Annual Screening Mammography for Breast Cancer in Women 75 Years Old or Older: To Screen or Not to Screen

Women s Imaging Original Research Hartman et al. Annual Screening Mammography of Women 75 Years Old or Older Women s Imaging Original Research Maya Hartman 1 Michele Drotman Elizabeth Kagan Arleo Hartman M, Drotman M, Arleo EK Keywords: breast cancer, older women, screening mammography, U.S. Preventive Services Task Force (USPSTF) guidelines DOI:10.2214/AJR.14.13394 Received June 23, 2014; accepted after revision October 7, 2014. Presented at the Radiological Society of North America 2014 annual meeting, Chicago, IL. 1 All authors: Department of Radiology, New York Presbyterian Hospital/Weill Cornell Medical Center, 425 E 61st St, 9th Fl, New York, NY 10065. Address correspondence to E. K. Arleo (ela9033@med.cornell.edu). AJR 2015; 204:1132 1136 0361 803X/15/2045 1132 American Roentgen Ray Society Annual Screening Mammography for Breast Cancer in Women 75 Years Old or Older: To Screen or Not to Screen OBJECTIVE. The purpose of the study was to review screening mammography examinations at our institution from 2007 through 2013 with the primary endpoint of determining the incidence of breast cancer and the associated histologic and prognostic features in women 75 years old or older. MATERIALS AND METHODS. Patients who presented for screening mammography who ultimately received a BI-RADS assessment of category 4 or 5 for a suspicious abnormality were followed retrospectively through completion of care and were analyzed with respect to pathology results, treatment, and family history. RESULTS. From 2007 through 2013, 68,694 screening mammography examinations were performed. Of these screening examinations, 4424 (6.4%) were performed of patients 75 years old or older. On the basis of these examinations, 64 biopsies were recommended. Sixty biopsies were performed, and these biopsies detected 26 breast cancers. These results correspond to a breast cancer detection rate of 5.9 per 1000 screening examinations and a positive predictive value 2 (PPV2), defined as the probability of breast cancer after a BI-RADS assessment category of 4 (suspicious abnormality) or 5 (highly suggestive of malignancy), of 40.6%. Approximately 85% (22/26) of the screening-detected cancers in the women in this age group were invasive. For those with known genetic status (18 of 26), 33% had a first-degree relative with breast cancer. CONCLUSION. Although women 75 years or older accounted for less than 10% of the total screening population during the study time period, the breast cancer detection rate in this cohort was 5.9 per 1000 screening examinations, which is compatible with the American College of Radiology s recommendations, and most of these breast cancers were invasive. These results are relevant when considering appropriate age ranges for annual screening mammography. T he motivation for annual screening mammography is to detect breast cancer in its earliest stage because early-stage disease correlates with smaller tumor size and better prognosis [1]. The risk of breast cancer increases with age: Specifically, 3.84% of women who are currently 70 years old will develop breast cancer sometime during the next 10 years, which is a larger percentage of women affected than would be affected among women who are currently 40, 50, or 60 years old [2]. However, in 2009, the U.S. Preventive Services Task Force (USPSTF) issued the following statement with their recommendations [3]: The USPSTF concludes that the current evidence is insufficient to assess the benefits and harms of screening mammography in women 75 years and older. This statement was classified as an I grade statement, which the USPSTF defines as follows [3]: Evidence is lacking, of poor quality, or conflicting, and the balance of benefits and harms cannot be determined. Although multiple studies have shown that screening mammography reduces breast cancer mortality for women older than 40 years, few clinical trials have focused specifically on the older members of this population despite the fact that the number of older breast cancer patients is rapidly increasing [4]. For example, the highly publicized Swedish Two-County Trial [5, 6], which showed a 30% risk reduction in breast cancer mortality over a 3-decade follow-up period, evaluated only patients 40 74 years old. According to another study, 1132 AJR:204, May 2015

Annual Screening Mammography of Women 75 Years Old or Older only 9% of patients enrolled in breast cancer trials were 65 years old or older [7]. Thus, although the breast imaging community [8] may disagree with the 2009 USPSTF recommendation for biennial screening mammography of women 50 74 years old [3], the USPSTF may be correct that the level of evidence available regarding annual screening mammography of women 75 years old or older is insufficient. Nevertheless, earlier task forces were more sanguine about screening older women and argued that, despite the lack of data, the benefits of screening mammography observed in younger women would likely accrue to older women as well [3]. Furthermore, the American Cancer Society (ACS) is clear about recommending annual screening mammography starting at age 40 years and continuing for as long as a woman is in good health [9]. The objective of this investigation, therefore, was to obtain much-needed data about screening mammography for the detection of breast cancer in women 75 years old or older that would contribute to determining whether this cohort should undergo annual mammography screening or not. This study is a followup to our 2013 study [10] in which we reported that patients in their 40s who according to the USPSTF [3] should not be routinely screened accounted for nearly 20% of the screening-detected breast cancers. The specific purpose of the current study was to review the screening mammography examinations performed at our institution from 2007 through 2013 with the primary endpoint of determining the incidence of breast cancer and the associated histologic and prognostic features in women 75 years old or older. Materials and Methods Subjects This study was a HIPAA-compliant retrospective cohort study and was approved by the institutional review board at our institution. The study group was assembled from all screening mammography examinations performed at our institution from February 2, 2007, through December 31, 2013. The specific inclusion criteria were age of 75 years or older at the time of screening mammography, no personal history of breast cancer, and no palpable area of concern. The specific exclusion criteria were age of younger than 75 years at the time of screening mammography, a personal history of breast cancer, and a palpable area of concern. Procedures All of the screening mammography examinations performed from 2007 through 2013 were sorted by patient age on the date of mammography to determine the number of women 75 years old or older who underwent screening mammography during the study time period. In addition, all the biopsy-proven breast cancers diagnosed during the study time period were sorted by patient age on the date of mammography to identify the patients 75 years old or older diagnosed with breast cancer during the study period. These cases were then reviewed to see if the workup leading to the diagnosis of breast cancer originated from a screening examination: If so, the histopathology and treatment were recorded; and, if not, the case was excluded. The number of screening-detected cancers in women 75 years old or older was then divided by the number of screening mammography examinations performed of women 75 years old or older to determine the cancer detection rate for this cohort during the study time period. Terms and Measures The following terms and measures from established performance benchmarks for screening mammography were used [11]: positive predictive value (PPV), PPV1, PPV2, PPV3, true-positive mammography study, false-positive mammography study, false-negative mammography study, and cancer detection rate. The PPV was defined as the number of truepositive mammography studies divided by the sum of true-positive and false-positive mammography studies. PPV1 was defined as the probability of breast cancer after a positive mammography examination. PPV2 was defined as the probability of breast cancer after a BI-RADS assessment category of 4 (suspicious abnormality) or 5 (highly suggestive of malignancy). PPV3 was defined as the probability of breast cancer in patients actually having a biopsy after a BI-RADS assessment category of 4 or 5. A true-positive mammography examination with a positive interpretation followed by the diagnosis of invasive breast carcinoma or ductal carcinoma in situ (DCIS) within 1 year. A false-positive mammography examination with a positive interpretation and no breast cancer diagnosis within 1 year. A false-negative mammography examination with a negative interpretation but followed TABLE 1: Tumor Characteristics of Women 75 Years Old or Older Diagnosed With Breast Cancer at Our Institution, 2007 2012 Characteristics No. (%) of Patients (n = 26) Stage Unknown 9 (35) Known 17 (65) 0 3 (18) 1 12 (70) 2 2 (12) 3 0 (0) Histology DCIS 4 DCIS and invasive ductal 13 Invasive ductal cancer 4 Invasive mammary cancer NOS 1 Invasive lobular cancer 4 Grade Unknown 1 Known 25 High 5 Intermediate to high 3 Intermediate 11 Low to intermediate 3 Low 3 Note DCIS = ductal carcinoma in situ, NOS = not otherwise specified. AJR:204, May 2015 1133

Hartman et al. by a breast cancer diagnosis within 1 year. The lack of access to registry data precluded calculation of this measure. The cancer detection rate was defined as the number of cancers diagnosed after a positive mammography examination divided by the total number of screening mammography studies performed. Statistical Tests and Analysis Descriptive statistical analyses were performed to describe the study cohort. The cancer detection rate and PPVs were calculated according the definitions provided earlier. Results During the approximately 7 years (2007 2013) included in the study, 68,694 screening mammography examinations were performed at our institution. Women who underwent screening ranged from 25 to 96 years old at the time of screening. Of these 68,694 mammography studies, 4424 (6.4%) were of women 75 years old or older. Of the 4424 women 75 years old or older who underwent screening mammography, 64 ultimately received a BI-RADS assessment category of 4 or 5 (64/4424 = 1.4%), and 60 biopsies were performed (60/4424 = 1.4%). Of the 60 biopsies performed, 26 yielded malignant results (i.e., either DCIS or invasive breast carcinoma). These results correspond to a cancer detection rate of 5.9 per 1000 screening examinations, a PPV2 of 40.6% (26/64), and a PPV3 of 43.3% (26/60). The tumor characteristics for the 26 women 75 years old or older with screening-detected breast cancer are summarized in Table 1. The disease stage was known for 17 of the 26 women: 88% (15/17) of the women had stage 0 (n = 3) or stage I (n = 12) disease and 12% (n = 2) had stage II disease. Histologic analysis of the carcinomas showed a ductal predominance (21/26 = 81%), with an invasive component in 85% (13 DCIS with invasive ductal carcinoma, four invasive ductal carcinomas, four invasive lobular carcinomas, and one invasive mammary carcinoma not otherwise specified = 22/26 = 85%). The disease grade was known for 25 of the 26 women: 76% (19/25) had intermediate to high nuclear grade. In terms of hormone receptor status 96% (23/24) were estrogen receptor (ER) positive, 75% (18/24) were progesterone receptor (PR) positive, and 10% (2/24) were definitively (HER2/neu) positive (Table 2). Treatment was known for 21 of the 26 patients; the remaining five patients underwent treatment outside our institution, so the treatment details are not known. Most patients (18/21 = 86%) underwent lumpectomy, one had a mastectomy, and two had no surgical treatment; the latter two patients received hormonal therapy only. Family history was known for 18 of the 26 patients: six (33%) had a first-degree TABLE 2: Hormone Receptor Status of Women 75 Years Old or Older Diagnosed With Breast Cancer at Our Institution, 2007 2012 Hormone Receptor Status No. (%) of Patients ER status Unknown 2 (8) Known 24 (92) Positive 23 (96) Negative 1 (4) PR status Unknown 2 (8) Known 24 (92) Positive 18 (75) Negative 6 (25) HER2/neu status Unknown 6 (23) Known 20 (77) Positive 2 (10) Negative 15 (75) Indeterminate 3 (15) Note ER = estrogen receptor, PR = progesterone receptor. family member with a history of breast cancer. With respect to breast density, 50% (13/26) had heterogeneously (n = 12) or extremely (n = 1) dense breasts, 38.5% (10/26) had scattered fibroglandular densities, and 11.5% (3/26) had predominantly fatty breasts. Discussion The purpose of this research was to study screening mammography in women 75 years old or older at our institution. The principal findings of our study were that, although women in this age bracket constituted only 6.4% of our screening population, the cancer detection rate in this cohort was nevertheless a significant 5.9 cancer cases per 1000 screening mammography studies. Furthermore, most of the detected cancers were intermediate to high nuclear grade and were invasive (> 85%). Our interpretation of these findings is threefold. First, 6.4% of the screening mammography examinations performed during our study time period (2007 2013) were in women 75 years old or older; however, this percentage is less than the 10% reported by the National Cancer Institute s Breast Cancer Surveillance Consortium based on data as of 2009 [12]. This discrepancy raises the possibility that the USPSTF 2009 guidelines may have had an impact on the number of women in this age group presenting for screening. This discrepancy is consistent with the results of a study by Sharpe et al. [13], which showed that the USPSTF recommendations resulted in a decrease in the utilization of screening mammography in the Medicare population. Second, the cancer detection rate in women 75 years old or older (5.9 per 1000 screening mammography examinations) meets ACR s desired goals for medical audit data (2 10 per 1000 screening mammography examinations) [14]: This result suggests that screening women in this older age bracket should be considered, especially given that the incidence of breast cancer increases with age [2] and the life expectancy of a 75-year-old woman in the United States is 12.5 14.1 more years of life [15]. Third, the screening-detected cancers in women 75 years old or older were predominantly early stage (88% stage 0 or stage I disease). This result is consistent with or slightly better than the government benchmarks for screening mammography, which recommend that screening mammography examinations yield 76% stage 0 or I cancers) [16]. In addition, our PPVs (PPV2 = 40.6%, PPV3 = 43.3%) are within acceptable interpretive 1134 AJR:204, May 2015

Annual Screening Mammography of Women 75 Years Old or Older performance criteria for screening mammography [17]. These results support the validity of our practice. Our finding that 88% of the screening-detected breast cancers in our older cohort were early stage is consistent with the studies in the literature. For example, in a 2007 comprehensive review article analyzing the available data in the literature on routine screening mammography in women older than 74 years, Galit et al. [18] found that screening women in this older cohort may be associated with the detection of disease at an earlier stage and also with a lower mortality from breast cancer. More recently, Malmgren et al. [19] analyzed 1162 patients with breast cancer 75 years old or older and found that screening-detected breast cancer in this cohort was diagnosed at an earlier stage and had better disease-specific survival than patient- or physician-detected breast cancer. Critics of screening mammography might argue that 88% early-stage disease indicates overdiagnosis. Our retort is that most of these screening-detected breast cancers were intermediate to high nuclear grade and were invasive (85%), so these cases are not cases of overdiagnosis. On the other hand, more recently, the so-called aggressiveness of disease as defined by Ki-67, a cellular marker for proliferation, and by the types of cancers (e.g., ER, PR, HER2/neu) detected may affect mortality more than invasiveness or may be better predictors of mortality from breast cancer than invasiveness [20, 21]. Furthermore, it could be argued that a steadily greater fraction of women 75 years old or older will die of other causes; therefore, a woman who is 75 years old or older with good expected longevity could have breast cancer detected by screening but still die unexpectedly within a period when, without having undergone screening, her breast cancer would have remained unknown; arguably, this scenario would constitute overdiagnosis. This argument is in concert with a recent review in which Walter and Schonberg [22] concluded that the potential harms of continued screening in older women include overdiagnosis in approximately 13 of every 1000 women screened. Interestingly, Leach et al. [23] found that many physicians reported that they would still recommend mammography to a woman older than 80 years with terminal lung cancer; this finding suggests overrecommendation and overutilization and supports the ACS recommendations that annual screening mammography should be continued only if a woman is in good health [9]. Although the clinical implication of our findings is that screening mammography should be considered in women older than 74 years, the financial implications of this practice would be substantial. Because the population is rapidly aging [24] and women tend to live longer than men [25], screening women older than 74 years old would mean more health care dollars being spent screening a population that was not previously routinely screened, with Medicare largely footing the bill. Would factoring in family history help? Probably not. Annual screening mammography starts earlier (as early as 25 years old) in women with a firstdegree relative with breast cancer, but if annual screening mammography continued later (beyond 74 years) only in women with a first-degree relative with breast cancer, the results of our study suggest that a significant percentage of breast cancers would still be missed: In our study, only one third of the women 75 years old or older diagnosed with screening-detected cancers had a first-degree family member with a history of breast cancer. The retrospective study design is a limitation of this research. Furthermore, our study fails to address what percentage of all screening-detected breast cancers during the study period were in women 75 years old or older because the total number of screening-detected BI-RADS 4 and 5 cases that ultimately were diagnosed as breast cancer was unavailable to us by computer data analysis and manually sorting through thousands of charts was determined to be not feasible; an associated weakness is that the PPV1 therefore could not be calculated. Furthermore, the possibility of self-selection should also be considered: For example, older women with denser-than-expected breasts may continue to be screened on basis of radiologists recommendations or their primary physicians recommendations or, alternatively, because they know or assume that they are at a higher-than-average risk. Thus, it is possible that the studied population is not representative of the general population 75 years old or older. Finally, without a detailed comparison of the distributions of growth markers and the types of cancers detected, the results of this study may not necessarily be generalized to the expected actual benefits of periodic screening (i.e., mortality reduction) in this cohort of older women. In conclusion, although women 75 years or older accounted for less than 10% of the total screening population during our study time period, the breast cancer detection rate in this cohort was 5.9 per 1000 screening examinations (compatible with ACR recommendations) and most of the screening-detected breast cancers were invasive. The clinical implication of these results is that screening mammography should be considered for healthy women 75 years old or older. The logical next step is to further study this often ignored population and to investigate whether guideline updates are indicated. References 1. Cardenosa G. Breast imaging companion, 3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 2008 2. Centers for Disease Control and Prevention website. Breast cancer risk by age. www.cdc.gov/cancer/ breast/statistics/age.htm. Accessed October 6, 2014 3. U.S. Preventive Services Task Force website. Breast cancer: screening. www.uspreventiveservicestask force.org/uspstf/uspsbrca.htm. Published November 2009. Accessed October 6, 2014 4. Hurria A, Leung D, Trainor K, Borgen P, Norton L, Hudis C. Factors influencing treatment patterns of breast cancer patients age 75 and older. Crit Rev Oncol Hematol 2003; 46:121 126 5. Tabár L, Vitak B, Chen TH, et al. Swedish twocounty trial: impact of mammographic screening on breast cancer mortality during 3 decades. Radiology 2011; 260:658 663 6. Yen AM, Duffy SW, Chen TH, et al. Long-term incidence of breast cancer by trial arm in one county of the Swedish Two-County Trial of mammographic screening. Cancer 2012; 118:5728 5732 7. Kemeny MM, Peterson BL, Kornblith AB, et al. Barriers to clinical trial participation by older women with breast cancer. J Clin Oncol 2003; 21:2268 2275 8. Smith RA, Duffy SW, Tabár L. Breast cancer screening: the evolving evidence. Oncology (Williston Park) 2012; 26:471 475, 479 481, 485 486 9. American Cancer Society website. American Cancer Society guidelines for the early detection of cancer: breast cancer. www.cancer.org/healthy/findcancerearly/cancerscreeningguidelines/americancancer-society-guidelines-for-the-early-detectionof-cancer. Revised October 29, 2014. Accessed October 6, 2014 10. Arleo EK, Dashevsky BZ, Reichman M, Babagbemi K, Drotman M, Rosenblatt R. Screening mammography for women in their 40s: a retrospective study of the potential impact of the U.S. Preventive Service Task Force s 2009 breast cancer screening recommendations. AJR 2013; 201:1401 1406 11. Rosenberg RD, Yankaskas BC, Abraham LA, et al. Performance benchmarks for screening mammography. Radiology 2006; 241:55 66 12. National Cancer Institute website. Breast Cancer Surveillance Consortium: clinical demographics for 2,264,089 screening mammography examinations AJR:204, May 2015 1135

Hartman et al. from 2002 2006 based on BCSC data as of 2009. 16. National Cancer Institute Breast Cancer Surveil- after breast cancer surgery. Breast J 2010; 16(sup- breastscreening.cancer.gov/data/benchmarks/ lance Consortium website. Benchmarks for cancers pl 1):S29 S33 screening/2009/table2.html. Accessed October 6, for screening mammography examinations from 21. Falato C, Lorent J, Tani E, et al. Ki67 measured in 2014 2004 2008. breastscreening.cancer.gov/statistics/ metastatic tissue and prognosis in patients with 13. Sharpe RE Jr, Levin DC, Parker L, Rao VM. The benchmarks/screening/2009/table6.html. Accessed advanced breast cancer. Breast Cancer Res Treat effect of the controversial U.S. Preventive Ser- February 4, 2015 2014; 147:407 414 vices Task Force recommendations on the use of screening mammography. J Am Coll Radiol 2013; 10:21 24 14. American College of Radiology website. ACR Appropriateness Criteria: breast cancer screening. www. acr.org/~/media/acr/documents/appcriteria/ Diagnostic/BreastCancerScreening.pdf. Published 2012. Accessed October 6, 2014 15. Centers for Disease Control and Prevention website. Table 22: life expectancy at birth, at age 65, and at age 75, by sex, race, and Hispanic origin United States, selected years 1900 2010. www. cdc.gov/nchs/data/hus/2011/022.pdf. Accessed October 6, 2014 17. Carney PA, Sickles EA, Monsees BS, et al. Identifying minimally acceptable interpretive performance criteria for screening mammography. Radiology 2010; 255:354 361 18. Galit W, Green MS, Lital KB. Routine screening mammography in women older than 74 years: a review of the available data. Maturitas 2007; 57:109 119 19. Malmgren JA, Parikh J, Atwood MK, Kaplan HG. Improved prognosis of women aged 75 and older with mammography-detected breast cancer. Radiology 2014; 273:686 694 20. Rausei S, Rovera F, Dionigi G, et al. Predictors of loco-regional recurrence and cancer-related death 22. Walter LC, Schonberg MA. Screening mammography in older women: a review. JAMA 2014; 311:1336 1347 23. Leach CR, Klabunde CN, Alfano CM, Smith JL, Rowland JH. Physician over-recommendation of mammography for terminally ill women. Cancer 2012; 118:27 37 24. U.S. Census Bureau website. The next four decades: the older population in the United States 2010 to 2050. www.census.gov/prod/2010pubs/p25-1138. pdf. Published May 2010. Accessed October 6, 2014 25. Kirkwood T. Why women live longer: stress alone does not explain the longevity gap. Sci Am 2010; 303:34 35 1136 AJR:204, May 2015