarbazine had a median survival of 274 days, compared with 238 days for dacarbazine alone (p = 0.18). Secondary endpoints did show a significant effect: PFS was 74 days for oblimersen/dacarbazine, compared with 49 days for dacarbazine (p = 0.003). Antitumour response rate was 11.7% for oblimersen/dacarbazine, compared with 6.8% dacarbazine (p = 0.019).

However, the FDA said that efficacy results based on these secondary endpoints were only to be considered supportive or exploratory if there was a statistically significant finding in overall survival. The agency cited Schering-Plough's application for approval of temozolomide (Temodar) for melanoma in 1999, which was not approved under similar criteria.

Another concern was the control of the trial conduct. In particular, the agency noted that the five-day difference in assessment times between the two treatment groups could potentially bias the estimation of treatment effect and could lead to a false positive inference in a large study.

“These were secondary outcomes that we are looking at, so that the way that one would rigorously define these and ascertain them is somewhat missing,” said committee voting consultant Ralph D'Agostino, Professor of Mathematics, Statistics and Public Health at Boston University. “I am stuck with the ... difficulty with progression-free survival and how it can move around depending on assumptions.”

The decision has proved costly for Genta. The company withdrew its New Drug Application (NDA) for oblimersen and has stopped marketing gallium nitrate (Ganite) for hypercalcaemia, which was intended to help establish its sales force before the launch of oblimersen. Genta says this will allow it to devote its remaining money — an estimated US $67 million — to re-submitting approval data for oblimersen.

Although the advisory committee rejected the treatment, it did hint that there might be an approvable effect. “There might be something here, but it just isn't clear,” noted committee member Stephen George, professor of biostatistics at Duke University Medical Center, Durham, North Carolina.