Tuesday, June 09, 2009

How Close Avastin Really Came To Adjuvant Colorectal Cancer Use

Six events. Six additional cases of recurrence out of the 2,710 patients being treated in the C-08 trial of Roche/Genentech’s Avastin in adjuvant colorectal cancer, and there would have been an entirely different outcome.

Six more cases at the interim look and everyone would be using Avastin in adjuvant colorectal cancer patients. That’s how close the study was to meeting the early-stopping rule at one year, lead investigator Carmen Allegra said at Roche/Genentech’s on-site ASCO analyst event.

To be sure we hit this home hard enough: if just an additional six events had occurred at the one-year interim look, than the trial would have been stopped early. As the full analysis of data presented at ASCO made clear, at one year (not coincidentally, the time period that patients received bevacizumab), there was a significant benefit for the drug. A 40% advantage, to be precise. Lots of zeroes in the p-value to make the statisticians happy. More than enough benefit to drive utilization even before FDA approval.

Contrast that to the actual end of the study. At the pre-specified three-year endpoint, when there were 603 events, disease-free survival had dropped out of significance. The end result for Avastin was 77.4% versus 75.5% for chemo alone. But the eventual failure is old news, previewed in April and heard round the world.

The study investigators, and Roche/Genentech executives, were very keen on the one year results in unveiling the full dataset from C-08, stressing that the drug was extremely effective while being given and it was only after bevacizumab was stopped that the benefit diminished. Their take-away was that more study was needed with a longer duration of treatment with bevacizumab.

That’s not to say there wouldn’t have been valuable and appropriate questions raised about long-term tolerability and about the level of benefit versus alternatives and about the cost-effectiveness. However, it’s awfully close to taking what was ultimately a negative trial and finding it instead to be a runaway success based on an interim peek.

What the experience does show is exactly why long-term follow-up is important (both to see whether there’s a spike in hypertension, or hey, that benefit seems to disappear pretty quickly). It also underscores the value of designing a trial for a more clear-cut look at overall survival, which wouldn’t leave us with the potential variability from the disease-free surrogate. It’s also much harder to make a cost-effectiveness argument when you have data in hand to show that lives are saved.

For more on the implications of the C-08 data release, including the potential to establish Avastin as adjuvant therapy in other cancers (and possibly still colorectal), and the potential need for caution given tolerability and cost concerns, check out this week’s “The Pink Sheet” here and here.--Mary Jo Laffler

image from flickr user Shovelling Son used under a creative commons license


Anonymous said...

It's actually colon cancer.

Anonymous said...