Every drugmaker wants to tell you its latest data changes things. The harder trick is convincing the people who actually decide: the regulators who sign off and the analysts who price the stock.
Two of the largest cancer franchises in the world are running into that wall right now. AstraZeneca has a breast cancer dataset its own FDA panel already waved off. Pfizer has a lung cancer drug in a field so crowded the experts can no longer tell the contenders apart.
Good data, it turns out, is the easy part.
AstraZeneca's FDA Panel Said No. It's Taking The Case Abroad.
AstraZeneca's confidence in camizestrant has been boosted by the SERENA-6 results, along with the unmet medical need, after the FDA's Oncologic Drugs Advisory Committee voted 6-3 on April 30 that the trial did not demonstrate clinically meaningful benefit. The SERD is in patients with ESR1-mutated, HR+/HER2- metastatic breast cancer.
WHAT HAPPENED
The data from earlier this month showed that camizestrant plus investigator's choice of a CDK4/6 inhibitor maintained PFS with longer follow-up and showed statistically significant and clinically meaningful improvement in second PFS (PFS2) versus a CDK4/6 inhibitor combined with an aromatase inhibitor, while exploratory analyses showed reduction in circulating tumor DNA.
"We've been engaging the regulatory agencies globally about these results, and look - these are the two elements that we are fundamentally using in the discussion with the agency: very strong dataset and high unmet medical need in this patient population," Gianluca Pirozzi, head of development, regulatory and safety for Alexion, had said.
THE CHATTER
The 6-3 vote split the room on a question becoming central to ctDNA-driven design: whether intervening on a molecular signal before a patient has clinically progressed delivers benefit a patient can feel, or simply moves a curve. AstraZeneca is betting agencies outside the US weigh that differently than the panel did. "Very strong dataset and high unmet medical need" is the language of a company arguing around an endpoint, not from it.