Trio of mammography studies demonstrate enhanced screening effectiveness
Mary Beth Nierengarten · Cancer · 2021
As the standard tool for breast cancer screening, mammography is recommended by most US guidelines to be started after the age of 40 years, or no later than the age of 50 years, and to be continued to at least the age of 74 years for women with an average risk of breast cancer. Beyond the age of 74 years, recommendations vary, with some guidelines citing insufficient evidence on the benefits and arms of screening in this older population. Insufficient evidence is also cited by most guidelines as the reason to not recommend routine screening for women with dense breasts, in which lesions are more difficult to detect through mammography alone.1 A 2020 study published in Radiology: Artificial Intelligence has contributed to growing research on the use of algorithms based on deep learning to augment digital mammography.2 In that study, investigators reported on the use of an artificial intelligence (AI) system designed to identify regions of the breast that look suspicious for cancer on 2-dimensional digital mammograms and to assess their likelihood of being malignant. Using a data set of 240 digital mammography images acquired from 2013 to 2016, 14 radiologists first read half of the data set without AI and then read the other half with the AI system. This was followed by a second reading in which the radiologists read the first half of the data set with AI and the second half of the data set without AI. The data set included different types of abnormalities, including false-negative cases. Using the area under the receiver operating characteristic curve (AUC) to assess whether the radiologists' reading performance was superior with the help of AI versus no AI, the investigators found that 3% more cancers were detected with the addition of the AI tool to mammography as indicated by the average difference in AUC of 0.028 (95% CI, 0.002-0.055; P = 0.03). The tool also significantly increased the average sensitivity of the readings by 0.033 (P = 0.02). The lead author of the study, Serena Pacilè, PhD, clinical research manager of Therapixel in Nice, France, says that the inclusion of false-negative cases in the data set was an important novelty of the study. “We believe that AI can help in the detection of very early signs of cancer that may be overlooked in a regular screening setup, and these results demonstrated it,” she says, adding that prospective studies are now needed to confirm these results. For Avice M. O'Connell, MD, professor of imaging sciences and director of women's imaging at the University of Rochester in New York, the key finding of the study is the reduction in false negatives with the addition of the AI tool. “We've probably reached the limit of what mammography currently can do to reduce false negatives, so if this new technology using deep learning can trigger us to reduce false negatives, that will save lives,” she says. Dr. O'Connell believes that doing everything possible to reduce the high rate of false negatives, a particular problem in omen with dense breasts, is essential, although this may mean an increase in false-positive results as well. “We need to balance this all the time,” she says. In another study, investigators examined the use of abbreviated breast magnetic resonance (AB-MR) imaging after a negative 3-dimensional mammography or digital breast tomosynthesis (DBT) screening of women with dense breasts.3 The study included 475 asymptomatic women with dense breasts who had negative or benign readings after a DBT examination. When AB-MR imaging was applied to the screening, 420 lesions (88.4%) were still considered negative; a follow-up assessment was recommended for 13 lesions (2.7%) and biopsy was recommended for 42 lesions. Thirty-nine of the 42 biopsies were completed. The addition of AB-MR imaging resulted in 12 cancers being found: 7 were invasive carcinomas ranging in size from 0.6 to 1.0 cm, and 5 were ductal carcinomas in situ. At the 6-month follow-up, 1 additional patient was diagnosed with an invasive ductal carcinoma. Overall, the addition of AB-MR to DBT resulted in a cancer detection rate of 27.4/1000 at the patient level. “We need to start thinking about how to better screen women with dense breasts, and AB-MR is an effective and feasible option,” says lead study author Susan P. Weinstein, MD, an associate professor of radiology at the Hospital of the University of Pennsylvania, in a press release.4 “As even more data come out, there will be a lot of debate about how we should screen women with dense breasts, and how we should pay for it.” She underscores the need for more research to better understand the long-term benefits of improved cancer detection on outcomes such as survival. A third study examined the benefit of mammography in an older cohort of women. Conducted by Swedish researchers, the study assessed the effectiveness of mammography screening in women up to the age of 74 years versus limiting screening to women who are aged 69 years old or younger.5 In the cohort design study, women up to the age of 74 years (the study group) and women up to the age of 69 years (the control group) were invited to mammography screening in 1986-2012. To assess the effectiveness of mammography, investigators compared breast cancer mortality between the study and control groups with the breast cancer mortality ratio, which counts only breast cancer deaths in cases diagnosed among those aged 70 to 74 years. In the report on the 20-year follow-up, breast cancer deaths occurred in 1040 women in the study group in 1173 women in the control group. The breast cancer mortality rate ratio was 0.80 (95% CI, 0.75-0.85) for women in the study group in comparison with those in the control group; this represented a 20% lower mortality rate for women who underwent mammography between the ages of 70 and 74 years in comparison with those who were screened only up to the age of 69 years. The breast cancer mortality rate ratio for the study group alone was 0.73 (95% CI, 0.66-0.81), which represented a nearly 17% reduction in breast cancer mortality. Overall, the study supports the effectiveness of mammography screening for women up to the age of 74 years and highlights the need for more long-term data on the overall effect of screening for this older age cohort.