Abbreviated Breast MR
Author: Janice S. Sung, MD, FSBI
Background: Limitations of screening mammography
Mammography is the primary imaging modality used for breast cancer screening, as it is cost effective, widely available, and the only imaging modality validated by multiple randomized controlled trials and meta-analyses to reduce breast cancer mortality. The mortality reduction achieved by mammographic screening averages 30% in randomized trials, and approaches and even exceeds 40% in observational and case control studies of women actually screened (1-5). Early detection through breast cancer screening also results in less invasive and aggressive therapy (6).
Although mammography is an effective screening test, it is an imperfect test. A major limitation of mammography is that its sensitivity is decreased in dense breasts (i.e. those that are heterogeneously dense or extremely dense), which includes ~43% of women aged 40-74 years in the United States (7). Another limitation of mammography is the interval cancer rate. About 10-20% of all breast cancers are interval cancers, meaning that these cancers present symptomatically and are diagnosed within 365 days after a negative mammogram (8, 9). Interval cancers tend to be more aggressive and are associated with a worse prognosis than cancers detected by mammographic screening (8, 9).
Attempts to improve breast cancer screening:
Digital breast tomosynthesis (DBT) has been one attempt to improve full-field digital mammography (FFDM). The primary advantages of DBT are improved sensitivity and improved specificity. DBT increases the cancer detection rate by 0.7 to 2.7 cancers/1000 women (10-12). However, cancer detection may not be improved by DBT in women with extremely dense breasts (13, 14). DBT has the added benefit of reducing the recall rate by approximately 15% (10-12). For these reasons, DBT is replacing FFDM for breast cancer screening at many centers throughout the United States.
DBT is often considered “a better mammogram”, but still relies on the same principle of X-ray imaging and does not address the main limitations of mammography, particularly in women with dense breasts. Therefore other imaging modalities have been investigated as adjunct screening tests, such as screening breast ultrasound. In women with dense breasts, screening breast ultrasound detects an additional 2-4 mammographically occult cancers per 1000 women screened (15, 16). However, ultrasound also results in many false positives, with a consistently low positive predictive value of biopsies (PPV3) of approximately 8%, leading to a significant number of unnecessary biopsies, and, historically, a high rate of recommendation for short-term interval follow up especially on the first round of screening (15, 16). PPV3 has been shown to increase with radiologist’s experience and with the availability of prior studies for comparison. For example, in one study, the PPV3 increased from between 6-9% in years 1-3 of a screening breast ultrasound program to 20% in year 4 while maintaining an incremental cancer detection rate of 3-4/1000 (17).
Therefore, both DBT and screening breast ultrasound represent minor advances in breast cancer screening. MR is the most sensitive test for detecting breast cancer and is not limited by breast density. In a study of women at elevated risk who had 3 years of screening mammograms and screening ultrasonography, an additional 15 cancers/1000 women screened were identified from a single screening MR exam; in addition, the PPV3 of MR was comparable to that seen with mammography and higher than that of screening ultrasound (18). In another study, screening breast MRI in average risk women across all breast densities found an additional 16 cancers/1000 women after a negative screening mammogram (19). Moreover, in this study, the interval cancer rate was reduced to zero, even in women undergoing MR screening only every 2 to 3 years. The PPV was comparable to that observed for screening mammography. These studies demonstrate that there is a large reservoir of breast cancers not detected on screening mammography or ultrasound, and that there is a technology (MR) available that is able to depict these cancers without sacrificing specificity. In addition, MR is known to preferentially depict invasive cancers and high grade ductal carcinoma in situ (DCIS), which are more likely to be biologically significant (20). If these cancers were not detected early by MR, some would likely progress to cause interval cancers presenting at a later stage with less favorable outcomes. However, the use of screening MR has been restricted to women at the highest risk for breast cancer primarily due to its high cost, time to perform, and perceived low PPV. Some women elect not to undergo screening breast MR. In one study where women at elevated risk were offered to undergo a screening breast MR, the most common reasons for refusing were claustrophobia, time constraints, and financial concerns, all of which may be partially alleviated with abbreviated breast MR (21).
Abbreviated breast MR:
One way to expand the use of MR to benefit more women is to perform an abbreviated breast MRI (AB-MR). With AB-MR, the scan time is reduced substantially by decreasing the number of imaging sequences acquired. In addition, with fewer imaging series, the study interpretation time is shortened. This reduction of acquisition and reading times drives a reduction in the cost of performing the MR. The first study to evaluate the concept of an AB-MR was by Kuhl et al (22). In this reader study, an initial interpretation was rendered on what approximated an abbreviated MR protocol (one T1-weighted precontrast and one postcontrast series) and compared to the full diagnostic protocol for 606 screening breast MR exams (22). The diagnostic accuracy of the abbreviated protocol was equivalent to the full diagnostic protocol, yielding an incremental cancer detection rate of 18.2/1000. Using the abbreviated protocol, the acquisition time was reduced to 3 minutes, and the interpretation time reduced to 28 seconds. Subsequently, several other reader studies similarly reported comparable sensitivity and specificity between an abbreviated and full diagnostic protocol (23-25). The sequences included in these abbreviated protocols varied slightly. All included a pre-contrast series followed by at least one post-contrast series. Some studies also included a T2-weighted or STIR sequence and/or a second post-contrast series.
To further evaluate the clinical utility of AB-MR in a multi-center setting, there is an NCI-approved research trial comparing AB-MR to digital breast tomosynthesis, the ECOG-ACRIN 1141 trial (http://ecog-acrin.org/clinical-trials/ea1141-educational-materials) (26). In this multicenter trial, approximately 1500 women with mammographically heterogeneously or extremely dense breasts (as determined on their prior mammogram) will have both an AB-MR and DBT for two years and will then be followed for 3 years. The study will compare the detection rate of invasive cancers and compare the types of cancers detected by the two modalities. In this trial, the AB-MR protocol includes a pre-contrast series, a T2 weighted sequence, and a single post-contrast series. This study completed accrual in December 2017, and preliminary results from the first round of screening will soon be available.
Limitations of AB-MR:
The main limitations of AB-MR are the cost, the frequency of false positives, the requirement for intravenous gadolinium, and access to MR. All screening tests have false positives, and breast MR is no exception. However, the PPV of MRI and AB-MR is within the range that is considered acceptable for screening mammography (19, 22).
Another limitation of AB-MR is the requirement for intravenous gadolinium administration. Minute traces of gadolinium have been found in specific areas of the brain in people who have had multiple contrast-enhanced MR exams, in particular after use of linear gadolinium compounds (27, 28). Macrocyclic gadolinium compounds appear to be less likely to cause these depositions (29, 30). The consequences of the gadolinium deposition are unknown. Gadolinium based contrast agents were approved by the FDA in 1988. After almost 30 years of use and more than 400 million doses of gadolinium, there has been no evidence of clinical symptoms resulting from gadolinium injections. In May 2017, an FDA review found no adverse health effects from gadolinium retained in the brain and found that restricting gadolinium-based agents was not warranted. Subsequently, in December 2017, the FDA required a new class warning and other safety measures for all gadolinium-based contrast agents. All new patients are now required to be provided with educational information prior to receiving a gadolinium contrast injection. Further information on this topic may be found at https://www.fda.gov/Drugs/DrugSafety/ucm589213.html.
A final limitation of AB-MR is limited access to MR. MR units are still not widely available in many areas or are already at maximum capacity at other centers. Successful implementation of AB-MR will require increased access to MR across the country, including rural centers. Companies are currently developing dedicated, less expensive breast magnets in the range of $500,000 that would allow more centers to offer breast MR. The number of trained MR technologists also will need to increase. To simplify the process, companies are exploring “one touch” scanning, where patient centering and all parameters are automated, allowing faster positioning and patient throughput. Access to AB-MR is also limited by billing issues. Currently there is no CPT code for an AB-MR and, at this time, most insurance companies do not cover the cost of the study. Most centers performing AB-MR currently offer it to women as an out-of-pocket expense, limiting access.
Breast MR is the most sensitive test available for breast cancer screening. To date, screening MR has been restricted to high-risk women for a variety of reasons. With the advent of AB-MR, widespread screening MR can be made available at cost comparable to conventional imaging techniques. AB-MR represents a promising new screening tool and may be the optimal supplemental imaging for average and intermediate risk women, especially in women with dense breasts. Given its high cancer detection rate, biennial screening with AB-MR could prove to be a more efficient and cost-effective supplement to mammography compared to annual whole breast screening ultrasound. Future studies will also evaluate the possibility of AB-MR as a standalone screening test, replacing mammography altogether.