Click to enlarge
A good screening test is capable of identifying a precursor or early-stage disease in an asymptomatic patient and target a disease for which early treatment is available to improve the patient’s ultimate outcome. Controversies about mammography screening guidelines for early detection of breast cancer include the appropriate age to start and stop screening, the frequency of testing, and how to manage and notify patients with dense breasts. These controversies stem from the need to balance the benefits of early detection (with the aim of reducing breast-cancer mortality) against the costs of screening (including the cost of the test as well as the financial and emotional costs of follow-up testing such as biopsies, repeat imaging, and potential harms of overdiagnosis).
In a paper entitled “Breast-Cancer Tumor Size, Mammography, and Screening Effectiveness” in this week’s NEJM, investigators from Dartmouth Medical School and the National Cancer Institute aimed to assess the real-world effectiveness of mammography and prevalence of overdiagnosis of breast cancers in the United States. Using data from the large Surveillance, Epidemiology and End Results (SEER) database, the investigators calculated the size distribution and size-specific incidence of breast cancer among women age 40 and older. They then compared size-specific cancer case fatality rates from the era before widespread screening (1975–1979) to rates from the most recent data with 10 years of available follow-up (2000–2002).
The authors hypothesized that an effective screening program would reduce the number of large cancers detected and proportionally increase in the number of small cancers detected, implying that screening identified cancers earlier, while at a smaller size. They found that screening did increase the proportion of small breast cancers (<2 cm) detected from 36% to 68% and decreased the proportion of larger tumors (≥2 cm) detected from 64% to 32%. The incidence of larger tumors decreased by 30 cases per 100,000 women, while the incidence of smaller tumors increased by 162 per 100,000, resulting in a rate of overdiagnosis of 132 per 100,000 (representing small cancers that were not likely to progress to larger cancers).
Using the size-specific case-fatality rates of the larger cancers to estimate mortality reduction, the authors concluded that improvements in breast cancer treatment are responsible for at least two thirds (68%) of the reduction in mortality over time, while screening is responsible for the remainder. The authors acknowledge that tumor biology and genetics are more important than size, and that size is only an imperfect proxy for disease and prognosis. One possible explanation for the relatively small contribution of screening to improved survival is that many small cancers detected by screening have favorable biology, with slow growth and good response to therapy, and therefore prognosis may not be improved by earlier detection. A limitation of the study is that the authors had to assume that the disease incidence was stable over time because true breast cancer incidence is unknown. Although that assumption seems likely to be true, it has not been confirmed.
In an accompanying editorial, Dr. Joann Elmore from the Department of Medicine at the University of Washington cautions that the authors, “rely on data with extensive missing values, make assumptions about underlying disease burden that cannot be verified, and acknowledge that their estimates are imprecise.” However, she concludes that we all must agree that overdiagnosis exists, even if our estimates of its magnitude are not accurate. She adds, “One way to reduce overdiagnosis is targeted, precision screening of persons who have a higher risk of breast cancer rather than screening large populations in which the majority of persons are at a lower risk for harmful disease.”
Dr. Dan Longo, NEJM Deputy Editor adds, “It would seem that the main way forward to resolve the debates on the usefulness of screening mammography is the development of better methods of predicting the biology of mammogram-detected cancers, probably through genetic analysis.”
This study concludes that current population-based mammography screening guidelines result in overdiagnosis of cancers unlikely to affect the life span, and that the mortality reduction gained from improved breast cancer treatments is likely greater than the mortality reduction from screening. These data will help inform efforts to improve the effectiveness of screening programs to maximize the benefits and minimize harms.
Browse more From Pages to Practice »
Rebecca is a 2016-2017 NEJM Editorial Fellow and a hospitalist at Massachusetts General Hospital. She graduated from Columbia University College of Physicians and Surgeons in 2013 and completed internal medicine residency at Massachusetts General Hospital in 2016. Her interests include medical education, quality improvement, patient safety, health care delivery innovation, and teaching value in health care.