The outcry among many physicians and patients over a government panel’s recent announcement that healthy men should no longer receive P.S.A. blood testing to detect prostate cancer is rooted in a long and impassioned history among cancer screening advocates that early detection must always save lives. But as science has taught us, that’s not always the case.
As early as 1913, physicians and laypeople formed the American Society for the Control of Cancer, which later became the American Cancer Society, bearing this hopeful message: “With early recognition and prompt treatment, the patient’s life may often be saved.” The idea had some scientific basis. Patients whose cancers appeared to be less extensive at diagnosis lived longer, on average, than those whose cancer was more widespread. The organization put this philosophy in action, publicizing a series of “danger signals” that suggested possible early warning signs of cancer, including breast lumps, irregular bleeding, sores that did not heal and persistent weight loss. “Delay kills!” posters bluntly warned.
Early detection achieved perhaps its greatest triumph with the introduction of the Pap smear in the 1940s. Annual scraping of a woman’s cervix detected precancerous cells that could be eliminated before actual cancer emerged. Thanks in large part to Pap testing, deaths from cervical cancer fell sharply.
Armed with this example, cancer society officials and cancer specialists after World War II began exploring other technologies that could find cancers earlier, like breast self-examination and mammography for breast cancer and rectal examinations to pick up cancerous lumps in men’s prostates.
Yet around this same time, research emerged that questioned the cancer society’s original assumption that cancer was a local disease that spread in a gradual and orderly fashion. Scientists had found cancer cells in the blood of patients with seemingly tiny, localized cancers, suggesting that cancer cells could spread silently early in the course of disease. In that case, so-called early detection might not really be early, or of much value. Researchers coined the term “biological predeterminism” to underscore how the cellular makeup of a given cancer — rather than when it was found — was most important in determining survival.
This was not a message, however, that people wanted to hear. The “war on cancer,” which was formally declared in the 1970s, was predicated on the optimistic message of early detection and treatment. How could it be reconciled with the idea that nothing could be done to better one’s chances of survival?
Patients whose cancers had been detected by new screening strategies like mammograms and Pap smears, and then cured, were particularly upset. To many, the notion that their proactive efforts had not saved their lives — that they would have done just as well if their cancers had been picked up later — defied sense.
In recent decades, we have learned just how complex and variable cancer may be. In some instances, malignant or precancerous cells disappear by themselves. Cancers of the same organ, whether it’s the breast, prostate, lungs or otherwise, can be very sensitive — or stubbornly resistant — to the same treatment.
Some screening tests, too, are better than others. Physicians had long obtained chest X-rays and sputum samples to screen for lung cancer in healthy patients, but studies in the 1960s showed that these tests did not lower mortality from lung cancer; researchers are now studying whether CT scans may be a better screening option. Routine colonoscopy, on the other hand, has been shown to lower the death rate from colon cancer in people over age 50. Earlier is sometimes better, sometimes not.
So where do we go from here? During my training as a physician, I learned the early detection mantra. I have ordered mammograms for women in their 40s and P.S.A. tests for men in their 50s. The notion that there might be undetected cancer in any of my patients makes me uncomfortable.
But it would be wrong to ignore the government task force’s conclusions. Given that slow-growing cancers may not need to be detected early to be cured, and that fast-growing ones may be fatal regardless of when they are found, the fact that its review of the available epidemiological data shows that P.S.A. testing does not save lives from prostate cancer should not come as a huge surprise.
Dr. Arthur I. Holleb, a surgeon and medical director of the American Cancer Society in the 1960s and ’70s, and a staunch defender of early detection, once memorably wrote that epidemiology was “the practice of medicine without the tears.” He was frustrated because he had seen so many unscreened patients die, while patients who were screened were saved. To him, forgoing screening, even when large-scale studies showed that screening may not work, seemed unconscionable. But his logic was based on anecdote, and falsely assumed that cure was necessarily linked to screening. Prospective randomized clinical trials, like those done for P.S.A. testing, seek to avoid this type of bias.
We all wish that finding prostate cancer earlier would lead to better outcomes. But as I will tell my patients, that is no reason to keep on ordering the P.S.A. test.
Dr. Barron H. Lerner, professor of medicine and public health at Columbia University Medical Center, is the author of “The Breast Cancer Wars” and, this month, “One for the Road: Drunk Driving Since 1900.”