Replies to post #44286 on Biotech Values

[imnot6]

"The toxicity profile seen with panitumumab was similar to that seen with Erbitux, [the FDA's] Hughes said...


...., noting that "the apparent lower incidence of infusion-related reactions may relate to how data were collected in the two applications."

Ouch!

(no pun intended)

[caravon]

FDA approval process is terribly flawed

<<"The applicant has not provided sufficient data to evaluate the efficacy of panitumumab therapy," Division of Biologic Products Team Leader Kaushik Shastri wrote in a Sept. 26, 2006 memo, noting that Amgen did demonstrate improvement in PFS. Despite the modest effect size, FDA reviewers rationalized that for this patient population, there are no other treatment options, and the benefit-risk assessment is favorable under accelerated approval guidelines.>>
There is another treatment option. It is Erbitux being already approved by the same FDA group.



<<In his statistical review of overall survival, Lead Mathematical Statistician Mark Rothman disagreed with this analysis and the usefulness of using PFS as a predictor of overall survival . "Among the 232 patients on the BSC arm, 175 received panitumumab after investigator ascertainment of disease progression," he said. "For the patients in the study, overall survival was practically equal between the two arms. So not giving the patients on the BSC arm any anti-cancer therapy until they had an investigator ascertainment of disease progression or died, and then giving 175 of the survivors panitumumab led to the same overall survival as giving panitumumab upfront to the patients on the panitumumab arm. This does not say much for the meaningfulness of a PFS event in its relationship with overall survival."

It is obviously clear that, in their (FDA) desire to approve panitumumab, FDA leadership decided to disregard and ignore the statistical science embedded in the foundation of the approval process.

[gofishmarko]

On Vectibix , PCYC , DNDN and the FDA

The FDA deserves be on the hot seat right now as other companies ask for a second look at their data , since the lack of logic ( and integrity ? ) in the accelerated approval process is now too obvious to ignore.

This is what they said about Vectibix :

"Although the pivotal trial failed to show evidence of an impact on overall survival, FDA Division of Biologic Oncology Products Director Patricia Keegan noted in her recommendation to approve the biologic that "improvements in [progression-free survival] are generally accompanied by improvement in overall survival."

"Generally" , AKA : "sometimes". Not this time , though , in the case of Vectibix.

The purpose of Accelerated Approval is to get drugs on the market that are very likely to improve survival based on "surrogate data" , with the understanding that they will get full approval only if they subsequently demonstrate the survival benefit.

Now , suppose you had a bunch of NDAs or BLAs to review , with data packages similar to either :

1) The Vectibix data , on PFS.

or

2) The DNDN data , on actual survival.


Which group would have a better chance of eventually demonstrating improved survival ? I'd bet my money on the group "2" drugs , and I'd call the person crazy that said ONLY the group "1" drugs are eligible for Accelerated Approval , the situation that now pertains at the FDA.

Yeah , I hope there's some sweaty brows at the FDA right now as they try to decide what to do with DNDN. Maybe they'll decide that the best decision will be the one that makes the most sense.

In other words , time to start doing things differently.