To Screen or Not to Screen Older Women for Breast Cancer: A New Twist on an Old Question or Will We Ever Invest in Getting the Answers?

Abstract

Women 65 or older (hereinafter referred to as older women) constitute one half of new breast cancer patients each year. With the graying of America, the absolute number of patients in this age group will double by the year 2030. The efficacy of screening mammography for improving survival of this growing older population has never been demonstrated in a randomized trial, leading most groups to qualify their screening recommendations for older women, especially women older than age 74 or 79. However, annual mammography is recommended for breast cancer screening surveillance for all women after treatment for breast cancer, despite the absence of clinical trial evidence for effectiveness in older women (or any other age group). Filling this evidence gap was the goal of the study in this issue of the Journal of Clinical Oncology by Lash and colleagues. Lash et al used data from a large cohort of older breast cancer survivors with stage I and II disease receiving care in managed care settings to examine the protective effect of mammography on breast cancer and all cause mortality. They found that there was a striking 31% (95% CI, 18% to 48%) reduction in the odds of death for each additional surveillance mammogram, after considering tumor characteristics, comorbidity, surgery, and other covariates. Unfortunately, even with guaranteed access in the managed care environment, use of mammography among survivors fell well below the recommended annual recommendation, with 40% of decedents having no mammograms, 32% having only one, and only 41% of survivors having four or more mammograms in the 5-year follow-up period. Given these rates of surveillance and an estimated prevalence of older breast cancer survivors of 2.3 million, the results of Lash and colleagues suggest that as many as 500,000 older women will die of their breast cancers due to suboptimal mammography surveillance. One assumes that the survival advantage afforded to older breast cancer survivors by surveillance mammography is a result of the detection of curable new primaries in the contralateral breast. There is also growing evidence that survival is improved when surveillance detects early locoregional recurrences of the primary breast cancer. The fact that overall nonbreast cancer–specific survival was greater among women who underwent surveillance mammography in the Lash et al study also suggests that part of the benefit observed may have been due to observational bias (eg, unobservable or unmeasured selection of healthier women to mammography). However, the authors did conduct statistical analyses to try to correct for any such bias (propensity scoring), and the results of a strong survival benefit were robust. Randomized trials can avoid potential observational biases, but we are left with the persistent dilemma that older women have not been included in trials in sufficient numbers to provide the necessary evidence base. Moreover, as new technology evolves, this situation is being perpetuated. For example, a recent high-profile article from the American College of Radiology Imaging Network (ACRIN) trial demonstrated that magnetic resonance imaging (MRI) can detect new breast cancers in survivors that were not seen on clinical examination or mammography. The average age in the MRI study was 53.3 11.4 years; therefore, there were no women 65 or older included, leaving us to wonder if MRI should be applied in older patients. With continued gains in life expectancy, improvements in survival from breast cancer (and other major cancer sites), and increases in incidence with age, clinicians will be caring for an ever-increasing number of older breast cancer patients and survivors. Without a major shift in emphasis in clinical trials and new investments in understanding the impact of therapy and technology on older populations, we will continue to practice with limited trial evidence. In this situation, as demonstrated by the study by Lash et al, well-designed observational studies can provide a reasonable level of evidence. This high-quality observational research was facilitated by the development of the Cancer Research Network, a National Cancer Institute (NCI) –funded collaboration between 12 large, mature, managed care systems, including Group Health Cooperative, the Harvard Pilgrim Health Plan, Henry Ford Health Systems, Fallon Community Health Plan, and Kaiser Permanente in six regions. Managed care research networks have several potential advantages for studying questions about cancer care in older women. First, access is removed from the equation in survival outcomes. Second, they include large proportions of the population from almost all regions of the United States, providing an alternative universe for population-based studies. Third, the availability of computerized administrative databases linked to laboratory and pathology data and often electronic medical records allows careful delineation of disease and interventions and comprehensive, cost-efficient long-term follow-up. Thus, this network JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 25 NUMBER 21 JULY 2