The Oncolytics Biotech (ONCY) study of Reolysin in head and neck cancer patients failed. The company issued a press release Thursday claiming the Reolysin study results were positive, but CEO Brad Thompson is shading the truth.
It's easy to be confused because Oncolytics throws out a bunch of numbers in this morning's announcement to make Reolysin appear effective. It's smoke and mirrors, so let's cut through the B.S.
For starters, a brief recap of the study design. Oncolytics enrolled 167 patients with advanced head and neck cancer and randomized them to receive either 1) Reolysin plus carboplatin and paclitaxel or 2) placebo plus carboplatin and paclitaxel. The primary endpoint of the study was overall survival, with progression-free survival (PFS) and overall response rate important secondary endpoints. A description of the study can be found at ClinicalTrials.gov.
This morning, Oncolytics reported results from an "intent-to-treat" analysis of 118 patients with "loco-regional head and neck cancer with or without distant metastases."
The first red flag: An "intent-to-treat" analysis cannot be conducted on a subgroup of patients. Oncolytics is omitting from its analysis 49 head and neck cancer patients with distant metastatic disease.
Why is Oncolytics dividing up patients in the study? Last year, the company claimed patients with localized disease (tumors primarily restricted to the head and neck) were responding differently to Reolysin compared to patients with cancer that had spread to distant parts of the body. How Oncolytics came to this conclusion while remaining blinded to study data has never been explained, but nonetheless, the company insisted on splitting the study into two parts, with loco-regional and distant metastatic patients analyzed separately.
On Thursday, Oncolytics reported overall survival results on 88 loco-regional patients who "did not receive additional therapy following discontinuation of study treatment." The 50 patients in the Reolysin arm reported a median overall survival of 150 days compared to 115 days in the control arm.
The second red flag: Oncolytics was supposed to report overall survival results on all 118 loco-regional patients. Instead, the company gives us results on 88 patients who didn't receive any other therapy after leaving the study. Thirty patients were disappeared from the overall survival analysis.
The third red flag: The subgroup analysis of overall survival in the 88 patients, which isn't even credible, wasn't statistically significant.
Oncolytics CEO Thompson, on a conference call Thursday morning, was asked to provide the real overall survival results from all 118 loco-regional patients. He refused. Red flag number four.
What about the overall survival analysis of the 49 head and neck cancer patients with metastatic disease? Metastatic disease is more advanced, so these patients should actually have shorter survival times than patients with localized disease.
Oncolytics says it remains blinded to the data on the metastatic patients. Call that red flag number five.
On this morning's call, Thompson made the argument that progression-free survival -- and not necessarily overall survival -- was the most important efficacy measure from the Reolysin study.
But that's not what he said in May when speaking to investors at a healthcare conference sponsored by Merrill Lynch:
The primary endpoint was and is remains overall survival, and that's event-driven. And we're currently awaiting events to get to the predicted event range to actually do the survival data. And the secondary endpoints, which we will also report on, are PFS, looking at best response, which is, in this case, but actually by RECIST, so there's really a true response rate....
That's red flag number six.
Now, about those PFS data. Oncolytics reported Thursday a median PFS of 94 days for the Reolysin patients compared to 50 days for control patients. Again, these results were just from the 118 head and neck cancer patients with loco-regional disease. The company once again "disappeared" data on the 49 metastatic disease patients.
Was this PFS result statistically significant? When asked, Thompson claimed use of "non proportional hazard ratios" to analyze results made it impossible to provide p values.
Translation: No, the PFS data were not statistically significant. Where are we now on the red flag count? Seven and eight.
I won't delve into the issues with the safety data in the Reolysin study except to point out that Oncolytics claims to have included all 167 head and neck cancer patients in the safety analysis.
If that's true, how is it possible for Oncolytics to remain blinded to efficacy results in the 49 patients with metastatic disease?
As you'd expect, Oncolytics has no answer to the question.
Red flag number nine.
-- Reported by Adam Feuerstein in Boston.
Follow Adam Feuerstein on Twitter.