JAMA comment challenges cancer screening benefits; radiologists, others disagree

October 26, 2009

A forward-looking commentary in the Journal of the American Medical Association and related news coverage in The New York Times have drawn public attention to the diagnostic limitations of mammography and prostate cancer screening and future opportunities to develop better tests.

A forward-looking commentary in the Journal of the American Medical Association and related news coverage in The New York Times have drawn public attention to the diagnostic limitations of mammography and prostate cancer screening and future opportunities to develop better tests.

"Rethinking Screening for Breast Cancer and Prostate Cancer," a special communication in the Oct. 21 JAMA reviewed 20 years of breast and prostate cancer screening research to identify what many radiologists already know are the shortcomings of the screening mammography and prostate-specific antigen (PSA) tests.

First author Dr. Laura Esserman, a professor of radiology and surgery at the University of California, San Francisco, credits the two procedures for boosting breast and prostate cancer detection rates. But she criticizes PSA screening for increasing healthcare costs and morbidity because of the overdetection and overtreatment of non–life-threatening cancers.

For evidence, she pointed out that widespread PSA testing in the U.S. has led to a far higher prostate cancer incidence rate than in the U.K., where testing was not widely adopted. Yet the mortality rate for the disease is not significantly different in the two countries.

Mammography fared somewhat better in Esserman's analysis. She gave it partial credit, along with adjuvant therapy, for the 20% to 30% decline in breast cancer mortality since 1980. But, she wrote, screening mammography has failed to realize its full potential by detecting biologically indolent tumors that do not warrant treatment and missing aggressive breast cancers that are life-threatening.

"Without the ability to distinguish cancers that pose minimal risk from those posing substantial risk and with highly sensitive screening tests, there is an increased risk that the population will be overtreated," she wrote.

To improve screening, Esserman and coauthors Dr. Ian Thompson, chair of urology at the University of Texas Health Science Center, San Antonio, and Yiwey Shieh, a medical student at UCSF, recommended developing new screening protocols based on molecular biomarkers capable of discriminating between minimal-risk and high-risk disease.

Better testing would enable clinicians to withhold treatment for minimal-risk disease, thereby lowering the cost and morbidity associated with follow-up procedures. Intensified screening and preventive interventions could be directed to patients at especially high risk of developing disease, according to their commentary.

The notoriety of these opinions rose markedly during their day of their publication when New York Times reporter Gina Kolata coaxed Dr. Otis Brawley, chief medical officer of the American Cancer Society to admit that the advantages of screening have been exaggerated. She reported that ACS is quietly planning soften its support of annual screening mammography for women over age 40. The new policy would purportedly urge at-risk women to discuss the pros and cons of screening with their physicians.

The position, if ever embraced by the ACS, was short-lived. Within hours of the article's publication, Brawley and the society issued a statement noting that some of the advantages of screening for some cancers have been overstated, but mammograms work and women should continue to get them.

For proof, the ACS statement referred to seven clinical trials finding mammography screening and clinical breast exams reduce the risk of breast cancer death. It stood by its recommendation that women age 40 and over should receive annual mammographic screening, and women at high risk should talk with their doctors about when to begin screening based on their family history.

Despite the hubbub, breast imaging researchers saw nothing new in the findings of Esserman et al. Dr. Daniel Kopans, a professor of radiology at Massachusetts General Hospital, called her observation about mammography catching nonlethal cancers but missing aggressive ones "a fundamental concept that is older than I am."

Mammography saves lives by finding moderate and slow-growing cancer that will kill in five or more years without diagnosis and treatment, according to Kopans.

"Saving these lives is no less important than saving someone from a fast-growing cancer," he said.

Dr. David Dershaw, director of breast imaging services at Memorial Sloan Kettering Cancer Institute, noted that the need to customize breast cancer screening to a women's individual risk is widely recognized and is part of clinical practice. Women with a family history of breast and cervical cancer now receive genetic testing with BCRA-1 and BCRA-2 mutations. Women deemed especially susceptible to breast cancer are prescribing a rotating regimen of breast MRI and digital x-ray mammography screenings every six months.

Though compliance with policy for routine annual mammograms has fallen slightly, Dershaw argued that most women who should be screened comprehend the procedure's importance regardless of short-lived controversies played out in the press.

"They understand it is advantageous. And increasingly they have friends, family members, or they themselves have seen the advantage of early diagnosis of breast cancer with mammography," he said.

Dr. Etta Pisano, a professor of radiology and biomedical engineering at the University of North Carolina Medical School, agreed with Esserman about the need to develop and validate biomarkers that would help radiologists differentiate between low- and high-risk breast cancers.

But much more research is needed before functional and anatomic data may be used together to assess breast cancer, Pisano said. Researchers believe they know what molecules on the surface of breast cancer cells correlate with aggressive pathologies, but more experimentation is necessary before that knowledge is applied clinically, she said.

"Anyone in this field would prefer that we only treat cancers that are 100% likely to progress," Pisano said. "In reality we can't do that. We don't have the tools to do it."