Clovis Oncology has decided to halt development of its investigational agent CO-101 as an alternative for pancreatic cancer patients who are poor responders to the standard chemotherapy agent gemcitabine after a clinical trial failed to show any improvement in survival time in a molecularly defined subset of patients.
The trial results also disproved previous retrospective studies that indicated a link between hENT1 expression and survival for patients on gemcitabine. While previous studies suggested that patients with low hENT1 expression responded poorly to the chemotherapeutic — thereby indicating a potential treatment population for CO-101 — the study found that hENT1 status had no impact on survival for patients on gemcitabine.
In the prospectively designed trial, called LEAP, for Low hENT1 and Adenocarcinoma of the Pancreas, metastatic pancreatic cancer patients with low hENT1 expression showed no difference in survival when treated with gemcitabine or CO-101. There was also no survival difference between the treatment arms in the overall population.
Clovis reported that median survival for patients on either treatment was approximately six months. Drug-related toxicities were similar for each group.
As recently as last month, the company anticipated commercializing the drug in early 2013, so the results of the trial took officials by surprise (PGx Reporter 10/10/2012).
"What we want in this industry, and what we want in this company, is a clear answer and to design a trial that gives us an unambiguous result. To our astonishment, this was even more unambiguous than we could have imagined. These drugs behaved exactly the same," Patrick Mahaffy, CEO of Clovis Oncology, said during a call with analysts this week. "I am, at least as a developer and as a human, grateful that our drug did no harm. It was as good as gemcitabine in the patient populations it addressed."
The median six-month survival seen in the study is consistent with the outcomes seen in patients enrolled in past gemcitabine trials. Clovis added that prognostic variables, such as performance status and age, were balanced between the hENT1-low and -high groups.
"[CO-101] showed no evidence, none, of an improved response rate, of an improved progression-free survival, or an improved outcome in subset analysis," Mahaffy said. "Probably, if we wanted to, we could data dredge and tell you a story about [if we] look at this group or that group we might have seen something. It's not helpful. This drug was just not better than gemcitabine."
Beyond this stark acknowledgement of events, the company did not provide any mechanistic explanations for why CO-101 did not perform better than gemcitabine in hENT1-low patients. Gemcitabine, currently the first-line standard chemotherapy treatment for metastatic pancreatic cancer patients, requires a transport mechanism, such as the protein cellular transporter hENT1, to help it enter tumor cells. Previously published retrospective studies suggested that patients with high hENT1 expression respond well to gemcitabine, while those with low expression — about two-thirds of pancreatic cancer patients — respond poorly to the chemotherapeutic.
Researchers at Clovis had hypothesized that unlike gemcitabine, CO-101, a gemcitabine-lipid conjugate, would enter tumor cells through passive diffusion and as such the drug's efficacy wouldn't be hindered by hENT1 expression.
In LEAP, Clovis was using a companion test to hone in on whether hENT1-low patients would respond better to CO-101 than to gemcitabine. If this proved to be the case, the company was hoping to make a strong cost-effectiveness case to help garner reimbursement of CO-101 for that population.
However, "the hENT1 hypothesis in this prospective study appeared to be absolutely irrelevant to outcomes," Mahaffy said. "At the very least, we had hoped to prove the hENT1 hypothesis to better direct gemcitabine to the patients who would benefit. Sadly that is not the case."
The development program for the companion diagnostic, being developed by Roche subsidiary Ventana Medical Systems, will most likely also be halted.
"We are completely perplexed by the evident failure of hENT1 status to be relevant to gemcitabine survival," Mahaffy said. "To our surprise and contrary to the results of numerous retrospective studies, there is no difference in survival observed between the gemcitabine hENT1 high and hENT1 low patients, suggesting that the hENT1 biomarker is not predictive for gemcitabine outcomes."
Clovis projected that it would end the year with $140 million in cash, which would be sufficient to fund Phase I studies for the other products in its pipeline, mainly CO-1686 and rucaparib.
CO-1686 is a selective covalent EGFR inhibitor that the firm is exploring in patients with non-small cell lung cancer. Currently Clovis is conducting a dose-finding Phase I/II trial involving CO-1686 in NSCLC patients with T790M mutations. Patients with these "gatekeeper" mutations become resistant to treatment to widely prescribed EGFR-inhibiting NSCLC drugs, Roche/Genentech's Tarceva and AstraZeneca's Iressa.
Rucparib, meantime, is an inhibitor of PARP 1 and PARP 2 that the company licensed from Pfizer. Rucaparib is currently undergoing Phase I/II trials in breast and ovarian cancer. Clovis is investigating the efficacy and safety of the drug in patients with germline BRCA mutations, who lack the ability to repair damaged DNA that cancer cells need to thrive.
Clovis is planning to present data from early-stage studies involving these two drugs at a major oncology conference next year.