Do Cancer Screenings Really Save Lives?

Suzy Khimm analyzes the science behind recent recommendations to forgo some cancer screenings. Inside the fight to make the technology more accurate.

12.03.09 9:35 PM ET

Laura Vincent’s yearly mammograms never picked up on the tumor that could have killed her. For years, doctors mistook the cancerous growth in her left breast for normal dense tissue. It was only a breast self-exam that led the 49-year-old to bring a pea-sized lump to her doctor’s attention in May 2005. Within months, Vincent’s cancer progressed so quickly that doctors had to perform a mastectomy to save her life.

But while mammograms didn’t catch her deadly cancer, Vincent is incensed about the new recommendations from the U.S. Preventative Services Task Force that women under 50 should forgo the screenings. “We don’t have any alternatives,” says Vincent, a graphic designer who lives in the suburbs of northwest Illinois, and posts actively on the message boards at “Until you come up with a safe, cost-effective, viable alternative, don’t close us out—give us a choice.”

“We have more data than ever before,” says Dr. Peter Carroll of UCSF. “We’re just not applying it the way we should.”

Vincent’s criticisms point to a major problem underlying the recent controversy over cancer screenings: the early-detection technology at our disposal can’t reveal everything doctors and patients want to know, yet few patients are aware of emerging alternatives. For years, medical experts have acknowledged that the standard X-ray mammogram is a significantly less accurate tool for detecting cancer in women under 50 than in older women, whose breast tissue is less dense. The result has been an overabundance of false positives in younger women—as well as false negatives, as in Vincent’s case.

Mammography’s shortcomings are widely acknowledged, but it is the only screening method that has consistently proven to reduce the breast cancer mortality rate within the population at large. The basic procedure, in which low-dose X-rays are used to detect changes in breast tissue, hasn’t changed much since its wide-scale adoption three decades ago. The medical community, in the meantime, has developed a growing interest in developing better screening tools in hopes of catching fatal cancers while they’re still treatable—not just to fight breast cancer, but also ovarian cancer, prostate cancer, blood cancers, and many other forms of the disease.

But despite the billions spent annually to develop more advanced tests to detect cancer in its early stages, experts say the future of cancer screening is unlikely to be a silver bullet, stand-alone test. Rather, doctors will gather a complex set of data about their patients—using blood tests, genetics, digital imaging, and advanced computer technology—not only to detect cancer, but also to determine which patients should be screened in the first place.

This multi-faceted approach is, largely, a response to the inadequacies of current cancer screening tests. In the late 1980s, Dr. E. David Crawford, a professor at the University of Colorado Health Sciences Center in Denver, was among the doctors who helped popularize the use of the Prostate-Specific Antigen (PSA) test, which detects a substance in the blood linked to prostate cancer. Within years, the test became routine in doctors’ offices across the country, and “the number of advanced cases of prostate cancer that walked through the door decreased substantially.” Between 1993 and 2005, the death rate for prostate cancer declined by more than one-third—a drop that many in the medical community expected to be linked to the increasing prevalence of PSA testing.

The evidence, however, has been less than conclusive, as an elevated PSA level alone does not appear to be a reliable test for deadly incidence of the disease. One U.S.-based study published last March in the New England Journal of Medicine (which Crawford co-authored) revealed that the use of the test had no impact on the prostate cancer mortality rate. Just as with breast cancer and mammography, a growing number of doctors have come to believe that the current usage of the PSA test has resulted in widespread overdiagnosis as well as risky, invasive overtreatment.

Part of the problem is that there are other variables—such as localized infections—that can contribute to elevated PSA levels, and many men have non-fatal prostate cancers that shouldn’t require aggressive treatment. Taken by itself, then, a single PSA test is a weak indicator at best. “Most men that have [prostate cancer] die with it, not of it,” says Crawford. “We need to be able to have a test or series of tests besides PSA to help us ferret out who has cancer that’s significant.”

Technological advances like Magnetic Resonance Imaging (MRI) and Computerized Tomography (CAT scans) have yielded far more detailed images of the body. But it remains unclear whether such images can actually help doctors pinpoint fatal cancers and reduce the mortality rate. “MRIs can identify a huge number of things that might never go on to kill you,” says Dr. Diane Buist, an epidemiologist at the Group Health Research Institute in Seattle.

Ultrasounds and MRIs have had some success in discovering cancers overlooked by mammograms, and both are currently used as supplements to the standard X-ray. Digital mammograms hold promise in providing clearer images of younger women with denser breast tissue. But none of above technologies have yet proven to reduce the breast cancer mortality rate.

Aside from the development of advanced digital imaging techniques, one of the emerging cancer research areas is proteomics--the study of proteins, often through examining the bloodstream to find indicators of specific types of disease. As Thomas Goetz describes in Wired, researchers at places like the Canary Foundation are working to determine which of thousands of proteins—often described as biomarkers—might be evidence of cancer. Moreover, genetic testing may be able to isolate high-risk individuals who should be subject to more rigorous examination. Doctors have already begun closely monitoring individuals who carry the BRCA1 or BRCA2 genes, which researchers have linked to a high incidence for breast and ovarian cancer, prompting some women to undergo preventive mastectomies and hysterectomies. The medical community has also hailed the recent arrival of a blood test for colorectal cancer that detects a chemical substance produced by two genes present in cancerous cells.

Ultimately, there may never be any single test that can determine whether a patient has a certain type of deadly cancer. Rather, success in early detection will rely on the ability of doctors to analyze and monitor the increasingly wide range of information they’re able to gather about their patients. “We have more data than ever before…we’re just not applying it the way we should,” says Dr. Peter Carroll, chair of the urology department at the University of California San Francisco.

Private companies like Soar Biodynamics are using information technology to help solve this riddle. Soar uses a complicated series of metrics to help determine whether a man’s elevated PSA level is likely to mean the presence of fatal prostate cancer. In the future, rather than follow blanket recommendations for screenings or diagnoses, doctors could tailor their approach to individual patients--an extension of the movement toward “personalized medicine.”

To be sure, such a comprehensive approach carries its own set of limitations. While doctors routinely gather data like family histories at negligible cost, technologically advanced imaging tests, new blood tests, and genetic profiling can be significantly more expensive than traditional procedures. And in the absence of any standardized means of singling out the patients at greatest risk, it’s still difficult to convince anxious patients that tailoring early-detection tests to those at different risk levels—that is, screening some patients and not others—would be an improvement over blanket screenings.

But even cancer survivors who’ve benefited from the existing screening methods acknowledge that better early-detection methods are in order. Linda Ewing, a 48-year-old economist from Detroit, discovered she had a slow-growing tumor through a routine mammogram two years ago. While Ewing isn’t sure whether her cancer would have been detected had she not been screened, she recognizes that the current technology isn’t up to the task for women under 50. “Like a lot of patients and survivors in their 40s, I have a fair amount of ambivalence about the new guidelines,” says Ewing. “But the first step toward getting something better is to realize the limitations of what you’ve got.”

Suzy Khimm is a reporter-researcher at The New Republic. Her work has also appeared in The Economist, Slate, The Christian Science Monitor, and The Wall Street Journal Asia. Follow her on Twitter.