Each spring, tens of thousands of oncologists from around the world descend on the annual meeting of the American Society of Clinical Oncology where the latest findings are presented and discussed regarding the latest treatments and discoveries. Reporters and journalists attend in order to communicate some critical new discovery to an eager public.
Out of these many reports, six abstracts each year are chosen for presentation at the plenary session, which is attended in a huge room with 20,000 or more. Usually perhaps one of the talks is truly groundbreaking. But 1998 was a banner year. One talk introduced irinotecan, the first new drug effective against colorectal cancer since 5-fluorouracil in the 1950s. That paper and what it led to changed life expectancy for advanced colorectal cancer. Another talk was a trial of Herceptin (trastuzumab), showing its efficacy for metastatic HER2-positive breast cancer—it radically changed treatment for over 20% of breast cancer patients. A third talk reviewed the results of a major breast cancer prevention trial, showing that hormonal therapy could prevent breast cancer.
Out of all of these incredible studies, which was the only one to find its way to the front page of The New York Times, above the fold no less? To appreciate this, we need some background on the status of prostate cancer in 1998. Prostate-specific antigen (PSA) is a protein first described by Richard Ablin in 1969 at the University of Arizona and usually associated with prostate malignancies. It was initially approved for use as a tumor marker; elevated levels could be used to monitor the progress of disease in patients with prostate cancer. Higher levels signaled disease progression, while lower levels reflected improvement.
Attempts were soon made to utilize PSA in asymptomatic men to screen for the presence of cancer. For reasons beyond the scope of this article, this practice had enormous success in diagnosing cancer in these men. Starting around 1985, this practice became extremely widespread and, as a consequence, the incidence of prostate cancer in the U.S. almost tripled. However, prostate cancer mortality, the true purpose of screening, was unaffected. Other observational studies as well reflected confusion as to the benefits of PSA screening while its use continued to soar. What was needed was a randomized trial to settle the question.
Fernand Labrie (1937-2019) was a French-Canadian clinician-scientist, a well-known professor at Universite Laval in Quebec City. He was the author of the paper reported in The New York Times, and what his study, a randomized trial of PSA screening, purported to show was a 50% reduction in prostate cancer mortality—no wonder it was a plenary paper at ASCO and a front-page article in the Times!
It was a warm Monday afternoon when the plenary session took place and Labrie presented his study. I was sitting in one of the first few rows with a friend, Barry Kramer, director of cancer screening for the National Cancer Institute, to witness this critical talk. Labrie described that 46,000 men, members of the Canadian nationalized health care system, were randomized, half (23,000) to receive annual PSA screening and half (23,000) not. He then showed that 8,000 men underwent screening while 38,000 did not and that those who had the screening, after a five-year follow-up, had about 50% the prostate cancer mortality of those not screened.
While this was presented, Barry’s jaw dropped. After Labrie finished, a statistician critiqued and reviewed the study for the audience, a common practice at the plenary session. He explained that Labrie had violated one of the cardinal rules of statistical analysis. Of the 23,000 men randomized to screening, only 7,000 had actually come to the clinic for PSA screening while 16,000 never showed up. On the control side, 1,000 had somehow gotten a PSA test, while the other 22,000 had not been screened correctly. Labrie had combined all those who got a PSA and compared them to all those who did not get a PSA—8,000 versus 38,000—and had thus arrived at his conclusion. But any statistical novice is aware of the rule of intent-to-treat analysis—you analyze as you randomize. You must compare the two arms in total—23,000 versus 23,000 no matter whether they comply with the intervention or not. If you analyzed the data that way, there was no difference in mortality. In truth, because such a small fraction did comply, one could really conclude nothing from this study.
Suffice it to say, there was an uproar, and subsequently ASCO changed its rules for how to select plenary papers. Labrie’s paper did ultimately get published in some journal but is never seriously cited in the discussion regarding the efficacy of PSA screening (there have since been four or five other randomized trials). Whether the reader should get PSA screening or not is something to be discussed with his own health care provider.
If you read Labrie’s abstract in retrospect, it does seem that it was written in such a way as to purposely mislead the reviewers. Is there a word for chutzpah in Quebec French?
Alfred I. Neugut, MD, PhD, is a medical oncologist and cancer epidemiologist at Columbia University Irving Medical Center/New York Presbyterian and Mailman School of Public Health in New York.
This article is for educational purposes only and is not intended to be a substitute for professional medical advice, diagnosis, or treatment, and does not constitute medical or other professional advice. Always seek the advice of your qualified health provider with any questions you may have regarding a medical condition or treatment.
