Weill Cornell Physicians Rebut Recent Criticisms on Value of Mammography
Mammographic Screening Does Save Lives
Jan 30, 2002
The journal Lancet has just published in its February 2 issue an important analysis from six physicians at Weill Cornell Medical College, rebutting a recent article published in Lancet and subsequent commentary that cast doubt on the value of mammography for preventing death from breast cancer.
It is hoped that the latest analysis will properly re-open and refocus the debate. The Weill Cornell authors believe the debate has been misguided and has done a disservice to the general public, to physicians, and to women specifically.
The authors point out fundamental flaws in the prevailing orthodoxy in studying the usefulness of screening for cancer. One of these is the mistake of routinely acting as though screening were supposed to prevent death from cancer even in the earliest time after the screening's initiation. In reality, only early treatment and its resultant cure may occur immediately, whereas the thus-prevented death would have occurred several years later.
Another, even more fundamental flaw is the routine failure to address the magnitude of any meaningful measure of the screening's usefulness. The purpose of screening for a cancer is to enhance its curability and, thereby, to reduce the risk of dying from the cancer insofar as it develops. This risk is expressed by the cancer's case-fatality rate. A critically important measure of any cancer-screening's usefulness thus is its resultant proportional reduction in the cancer's case-fatality rate. The magnitude of this reduction can be estimated from trials of the usual type (screening vs. no screening, rather than early vs. late treatment) only if they involve sufficiently long-term screening, and, of course, only with appreciation of the time lags from the cures to the thereby prevented deaths.
Drs. Ole Olsen and Peter C. Gotzsche were the original authors of the Lancet meta-analysis, last October, that reviewed seven trials of mammographic screening and concluded that "there is no reliable evidence that screening for breast cancer reduces mortality." This conclusion was widely noted and publicized. On December 9, The New York Times said that "experts and women's health advocates . . . do not know what to think about the report."
And just last week, the furor began again, with an article in the Times reporting that, "An independent panel of experts said today that there was insufficient evidence to show that mammograms prevented breast cancer deaths."
In their rebuttal, the Weill Cornell physicians point out that the Olsen-Gotzsche study mixes "irrelevant experience" (data from the first few years after the start of screening, and total mortality data from all causes) with "relevant experience" (data from after six years of screening and pertaining only to deaths from breast cancer). Thus, they miss the proper conclusion.
The Weill Cornell authors are led by Drs. Claudia Henschke and David Yankelevitz of the Department of Radiology and Dr. Olli S. Miettinen of the Department of Medicine (who is also on the faculty of McGill University). They include Drs. Mark W. Pasmantier, James P. Smith, and Daniel M. Libby, all from Weill Cornell's Department of Medicine.
The authors argue that, with mammographic screening, "the idea is to achieve early diagnosis and, thereby, early treatment, presumed to be more commonly curative than the relatively late treatment in the absence of screening. The idea thus is to reduce the case-fatality rate." When there is screening and early diagnosis and treatment, and if death is averted, the averted death would have occurred some years "after the early diagnosis and its associated early treatment."
This means that the mortality data from a randomized control trial (RCT) comparing screening with no screening must be obtained many years after the onset of screening, and that any screening must continue sufficiently long to realize its benefit.
Instead, Drs. Olsen and Gotzsche cited the total mortality data from the two trials that they considered most valid—known as the Malmo and Canadian trials. In the Canadian trial, screening was performed only for 3 to 4 years after entry into the study. In the Malmo trial, over the entire 11 years studied, 63 of the screened cohort died of breast cancer, as did 66 of the unscreened cohort. On that basis, and with other such global numbers from other studies, Olsen and Gotzsche went on to their conclusion of "no reliable evidence."
The Weill Cornell authors carefully focus only on breast cancer deaths (not overall mortality) and break down the data by the years in which the deaths occur. They point out that in the Malmo study, in women who were 55 or older at entry into the study, there were 16 breast cancer deaths in the screened cohort in the first five years vs. 13 in the control cohort; then, in the sixth year, practically equal numbers of breast cancer deaths; but, from the eighth year onward, fewer breast cancer deaths in the screened cohort: 14, compared to 31.
Likewise, in the cohorts of women aged 45 to 54 at entry into the trial, there were 10 breast cancer deaths in the screened cohort in the first five years vs. 4 in the control cohort, but, from the eighth year onward, there were fewer deaths in the screened cohort: 11, compared to 15.
Thus, the benefits of regular screening appeared after six years, and the data—at least for women 55 and older—amount to "reliable evidence" that "mammographic screening provides for reduced mortality from breast cancer after a suitable delay, naturally."
Last week, it was reported that the so-called P.D.Q. screening and editorial board of the National Cancer Institute—an independent panel of experts—had reviewed the same seven studies as Drs. Olsen and Gotzsche and, like them, had come to entertain doubts about mammography's value.
But Dr. Henschke and her colleagues remain convinced of the significance of their own analysis. And they believe their type of analysis applies not only to breast cancer screening but to all cancer screening.
Of the earlier study in Lancet, they write, "This failure is not peculiar to these two authors; it is, alas, ingrained in today's orthodoxy surrounding trials on screening for cancer. . . . For the mammography trials and reviews, the object should have been designed to pertain to case-fatality rate and to be, specifically, the reduction in it resulting from the early interventions that are provided for by screening-based early diagnoses. Pursuit of reliable valid and statistically significant/precise evidence in terms of a wrong measure of screening's usefulness is not only useless; it misleads public policy and confuses not only the public but physicians also."