Health News Review

The latest example is news out of the European Cancer Congress (ECC) in Amsterdam.

Roche immunotherapy drug may be ‘game changer’ in lung cancer – Reuters reported.  Excerpt: 

Of 53 patients with NSCLC tumours treated with the drug, 23 percent saw their tumours shrink, according to results presented at the European Cancer Congress (ECC) in Amsterdam.

But the most encouraging numbers were among smokers, where the response rate was 26 percent compared with 10 percent of patients who had never smoked, said Professor Jean-Charles Soria of France’s Institut Gustave Roussy, who led the study.

So they drilled down into the data more closely, separating out the 81 percent of the 53 patients who were smokers or former smokers, and the 19 percent of them who were not.

“And bingo, this is the first targeted agent (drug) that shows more activity in smokers than in non-smokers,” Soria told reporters in a briefing at the ECC.

81 percent of 53 patients?  Why this awkward explanation?  Why not just do the math for readers?

Yes, this is a difficult cancer to treat.

But it’s not a game.

And even sportswriters would provide more about the numbers than this story gave.  Missing was any information about the degree of the response rate or tumor shrinkage.  And not a word about the period of followup, leaving one to wonder whether the tumor shrinkage that was observed might be a surrogate or intermediate endpoint that doesn’t equate to a truly meaningful clinical outcome such as living better or longer.

But we can’t know that from the sparse details provided.

26% response in 81% of 53 patients who were smokers or former smokers = some tumor shrinkage in 11 people.

Maybe we should keep the gamechanger pronouncement on the sideline until we get more data or more longterm experience in more people.


Follow us on Twitter:

and on Facebook.



Greg Pawelski posted on October 1, 2013 at 9:42 am

Virtually every medical oncologist knows the drill. The result: the average patient has an average outcome with the average treatment. That is by no means a “gamechanger.” While clinical trials are designed to identify average improvements for average patients, virtually every trial conducted has patients who live much longer than average. They constitute the tail on the survival curve (the outliers) and almost every trial has several. The job of medical oncologists is to identify those true responders and treat them appropriately. The term applied for these failures in the average patient paradigm are “beta errors,” meaning that the investigators missed the benefit of a given treatment.