There are potential benefits and harms of screening ultrasound (US) to supplement mammographic screening of women with dense breast tissue. We conducted a comprehensive literature review of studies assessing the efficacy of screening US to supplement mammography among women with dense breasts. From a total of 189 peer-reviewed publications on the performance of screening US, 12 studies were relevant to our analysis. The reporting of breast cancer risk factors varied across studies; however, the study populations tended to be at greater than average risk for developing breast cancer. Overall, US detected an additional 0.3-7.7 cancers per 1000 examinations (median, 4.2) and was associated with an additional 11.7-106.6 biopsies per 1000 examinations (median, 52.2). Significant improvements in cancer detection in dense breasts have been achieved with the transition from film to digital mammography. Thus adjunctive screening with ultrasound should be considered in the context of current screening mammography performance. Clinicians should discuss breast density as 1 of several important breast cancer risk factors, consider the potential harms of adjunctive screening, and arrive at a shared decision consistent with each woman’s preferences and values.
Mammographic breast density reporting is an important national issue. As of May 2014, 17 states have enacted laws requiring that women be directly notified if they have dense breast tissue on a screening mammogram, and similar federal legislation has been proposed. Proponents of breast density reporting legislation note that many women do not know their breast density or that dense breasts are associated with an increased risk of breast cancer and that dense breast tissue can mask breast cancers resulting in false negative mammographic examinations.
Mammographic breast density is categorized into 1 of 4 density categories according to the American College of Radiology (ACR) Breast Imaging Reporting and Data System (BI-RADS) (almost entirely fatty, scattered areas of fibroglandular density, heterogeneously dense, and extremely dense; Figure 1 ). This information is routinely included in mammography reports sent to referring clinicians.
Current and proposed breast density legislation mandates directly informing patients if they have dense breasts, defined as either heterogeneously dense or extremely dense breasts. Approximately 47% of women undergoing screening mammography have mammographically dense breasts. As more women begin to receive density notification letters, many will be turning to their health care providers for additional information to guide their decisions regarding supplemental breast cancer screening.
For women with dense breasts and whose mammograms show no abnormality, adjunctive screening tests may detect cancers masked by dense breast tissue. Whole-breast ultrasound (US) has been suggested to supplement mammography screening of women with dense breasts because of its improved ability to detect cancer in dense breast tissue, lack of associated ionizing radiation exposure, and its wide availability. Currently breast US is used primarily in the diagnostic setting to evaluate specific findings identified by physical examination or mammography. It is less widely used as a screening test in women who are at high risk for developing breast cancer and cannot tolerate a magnetic resonance imaging (MRI) and to date is infrequently used to screen women at average risk.
Automated breast US (ABUS) is a relatively new technology that standardizes imaging acquisition using an automated transducer, rather than a handheld transducer, and allows visualization of entire breasts in multiple planes. Some legislation suggests US as an adjunctive screening test, and 2 states, Connecticut and New Jersey, mandate payer support for its use.
The purpose of this review is to provide a summary of the evidence regarding breast US as an adjunctive screening test for use in counseling patients with mammographically dense breasts about their screening options.
Materials and methods
We performed a systematic review of all articles in the medical literature regarding screening US in women with dense breast tissue from January 2000 through April 2013. We searched the PubMed database using the key words screening, breast ultrasound, and dense breast to identify relevant articles. We included studies using either ABUS or handheld US (HHUS) as an adjunct to screening mammography in our analysis. We identified 189 references for further review. Two authors (J.R.S. and C.I.L.) independently reviewed the abstracts of these articles for relevance to the clinical performance of adjunctive screening breast US in women with dense breasts. We excluded abstracts reporting US use for diagnostic evaluation of symptomatic women, in male patients, reader studies, case reports, and review articles.
After applying these inclusion and exclusion criteria, we identified 17 studies and 3 authors (J.R.S., C.I.L., and B.L.S.) reviewed the full articles for data relevant to the diagnostic test performance of adjunct screening US in women with dense breasts. We excluded 2 studies because of overlapping patient populations with one another, 1 study because the majority of the study population did not have dense breasts and 2 additional studies with data from before the year 2000 using outdated technology (ie, US transducers 7-10 MHz or less). In total, we selected 12 studies for data abstraction and analysis: 10 studies of HHUS and 2 studies of ABUS.
For each study, we abstracted the following characteristics: years of study, study design, location, number of women, mean age, mammographic breast density, mammogram result, and percentage of the study population with a family or personal history of breast cancer. We also abstracted performance measures (biopsy rate and cancer detection rate) for adjunctive US performed after a negative screening mammogram when available or calculated them based on the reported data.
We defined the biopsy rate as the number of biopsies performed based on the US findings divided by the total number of screening US examinations. We defined the cancer detection rate as the number of cancers detected by US divided by the total number of screening US examinations. We report both measures per 1000 examinations. Additionally, for the cancers detected, we report the proportion of invasive breast cancers and node-negative cancers.
To assess performance differences between US used in routine clinical practices and in a multiinstitutional randomized clinical trial (American College of Radiology Imaging Network [ACRIN] 6666), we performed an additional subanalysis of 4 studies: 3 reporting data from practices in Connecticut compared with the ACRIN 6666 prospective multicenter clinical trial of combined mammography and US screening compared with mammography alone in women at increased breast cancer risk.
Connecticut was the first state to enact breast density legislation and require insurance coverage for screening US, with rapid adoption of screening US across a variety of clinical practices. ACRIN 6666 assessed screening US performed by experience breast imagers in a more homogeneous high-risk patient population, compared with the other studies. The performance measures included in the subanalysis were biopsy rate and cancer detection rate.
Finally, to provide additional context to our evidence report, we provide a review of current major medical society recommendations for adjunct screening US for women with dense breasts.
Results
Study design and patient population characteristics
Table 1 demonstrates the design and characteristics of each study included in the analysis. Of the 10 HHUS studies, 3 were prospective, 5 were retrospective and 2 did not report the study design. Both studies using ABUS were prospective. Four of the HHUS studies were performed in the United States, 4 in Italy, 1 in Singapore, and 1 in Israel. Both ABUS studies were performed in the United States. The studies were conducted from 2000 to 2010, with the more recent studies performed in the United States and Italy. Eight of the 10 HHUS studies and both of the ABUS studies included more than 1000 women, with the largest study including 9960 women.
| Study | Years of study | Study design | Location | Number of women | Mean age, y | Mammographic breast density | Mammogram result | Family history, % a | Prior breast cancer, % |
|---|---|---|---|---|---|---|---|---|---|
| HHUS | |||||||||
| Crystal et al, 2003 | 2000-2002 | NR | Israel | 1517 | 52.1 | Scattered, heterogeneously and extremely dense | Negative | 21 b | 21 b |
| De Felice et al, 2007 | 2000-2006 | Prospective | Italy | 1754 | 65 | Heterogeneously and extremely dense | Negative | NR | NR |
| Brancato et al, 2007 | 2003-2006 | NR | Italy | 5227 | 51.9 | Heterogeneously and extremely dense | Negative | NR | NR |
| Corsetti et al, 2008 | 2000-2007 | Retrospective | Italy | 9157 | 52 | Heterogeneously and extremely dense | Negative | NR | NR |
| Berg et al, 2008 (ACRIN 6666) | 2004-2006 | Prospective | United States | 2501 | 55 | Heterogeneously and extremely dense | Negative | NR | 53% |
| Leong et al, 2012 | 2002-2004 | Prospective | Singapore | 141 | 45.1 | Heterogeneously and extremely dense | Negative | 17% | 5% |
| Hooley et al, 2012 | 2009-2010 | Retrospective | United States (Connecticut) | 648 | 52 | Heterogeneously and extremely dense | Negative | 8% | 15% |
| Weigert and Steenbergen, 2012 | 2009-2010 | Retrospective | United States (Connecticut) | 8647 | NR | Heterogeneously and extremely dense | Negative | NR | NR |
| Parris et al, 2013 | 2009-2010 | Retrospective | United States (Connecticut) | 5519 | 52 | Heterogeneously and extremely dense | Allowed positive mammogram | 42% | 6% |
| Girardi et al, 2013 | 2009-2010 | Retrospective | Italy | 9960 | 51 | Heterogeneously and extremely dense c | Negative | NR | 10% |
| ABUS | |||||||||
| Kelly et al, 2010 | 2003-2007 | Prospective | United States | 4419 | 53 | Heterogeneously and extremely dense | Negative | 59% | 10 |
| Giuliano and Giuliano, 2013 | 2010-2011 | Prospective | United States | 3418 | 57 | Heterogeneously and extremely dense | Allowed positive mammography | 0% | 0% |
a Includes women reporting either family relatives with breast cancer or a genetic predisposition
b Women with family history or personal history of breast cancer were not separately reported
c Analysis on women with heterogeneously and extremely dense breasts extracted from study data that included all breast densities (n = 22,131).
Table 1 also details patient characteristics for each study. The mean patient ages ranged from 45 to 65 years for HHUS studies and from 53 to 57 years for ABUS studies. Nine of 10 HHUS studies focused on women with heterogeneously or extremely dense breasts; 1 study also included women with scattered fibroglandular densities. Nine of 10 HHUS studies provided data on US performance after a negative mammogram. Parris et al included some women with a positive mammogram if the US finding was in a different quadrant of the breast than the abnormal clinical or mammographic findings.
Whereas both ABUS studies provided data specific to women with heterogeneous or extremely dense breasts, only Kelly et al limited their study to women with a negative mammogram. Giuliano and Giuliano included all women who had both ABUS and a mammogram performed, with performance of each modality reported independently.
Six of 10 HHUS studies reported the proportion of women in the studies with either a family history of breast cancer or a personal history of breast cancer. However, different definitions for a family history of breast cancer were used. Although some studies included only first-degree relatives, in the family history category, others included any family member with a breast cancer history. Additionally, 1 study reported BRCA mutation carriers as a separate group from family history, and some studies combined the family and personal history of breast cancer into the same or overlapping categories.
Approximately 15% of women in the United States have a family history of breast cancer, defined as having a first-degree relative (mother, sister, or daughter) with breast cancer, and approximately 6% have a personal history of breast cancer. Based on reported risk factors, the underlying breast cancer risk in study populations varied widely but tended toward populations at increased risk.
The patient population of the largest prospective study of US screening (Berg et al ) was at substantially increased risk of breast cancer compared with the general population of women in the United States. Although less than 1% of study participants were BRCA mutation carriers, 19% had a greater than 25% lifetime risk for developing breast cancer by the Gail or Claus model, and a majority (53%) had a prior history of breast cancer.
Leong et al reported that 17% of their population had a first-degree relative with a history of breast cancer and no known BRCA mutation carriers. Girardi et al reported that 10% of their population had a prior history of breast cancer but did not report the family history of their population. Parris et al reported that 6% had a personal history and 42% of their population had a family history of breast cancer; however, this included participants with any relative with breast cancer. Crystal et al reported that 21% of their population had either a first-degree relative or a personal history of breast cancer.
For the ABUS studies, Kelly et al reported a family history of breast cancer in their study population of 30% with first-degree and 29% with second-degree relatives, whereas Giuliano and Giuliano reported excluding patients with personal or family history of breast cancer from their study.
Performance of adjunctive screening US
Table 2 demonstrates the biopsy rate and cancer detection rate of adjunctive screening US. The biopsy rate for adjunctive HHUS ranged from 11.9 to 106.6 per 1000 examinations (median, 56.1; Figure 2 ). This biopsy rate is in addition to the approximately 10.2 biopsies per 1000 examinations recommended based on screening mammography findings alone.
| Study | Number of examinations | Biopsy rate (per 1000 examinations) a | Total # cancers detected, n, % | Cancer detection rate (per 1000 examinations) b | Invasive cancers, n, % | Mean invasive cancer size, mm | Node negative, n, % c |
|---|---|---|---|---|---|---|---|
| HHUS | |||||||
| Crystal et al, 2003 | 1517 | 25.0 | 7 | 4.6 | 7 (100) | 9.6 | 6 (86) |
| De Felice et al, 2007 | 1754 | 106.6 | 12 | 6.8 | 10 (83) | 10.0 | NR |
| Brancato et al, 2007 | 5227 | 11.9 | 2 | 0.3 | 2 (100) | NR | NR |
| Corsetti et al, 2008 | 9157 | 56.1 | 37 | 4.0 | 36 (97) | NR | 29 (86) |
| Berg et al, 2008 ACRIN 6666 | 2501 | 68.0 | 11 | 4.4 | 10 (91) | 10.0 | 8 (89) |
| Leong et al, 2012 | 141 | 99.3 | 2 | 14.0 | 1 (50) | 13.0 | 1 (100) |
| Hooley et al, 2012 | 648 | 71.0 | 3 | 4.6 | 2 (67) | 6.5 | 2 (100) |
| Weigert and Steenbergen, 2012 | 8647 | 48.3 | 28 | 3.2 | 24 (85) | 19.0 | NR |
| Parris et al, 2013 | 5519 | 32.8 | 10 | 1.8 | 9 (90) | 9.7 | 7 (78) |
| Girardi et al, 2013 | 9960 | NR | 22 | 2.2 | NR | 8.0 | NR |
| ABUS | |||||||
| Kelly et al, 2010 | 6425 | 11.7 | 23 | 3.6 | 22 (96) | NR | NR |
| Giuliano and Giuliano, 2013 | 3418 | NR | 42 | 12.3 | 42 (100) | 14.3 | 41 (98) |
a Calculated as the number of biopsies performed divided by total number of US examinations
b Calculated as the number of cancers detected by US only divided by the total number of ultrasound examinations
c Calculated as the number of detected invasive cancers with negative nodes divided by the total number of invasive cancers with nodal status known.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
