Discuss as:

Rethinking how we confront cancer: Bad science and risk reduction

A new government report found the overall cancer death rate is falling, and the incidence of cancer deaths is declining in men and has leveled off in women. NBC's Robert Bazell reports.

 By Robert Bazell
Chief science and medical correspondent
NBC News

Two thought provoking and disturbing studies out Wednesday raise major questions about conduct of the “War on Cancer.” One examines  the quality of basic research and the other concludes that half of current cancer deaths could be prevented.

Almost 90 percent of early stage cancer research looking for improved treatments is wrong, according to scientists at biotechnology giant Amgen and the MD Anderson Cancer Center.  The researchers describe their findings as “shocking.”

Read Wednesday's news about the decline in cancer death rates.

The allegations about questionable research in the quest for treatments appear in the prestigious journal Nature.  C. Glenn Begley, the former head of cancer research at Amgen, and surgical oncologist Lee M. Ellis of MD Anderson Cancer Center in Houston describe how scientists at the Thousand Oaks, Calif.-based Amgen tried to replicate the results of 53 landmark cancer research papers.  By landmark, they mean papers cited by others as significant progress.  All were so-called “pre-clinical,” meaning they were studies in rodents or with cells in petri dishes. The scientists were able to replicate only 11 percent of the conclusions.  In science, replication is proof.  If a study can’t be reproduced reliably, it is wrong. 

Most of the papers in question describe gene mutations or other changes in cancer cells that could be potential targets for new cancer treatments.  Such research is obviously critical for companies like Amgen deciding how to spend hundreds of millions testing potential drugs in humans.  The findings at Amgen do not differ greatly from those at a team at Bayer HealthCare in Germany, which reported last year that it could not replicate 25 per cent of studies.

Begley and Ellis assume that fraud plays little or no role in the bad science. “These investigators were all competent, well-meaning scientists who truly wanted to make advances in cancer research,” they write.

So, what is the problem? Scientists often ignore negative findings that might raise a warning, cherry picking the results and putting the best face on their research. The practice involves many parties -- not just the scientists -- in the research process who turn blind eyes to questionable actions.

As Begley and Ellis detail it, “To obtain funding, a job, promotion or tenure, researchers need a strong publication record…Journal  editors, reviewers, and grant review committees [and I might add journalists—R.B.] often look for a scientific finding that is simple, clear and complete—a ‘perfect’ story.  It is therefore tempting for investigators to submit suspected data sets for publication, or even to massage data.” 

Whatever the motivation, the results are all too often wrong.

Begley and Ellis call for nothing less than a change in the culture of cancer research.  They demand more willingness to admit to imperfections and an end to the practice of failing to publish negative results. 

“We in the field,” the say, “must remain focused on the purpose of cancer research: to improve the lives of patients.”

While the Amgen report casts doubt on cancer research, a separate study concludes that fully half of all cancers occurring today are preventable.  It raises questions about the billions spent searching for treatments and concludes that “we must vigorously implement what we already know about preventing cancer.”

The article about prevention appears in the top-tier journal Science Translational Medicine. Epidemiologists Graham Colditz, Kathleen Wolin and public health researcher Sarah Gehlert of Washington University in St. Louis review the best data. 

According to the careful Washington University study, smoking remains the biggest cancer-causing environmental factor  -- responsible for  33 percent of cancer deaths, almost 189,000 lives a year in the U.S. alone.  Obesity now follows closely, causing 20 percent of cancer deaths, or 114,000 people a year.  Pollution and radiation (most of it from medical sources) each account for only about 2 percent of cancers. 

The argument about allocation of funds for reducing the risk of cancer versus treatment is as old as our efforts to confront cancer.  But as these authors show the evidence and the need to act on it grow ever stronger.

As a society, we have shown we can do a great deal — but not nearly enough  – about tobacco. Obesity is another story, but we must do better if we want to be serious about cancer – and all the other attendant diseases.

Even short of the huge social challenges in confronting tobacco and obesity, there are many proven relatively simple methods to cut cancer deaths.  They include effective screening tests, such as pap tests and colonoscopies. The vaccines against HPV and hepatitis B both prevent cancer-causing viral infections, and aspirin is looking ever better as a cancer control agent.

Only 1.5 percent of the U.S. cancer research budget now goes to risk reduction. The rest seeks to find treatments, an effort that Begley and Ellis show is seriously flawed.  As these two powerful studies out Wednesday show, it is high time we reorder our priorities.

Robert Bazell is NBC's chief science and medical correspondent. Follow him on Facebook and Twitter.

 Related:

Daily aspirin linked to reduced cancer risk

What kills one person every six seconds? 

More from Robert Bazell:  

Could weight-loss surgery end diabetes?