Recent Trends in Screening Breast MRI

Abstract

Objective

The objective of this study was to assess trends in screening breast MRI utilization among privately insured women in the U.S. from 2007 to 2017.

Methods

The utilization of screening breast MRI among women aged 25–64 years from January 1, 2007, to December 31, 2017, was obtained using the MarketScan Commercial Database. We used Current Procedural Terminology codes to exclude breast MRI exams performed in women with a new breast cancer diagnosis and in women imaged to assess response to neoadjuvant therapy in the preceding 90 days. During the 11-year study, 351 763 study-eligible women underwent 488 852 MRI scans.

Results

An overall 55.0% increase in screening breast MRI utilization was observed over the study period, with a steadily increasing trend. The greatest annual increase in percent utilization was from 2007 to 2008 at 16.6%. The highest utilization rate was in 2017, in which 0.4% of women aged 25–64 years underwent screening breast MRI. Of the women who underwent screening MRI with sufficient follow-up, 76.5% underwent only one examination during the study period.

Conclusion

Utilization of screening breast MRI has increased steadily in the past decade to a peak of 0.4% of adult women. However, an estimated 9% of U.S. women are eligible for high-risk breast MRI screening; thus, utilization falls short of optimal compliance. Further studies to evaluate the barriers to screening compliance may help optimize utilization.

Graphical Abstract

Graphical Abstract.

Key Messages.

• An overall 55.0% increase in screening breast MRI utilization was observed from 2007 to 2017, with a steadily increasing trend.

• The highest utilization rate occurred in 2017, in which 0.4% of women aged 25–64 years underwent screening breast MRI. However, an estimated 9% of U.S. women are eligible for high-risk breast MRI screening; thus, utilization falls short of optimal compliance.

• Of the women who underwent screening MRI with sufficient follow-up, 76.5% underwent only one examination during the study period.

Introduction

The American College of Radiology (ACR), the American Cancer Society (ACS), the United States Preventive Services Task Force (USPSTF), and the National Comprehensive Cancer Network (NCCN) all recognize that early detection of breast cancer decreases mortality (1–3). For the past several decades, the mainstay for early detection has been annual screening mammography. However, multiple studies have demonstrated higher sensitivity of breast MRI screening compared to mammography in high-risk women (4–7) and have found that nearly all (97%) incident cancers detected with high-risk MRI screening were stage 0 or 1 (8). In response, a 2007 expert panel assembled by the ACS provided recommendations for breast MRI screening as an adjunct to mammography for high-risk women (9). Recently, the ACR also added the recommendation of annual MRI screening to include women with a personal history of breast cancer and dense tissue or any woman diagnosed with breast cancer before the age of 50 (2).

According to several estimates, as many as 9% of all U.S. women may be considered at high risk for breast cancer and therefore eligible for screening MRI (10–15). This includes the more than 3.1 million estimated survivors of breast cancer, which represents approximately 2.6% of the adult female population, and the 9.6% of women who have a 20% or greater lifetime risk of breast cancer as determined by risk assessment models (10–13). Additional high-risk groups include survivors treated with moderate- to high-dose chest radiation for pediatric and young adult cancers (105 000 women) and those with a genetic predisposition to breast cancer, most commonly BRCA1 or BRCA2 (0.1%–0.2% of the population) (14,15). This estimate is offset by those women who choose bilateral mastectomies to prevent or treat breast cancer and those in whom breast MRI is contraindicated (16).

Despite unanimous consensus from the ACR, ACS, USPSTF, and NCCN that annual MRI screening is known to increase early detection of breast cancer in high-risk women compared to mammography alone, previous studies have shown utilization to be low in eligible women (17), and its use is frequently outside of established guidelines (18,19).

Although previous studies suggest that breast MRI utilization is increasing (19,20), the current utilization trends are not known. Such utilization data are important for both clinicians and insurers to optimize both patient care and value. In this study, we aim to examine breast cancer screening MRI trends using a comprehensive health care claims database.

Methods

This retrospective claims analysis used data from the IBM MarketScan Commercial Database from January 1, 2007 to December 31, 2017. The Commercial Database contains deidentified medical claims from more than 150 million employees, spouses, and dependents less than 65 years of age from employer-sponsored health plans in all 50 states and U.S. territories in the 11-year period. This administrative claims database includes a variety of fee-for-service, preferred provider organizations, and capitated health plans.

The MarketScan database includes all medical claims with billing codes based on the International Classification of Diseases, 9^th and 10^th Revision, Clinical Modification (ICD-9-CM/ICD-10-CM) and Current Procedural Terminology (CPT) codes. The Institutional Review Board considered this study exempt because the MarketScan database is a limited data set.

Inclusion criteria were women ages 25–64 years with continuous enrollment in a participating health plan for at least three months from January 1, 2007, through December 31, 2017. Data from September 1, 2006, through December 31, 2006, was used to establish that participants in early 2007 met these inclusion criteria. In an effort to avoid underestimating utilization in women with concurrent Medicare insurance coverage, the older age limit was capped at 64 years.

Bilateral contrast-enhanced breast MRI utilization was captured from the outpatient services file using CPT or Healthcare Common Procedure Coding System codes for breast MRI, 77058-59 and C8903-C8908. Because billing codes are not specific to the indication or purpose for breast MRI, we developed a claims-based algorithm that used procedure and diagnosis codes to conservatively identify screening breast MRI examinations. Current ACR guidelines recommend breast MRI for screening in patients at high risk, to evaluate the extent of disease in patients with newly diagnosed breast cancer, and to evaluate response to neoadjuvant chemotherapy (21–23). We excluded patients who had encounters associated with invasive breast cancer (ICD-9 codes 174.x; ICD-10 codes C50.X1Y) or carcinoma in situ (ICD-9 code 233.0; ICD-10 codes D05.X) and secondary malignant neoplasm of the breast (ICD-9 code 198.81; ICD-10 code C79.81) within 90 days prior to the MRI to exclude diagnostic breast MRI studies performed to evaluate a patient with newly diagnosed breast cancer. We also excluded patients who had encounters associated with neoadjuvant chemotherapy (96401-450, 96521-3, 96542, 96545, and 96549) within 90 days prior to the MRI study to evaluate the response to therapy. Breast MRIs are rarely performed on a hospitalized patient; therefore, MRIs performed in the inpatient setting were not considered (Figure 1). Individual breast cancer risk was not assessed as some specific breast cancer risk information is not included in the database.

Figure 1.

Flowchart describes the eligible population and number of women who underwent screening breast MRI.

To examine the effect of dense breast notification (DBN) legislation on breast MRI utilization, states that enacted the legislation were analyzed. Each patient was assigned to a region based on residence at the time the MRI was performed, matching to the year of DBT implementation.

Data Analysis

Results are reported in person-years of observation (PYO). This measurement considers both the number of people in the study and each person’s observation time based on continuous health insurance enrollment. This is necessary to adjust for participants changing insurance enrollment status throughout the study period. The PYO rate reflects the number of MRI studies performed per 100 000 PYO.

Temporal trends were analyzed using generalized linear models with reported P-values representing the effect of year. Pairwise comparisons for effect of DBN legislation employed weighting for PYO to account for loss of enrollment status; P-values are derived from comparisons of proportions using chi-square tests. All data programming and statistical analyses were performed in SAS version 9.4 (SAS, Cary, NC).

Results

During the 11-year study, 351 763 study-eligible women underwent 488 852 MRI scans. Women undergoing breast MRI ranged in age from 25 to 64 years, with a median of 50 years (Table 1). The median duration of health insurance enrollment for women who underwent MRI was 52 months (range 4–144 months), and the mean was 61 months (SD 38.5) (Table 2). The median duration of enrollment of the total eligible study population was 24 months (range 4–144 months) and the mean was 33 months (SD 30 months).

Table 1.

Number of Breast MRI Patients in Each Age Range at the Time the MRI Was Performed

Age at time of MRI	Number of Women
25–29	6055
30–34	13 564
35–39	28 623
40–45	53 735
45–50	69 082
50–55	70 772
55–60	61 804
60–64	48 128

Table 2.

Number of Women Who Underwent Breast MRI per Length of Health Insurance Enrollment

Years enrolled	Number of Women With MRI	Percent
3 months–1 year	8740	2.48
1–<2 years	42 763	12.16
2–<3 years	52 408	14.90
3–<4 years	43 138	12.26
4–<5 years	42 446	12.07
5–<6 years	40 168	11.42
6–<7 years	21 380	6.08
7–<8 years	29 790	8.47
8–<9 years	15 922	4.53
9–<10 years	12 938	3.68
10 years or more	42 070	11.96

From 2007 to 2017, breast MRI utilization increased by 55.0%, from 290.0 to 449.4 per 100 000 PYO (P = 0.0002). The year with the greatest increase in utilization was 2008 compared to 2007, rising from 290.0 to 338.1 per 100 000 PYO, a 16.6% increase (P = 0.0001). The overall trend in utilization was primarily a steady increase, with just one calendar year demonstrating a decrease in utilization, 2009–2010, with a reduction from 358.6 to 322.4 per 100 000 PYO (10.7% decrease, P < 0.0001). The lowest utilization rate was in 2007 at 290.0 per 100 000 PYO, and the highest utilization rate in 2017 at 449.4 per 100 000 PYO or 0.4% of the study population (Figure 2).

Figure 2.

Utilization rate of women undergoing screening breast MRI per calendar year per 100 000 person-years of observation (PYO).

When subdivided by age group (Figure 3), all ages demonstrated an increase in breast MRI utilization over the study period. The greatest percent increase in utilization over the study period was seen in the 25–29 age group, from 33.1 to 76.6 per 100 000 PYO (131.4% increase, P < 0.0001) from 2007 to 2017. The smallest increase in utilization over the study period was in the 60–64 age group, from 395.3 to 531.6 per 100 000 PYO (34.5% increase, P = 0.003) from 2007 to 2017. Women aged 25–29 had the lowest utilization at 33.1 per 100 000 PYO (P = 0.002) in 2007, and women aged 50–54 had the highest utilization rate at 656.9 per 100 000 PYO (P < 0.0001) in 2017.

Figure 3.

Utilization rate of women undergoing screening breast MRI per calendar year per 100 000 person-years of observation (PYO) by age group.

The lowest performance rate of breast MRI over the 11-year study period was in the 25- to 29-year-old age group at 29.0 per 100 000 PYO. The 60–64 age group had the overall highest screening MRI utilization of 591.8 per 100 000 PYO (Figure 4). Approximately 86% (303 521/351 763) of all screening MRIs were performed in the 40–64 age range.

Figure 4.

Regional increase in the rate of breast MRI one year before implementation compared to first and second years after the implementation of dense breast notification laws.

To determine the frequency with which women underwent MRI, women with sufficient continuous enrollment to have the opportunity for a subsequent MRI were examined. Of 327 225 study-eligible women with >13 months continuous enrollment, 76.4% (250 083/327 225) underwent only one screening breast MRI. Of the 196 119 women with >18 months of health insurance enrollment after the first MRI, 21.7% (42 525/196 119) had a second MRI during that time period. Of the 112 747 women with >36 months of health insurance enrollment after the first MRI scan, 33.6% (37.9/112 747) had a second MRI scan, and 14.8% (16 702/116 276) had a third MRI scan during that time period.

Thirty-three states adopted DBN laws over the course of the study (22). The rate of MRIs performed in the year prior to implementing the DBN legislation in these 33 states was compared to the years after implementation. Overall, an increase in breast MRI from 243.9 per 100 000 PYO in the year prior to legislation to 279.4 per 100 000 PYO in the year after legislation was found in the 33 states that enacted DBN laws (P < 0.0001). The utilization in these states further increased to 297.5 per 100 000 PYO two years after the implementation of DBN laws (P < 0.0001). Utilization and the increase in utilization varied by geographic location but increased significantly in each geographic region (P < 0.0001 for all). In the southern region of the U.S., utilization rose from 169.7 per 100 000 PYO in the year before legislation to 187.7 (+10.6%) per 100 000 PYO two years after legislation; in the northeast region, utilization rose from 384.4 to 432.3 (+12.5%), in the north central region from 201.1 to 234.8 (+16.8%), and from 231.2 to 292.8 (+26.6%) in the western regions of the U.S.

Discussion

Despite a 55.0% increase in breast MRI utilization from 2007 to 2017, implementation falls far short of the recommended utilization of annual screening breast MRI. Although breast cancer risk information is not included in the database, the highest utilization rate of 0.4% (403.7 per 100 000 PYO) in 2017 is well below the estimated 9% of eligible high-risk women. In addition to low overall utilization, most (76.4%) of women who participated in MRI screening with at least 13 months subsequent insurance enrollment only underwent one MRI examination, and <15% underwent three or more exams. Furthermore, the highest utilization rate for breast MRI over the study period was in the 50–54-year-old age group. This older group has relatively less opportunity to benefit from screening in terms of life-years gained by detecting and treating early cancers compared to younger women. The lowest utilization rates occurred in the 25–30-year-old age group, who potentially have the most life-years to gain by the early detection and treatment of breast cancer.

Many potential factors may contribute to these utilization trends and overall suboptimal utilization. In terms of year-to-year utilization variability, one important influence is the release of authoritative organizations’ recommendations. In the early 2000s, the availability of breast MRI services was limited, and use was rare (24). Following early increases in MRI utilization, the ACS released guidelines in 2007 on the recommended use of breast MRI (9,19), which included screening women with a high risk of developing breast cancer based on familial factors, predisposing genetic mutations, and history of mantle radiation. Our study observed an increase in MRI use from 2007 to 2009, corresponding to the implementation of these guidelines better defining the exam target audience.

In 2009, both the NCCN and USPSTF released statements regarding the use of screening breast MRI. The NCCN expanded the scope of high-risk women eligible for screening MRI to include women with a history of atypia or lobular carcinoma in situ (25). In 2009, the USPSTF released a statement claiming insufficient evidence of breast MRI screening utility to create a guideline (26). These conflicting statements may have driven providers to order fewer screening breast MRIs, contributing to the 10.7% decline from 2009 to 2010. Our findings are similar to a study among insured women in New England, which demonstrated an increase in utilization through 2009 and subsequent decline (19). Conflicting society guidelines surrounding screening mammography recommendations issued in 2009 may have also amplified the confusion regarding supplemental screening recommendations.

In addition to variations in screening guidelines, MRI availability may have also contributed to the variable utilization rates over the study period. Analysis of Medicare and Surveillance, Epidemiology, and End Results Program of the National Cancer Institute data have frequently noted geographic variation in diagnostic care (26–28). In 2007, almost 75% of a national sample of breast imaging facilities were found to offer breast MRI, and the availability of MRI scanning was growing rapidly (20,29). A 2008 survey of the Society of Breast Imaging members found that MRI was performed in 73.8% of the 754 practices surveyed, with 62.6% performing more than five MRIs each week (27). However, following the introduction of imaging facility accreditation by the Medicare Improvements for Patients and Providers Act of 2008 (MIPPA), the number of facilities offering breast MRI decreased. MIPPA passed in July 2008 and became effective in January 2012, requiring all nonhospital suppliers of advanced diagnostic imaging services, including MRI, to obtain accreditation to receive CMS reimbursement. This may provide one explanation for the two-year decrease in breast MRI utilization from 2010 to 2012. By 2018, less than 20% of ACR-accredited imaging facilities offered breast MRI (30), with a majority of those facilities in just eight states (Figure 5)—New York, California, Florida, Texas, New Jersey, Pennsylvania, Illinois, and Maryland—and 20 states have fewer than 10 accredited breast MRI facilities each. Access to accredited breast MRI facilities is particularly limited in rural areas west of the Mississippi. As resources are not currently distributed geographically, the lack of access to MRI equipment may be contributing to the slower than expected increase in utilization rates following the accreditation deadline and the overall disappointing rates of screening breast MRI utilization.

Figure 5.

Number of ACR-accredited breast MRI facilities, as of July 2018, calculated as the number of accredited breast MRI facilities per zip code. Accredited Facility Search. American College of Radiology. https://www.acraccreditation.org/accredited-facility-search. Data from the website used with permission from the American College of Radiology (30). Abbreviation: ACR, American College of Radiology.

After the 2012 deadline for accreditation, utilization rates increased again, but at a slower rate than in the years before the introduction of MIPPA. One may have expected a decrease in breast MRI utilization in 2012, given the decline in breast MRI accredited facilities, though simultaneous institutional and state factors may have partially offset this. Institutional pressure to use expensive capital equipment may have led to increased MRI utilization (31). In addition, state-mandated DBN laws may be handled variably depending on the institution, leading to additional imaging. When comparing the number of MRIs performed in the year before the implementation of DBN laws to two years after implementation, we found an increase in screening breast MRI utilization from 220 to 264.3 per 100 000 PYO in those states that adopted the laws. Similarly, Horny et al found that some states that implemented DBN legislation before 2014 had a statistically significant increase in breast MRI utilization (22). Taken together, these factors may compensate for the expected reduction in MRI utilization due to the various guidelines and accreditation standards set forth by MIPPA.

For the individual patient, out-of-pocket cost concerns may serve as a barrier to screening breast MRI utilization. The Patient Protection and Affordable Care Act passed in March 2010 mandated full coverage for screening mammography, estrogen modulator pharmaceuticals, and genetic testing, removing the financial barrier for breast health care and increasing the visibility of the importance of breast health. With more women receiving mammograms and genetic counseling, there would be an expected increased referral rate to breast MRI, as we have seen. However, because the Patient Protection and Affordable Care Act does not cover breast MRI, and women can be billed well over $2000, cost may remain a barrier for many women (32). A study of breast MRIs at 14 institutions found that concerns about MRI costs prompted the refusal of up to 17% of women from undergoing MRI (32). In addition, a study of more than 270 000 women with high deductible health plans, and therefore high out-of-pocket costs, found breast imaging noncompliance and delays due to costs (33). The concern for costs may explain the higher duration of enrollment in women who underwent MRI compared to the total study population, as insurance offsets can result in reduced out-of-pocket expenses. The cost barrier has been partially mitigated in two states, Connecticut and Missouri, which passed legislation into law that requires expanded health care coverage to include screening breast MRI for high-risk women. However, copay and deductibles remain an issue for many women.

Individual patient factors may also explain the low utilization rates of screening breast MRI. Recommendations call for annual screening breast MRI; however, 76.4% of patients who underwent screening MRI received only one breast MRI screening during the study period, and less than 15% with sufficient continuous health insurance enrollment underwent three or more examinations. A recent review of participation in an MRI screening program suggested that >40% of women did not undergo an MRI for various reasons, including claustrophobia (25%) and medical intolerance such as a contrast allergy (7%) (32). In addition, the use of MRI is limited in specific patient populations, including patients who develop renal disease and other comorbidities or become pregnant or breastfeed over the course of the study (34). Furthermore, referring physicians may not identify women as being at high risk for breast cancer and be referred for related services (35), representing a missed opportunity to uze supplemental screening to improve early detection.

The study is limited to the population included in the database, which may not fully represent the overall U.S. population because it only represents commercially insured women in individual and employer-sponsored health plans. We excluded Medicare-aged women over the age of 65, which represents approximately 20% of the U.S. population (36). The database does not include breast MRIs performed outside of insurance policies, including those who transitioned out of the private health insurance programs included in the database. Additionally, specific medical comorbidities and devices may preclude a patient from undergoing an MRI examination. Finally, to distinguish between diagnostic MRI exams performed for a new cancer diagnosis or a response to neoadjuvant therapy and screening breast MRI, we excluded women with a breast cancer clinical encounter or neoadjuvant therapy in the prior 90 days, as defined by ICD-9/10 diagnosis codes. Additional ICD diagnosis codes may not reliably distinguish the indications for breast MRI examination within this population. Rarely, there may be women within this population undergoing breast MRI outside recommended guidelines, such as part of a diagnostic workup for suspected breast cancer or a short-term follow-up of a probably benign finding (21). This would result in an overestimation of screening utilization.

Conclusion

In conclusion, an increase in screening breast MRI was observed over the study period, though utilization remains far below targeted compliance. Barriers to full utilization of screening breast MRI exist at many levels, ranging from systemic, including professional and governmental organizations, to the ordering providers and the individual patient. Emerging technologies such as abbreviated breast MRI may mitigate some barriers to screening breast MRI by improving the patient experience and potentially at a reduced cost. Ongoing individual risk assessments in all adult women would best facilitate the identification of women who would benefit from supplemental screening. Identifying trends in screening MRI use may provide insight into the effect of these barriers and help lead to additional targeted initiatives to bolster participation.

Funding

The Center for Administrative Data Research is supported in part by the Washington University Institute of Clinical and Translational Sciences grant UL1 TR002345 from the National Center for Advancing Translational Sciences of the National Institutes of Health and grant number R24 HS19455 through the Agency for Healthcare Research and Quality. IBM Watson Health and MarketScan are trademarks of IBM Corporation in the United States, other countries, or both.

Conflict of Interest Statement

Dr Appleton is a consultant for WhiteRabbitAI, Inc, and Dr Olsen is a consultant to Pfizer. Dr Olsen has received grant funding (to institution) in the past three years from Pfizer, Merck, and Sanofi Pasteur. Mr Stwalley has personal stock ownership in Abbvie, Inc, and Bristol-Myers Squibb. All other authors have no conflict of interest to declare.