Cancer meeting reports often not the final word

NEW YORK (Reuters Health) - Preliminary results of cancer treatment studies presented at major cancer meetings should be viewed with caution, because the final results may wind up being quite different, new research shows.

Dr. Christopher M. Booth of Queen's University in Kingston, Ontario, and colleagues found that nearly two-thirds of abstracts presented at cancer meetings had small differences in numbers of patients, follow-up times, or results when compared with final reports of randomized clinical trials published in medical journals.

And for 10% of the presented abstracts they reviewed, researchers came to the opposite conclusion when the trial was actually published. This means study authors may have initially concluded that a treatment should be adopted, but switched to not recommending a therapy in a published report, or vice versa.

Most of the time, Booth told Reuters Health, doctors don't change how they treat patients based on results presented at meetings. Nevertheless, "if you go to a conference and see an impressive presentation for a new therapy, it certainly is tempting to move those therapies into the clinic soon," he added.

Before a randomized controlled trial (RCT) gets published in a medical journal, it undergoes a process called peer review in which independent experts examine the findings. But the process for review of abstracts presented at medical meetings is different, he and his colleagues note in the Journal of Clinical Oncology, and such abstracts may also report preliminary results of a trial still in progress.

The researchers tested their hypothesis that preliminary results presented at meetings might be different from final published findings by tracing back conference abstract presentations for 138 RCTs published in six major journals from 2000-2004.

They looked at therapies for lymphoma, breast cancer, colorectal cancer or non-small-cell lung cancer.

Booth and his team found 303 abstracts related to the RCTs, 197 of which were eligible for their analysis. While 41 of the abstracts stated that the results were based on non-final analysis of the findings, 45 just implied that the results were not final but didn't explicitly state this. Even in the 12 abstracts reporting on studies for which patients were still being recruited, just 4 stated that the analysis wasn't final.

Presenting findings while recruitment is still underway, Booth and his team note, "has been referred to as 'inappropriate and undesirable,'" both because it can make it more difficult to recruit additional patients, and it may affect whether or not patients currently in the study will stick to their treatment.

The final results were different from preliminary findings for 124 of the abstracts, or 63%, and this "discordance" was more common when the early findings were based on a non-final analysis (78% of non-final analyses vs. 51% of final analyses were discordant).

In one out of 10 of the abstracts, the conclusions of the authors in the final published studies were "substantively different" than for the preliminary findings, the researchers found.

These findings show that doctors and patients need to be aware when findings are preliminary, and that researchers presenting findings based on non-final analyses should clearly state that this is the case, Booth said.

But that's not to say that the way cancer meetings are run needs to change radically, he added; most of the time, when real "blockbuster" findings get presented at meetings, they are based on final analyses and published in the peer reviewed literature shortly thereafter, Booth said.

Nevertheless, he added, given the heavy media attention now given to research on cancer treatments, it's important for both patients and doctors to be aware of when a publicized finding is the final word -- and when it isn't.

"One of the challenges for patients and clinicians is being able to stay on top of all of these advances," Booth added. "Things are moving so quickly, it really is a moving target."

Copyright © 2009 Reuters Limited. All rights reserved. Republication or redistribution of Reuters content, including by framing or similar means, is expressly prohibited without the prior written consent of Reuters. Reuters shall not be liable for any errors or delays in the content, or for any actions taken in reliance thereon. Reuters and the Reuters sphere logo are registered trademarks and trademarks of the Reuters group of companies around the world.