[Skip to Content]
[Skip to Content Landing]
Figure 1.
Screening Examinationsa With Follow-up Time and Associated Screening Episodeb Definitions
Screening Examinationsa With Follow-up Time and Associated Screening Episodeb Definitions

aOnly examinations with radiologist indication of screening were considered. For women without a personal history of breast cancer (PHBC), a screening examination (mammogram or magnetic resonance imaging [MRI]) was defined as a bilateral examination and without imaging of the same type (mammogram or MRI) in the prior 9 months. For women with a PHBC, a screening examination was defined as an examination without the following: a prior bilateral mastectomy, imaging of the same type (mammogram or MRI) within the prior 60 days, or a diagnosis of cancer within 6 months before the screen.

bAll analyses were conducted at the level of a screening episode. Each screening examination was turned into a screening episode. Each examination was followed for 90 days, unless there was another screening examination in the 90-day follow-up. When a screening examination was followed within 30 days, these were rolled up to the first index screening examination and followed for 90 days from the index examination.

cScreening examination occurred 31 to 90 days after index examination. The second screening examination started another screening episode with 90 days’ follow-up.

Figure 2.
Age-Adjusted Rate of Core and Surgical Breast Biopsies by Screening Modality and Breast Cancer Surveillance Consortium 5-Year Breast Cancer Risk Among 783 522 Women Without a Personal History of Breast Cancer
Age-Adjusted Rate of Core and Surgical Breast Biopsies by Screening Modality and Breast Cancer Surveillance Consortium 5-Year Breast Cancer Risk Among 783 522 Women Without a Personal History of Breast Cancer
Table 1.  
Screening Episode Modality by Biopsy Intensity and Biopsy Result Stratified by Women With a Personal History of Breast Cancer (PHBC)
Screening Episode Modality by Biopsy Intensity and Biopsy Result Stratified by Women With a Personal History of Breast Cancer (PHBC)
Table 2.  
Characteristics by Screening Modality,a Stratified by Personal History of Breast Cancer (PHBC)
Characteristics by Screening Modality,a Stratified by Personal History of Breast Cancer (PHBC)
Table 3.  
Unadjusted and Age-Adjusted Rates of Biopsy Intensity and Biopsy Result by Screening Modality and Personal History of Breast Cancer (PHBC)
Unadjusted and Age-Adjusted Rates of Biopsy Intensity and Biopsy Result by Screening Modality and Personal History of Breast Cancer (PHBC)
1.
Esserman  L, Shieh  Y, Thompson  I.  Rethinking screening for breast cancer and prostate cancer.  JAMA. 2009;302(15):1685-1692.PubMedGoogle ScholarCrossref
2.
Bleyer  A, Welch  HG.  Effect of three decades of screening mammography on breast-cancer incidence.  N Engl J Med. 2012;367(21):1998-2005.PubMedGoogle ScholarCrossref
3.
Myers  ER, Moorman  P, Gierisch  JM,  et al.  Benefits and harms of breast cancer screening: a systematic review.  JAMA. 2015;314(15):1615-1634.PubMedGoogle ScholarCrossref
4.
Nelson  HD, Pappas  M, Cantor  A, Griffin  J, Daeges  M, Humphrey  L.  Harms of breast cancer screening: systematic review to update the 2009 US Preventive Services Task Force recommendation.  Ann Intern Med. 2016;164(4):256-267.PubMedGoogle ScholarCrossref
5.
Marmot  MG.  Sorting through the arguments on breast screening.  JAMA. 2013;309(24):2553-2554.PubMedGoogle ScholarCrossref
6.
Saslow  D, Boetes  C, Burke  W,  et al; American Cancer Society Breast Cancer Advisory Group.  American Cancer Society guidelines for breast screening with MRI as an adjunct to mammography.  CA Cancer J Clin. 2007;57(2):75-89.PubMedGoogle ScholarCrossref
7.
National Comprehensive Cancer Network.  Breast Cancer. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Vol 1. Fort Washington, PA: National Comprehensive Cancer Network; 2015.
8.
Wernli  KJ, DeMartini  WB, Ichikawa  L,  et al; Breast Cancer Surveillance Consortium.  Patterns of breast magnetic resonance imaging use in community practice.  JAMA Intern Med. 2014;174(1):125-132.PubMedGoogle ScholarCrossref
9.
Stout  NK, Nekhlyudov  L, Li  L,  et al.  Rapid increase in breast magnetic resonance imaging use: trends from 2000 to 2011.  JAMA Intern Med. 2014;174(1):114-121.PubMedGoogle ScholarCrossref
10.
Haas  JS, Hill  DA, Wellman  RD,  et al.  Disparities in the use of screening magnetic resonance imaging of the breast in community practice by race, ethnicity, and socioeconomic status.  Cancer. 2016;122(4):611-617.PubMedGoogle ScholarCrossref
11.
Runowicz  CD, Leach  CR, Henry  NL,  et al.  American Cancer Society/American Society of Clinical Oncology Breast Cancer Survivorship Care Guideline.  J Clin Oncol. 2016;34(6):611-635.PubMedGoogle ScholarCrossref
12.
Buist  DS, Abraham  LA, Barlow  WE,  et al; Breast Cancer Surveillance Consortium.  Diagnosis of second breast cancer events after initial diagnosis of early stage breast cancer.  Breast Cancer Res Treat. 2010;124(3):863-873.PubMedGoogle ScholarCrossref
13.
Houssami  N, Abraham  LA, Miglioretti  DL,  et al.  Accuracy and outcomes of screening mammography in women with a personal history of early-stage breast cancer.  JAMA. 2011;305(8):790-799.PubMedGoogle ScholarCrossref
14.
Houssami  N, Abraham  LA, Kerlikowske  K,  et al.  Risk factors for second screen-detected or interval breast cancers in women with a personal history of breast cancer participating in mammography screening.  Cancer Epidemiol Biomarkers Prev. 2013;22(5):946-961.PubMedGoogle ScholarCrossref
15.
Lee  JM, Buist  DS, Houssami  N,  et al.  Five-year risk of interval-invasive second breast cancer.  J Natl Cancer Inst. 2015;107(7):djv109.PubMedGoogle ScholarCrossref
16.
Lee  JM, Ichikawa  L, Valencia  E,  et al.  Performance benchmarks for screening breast MR imaging in community practice.  Radiology. 2017;285(1):44-52.PubMedGoogle ScholarCrossref
17.
Lehman  CD, Arao  RF, Sprague  BL,  et al.  National performance benchmarks for modern screening digital mammography: update from the Breast Cancer Surveillance Consortium.  Radiology. 2017;283(1):49-58.PubMedGoogle ScholarCrossref
18.
Breast Cancer Surveillance Consortium website. 2016. http://www.bcsc-research.org/. Accessed June 11, 2017.
19.
Ballard-Barbash  R, Taplin  SH, Yankaskas  BC,  et al.  Breast Cancer Surveillance Consortium: a national mammography screening and outcomes database.  AJR Am J Roentgenol. 1997;169(4):1001-1008.PubMedGoogle ScholarCrossref
20.
BCSC Glossary of Terms, Version 2. 2017. http://www.bcsc-research.org/data/bcsc_data_definitions_version2_final__2017.pdf. Accessed July 31, 2017.
21.
Menes  TS, Kerlikowske  K, Lange  J, Jaffer  S, Rosenberg  R, Miglioretti  DL.  Subsequent breast cancer risk following diagnosis of atypical ductal hyperplasia on needle biopsy.  JAMA Oncol. 2017;3(1):36-41.PubMedGoogle ScholarCrossref
22.
Menes  TS, Rosenberg  R, Balch  S, Jaffer  S, Kerlikowske  K, Miglioretti  DL.  Upgrade of high-risk breast lesions detected on mammography in the Breast Cancer Surveillance Consortium.  Am J Surg. 2014;207(1):24-31.PubMedGoogle ScholarCrossref
23.
Tice  JA, O’Meara  ES, Weaver  DL, Vachon  C, Ballard-Barbash  R, Kerlikowske  K.  Benign breast disease, mammographic breast density, and the risk of breast cancer.  J Natl Cancer Inst. 2013;105(14):1043-1049.PubMedGoogle ScholarCrossref
24.
Kiluk  JV, Acs  G, Hoover  SJ.  High-risk benign breast lesions: current strategies in management.  Cancer Control. 2007;14(4):321-329.PubMedGoogle ScholarCrossref
25.
Tice  JA, Miglioretti  DL, Li  CS, Vachon  CM, Gard  CC, Kerlikowske  K.  Breast density and benign breast disease: risk assessment to identify women at high risk of breast cancer.  J Clin Oncol. 2015;33(28):3137-3143.PubMedGoogle ScholarCrossref
26.
Coca-Perraillon  M. Local and global optimal propensity score matching. Paper presented at: SAS Global Forum; April 16-19, 2007; Orlando, Florida. Paper 185-2007.
27.
Svahn  TM, Houssami  N, Sechopoulos  I, Mattsson  S.  Review of radiation dose estimates in digital breast tomosynthesis relative to those in two-view full-field digital mammography.  Breast. 2015;24(2):93-99.PubMedGoogle ScholarCrossref
28.
Gilbert  FJ, Tucker  L, Gillan  MG,  et al.  The TOMMY trial: a comparison of tomosynthesis with digital mammography in the UK NHS Breast Screening Programme—a multicentre retrospective reading study comparing the diagnostic performance of digital breast tomosynthesis and digital mammography with digital mammography alone.  Health Technol Assess. 2015;19(4):i-xxv, 1-136.PubMedGoogle ScholarCrossref
29.
Zuley  ML, Guo  B, Catullo  VJ,  et al.  Comparison of two-dimensional synthesized mammograms versus original digital mammograms alone and in combination with tomosynthesis images.  Radiology. 2014;271(3):664-671.PubMedGoogle ScholarCrossref
30.
Hardesty  LA.  Issues to consider before implementing digital breast tomosynthesis into a breast imaging practice.  AJR Am J Roentgenol. 2015;204(3):681-684.PubMedGoogle ScholarCrossref
31.
Lee  CI, Lehman  CD.  Digital breast tomosynthesis and the challenges of implementing an emerging breast cancer screening technology into clinical practice.  J Am Coll Radiol. 2013;10(12):913-917.PubMedGoogle ScholarCrossref
32.
Friedewald  SM, Rafferty  EA, Rose  SL,  et al.  Breast cancer screening using tomosynthesis in combination with digital mammography.  JAMA. 2014;311(24):2499-2507.PubMedGoogle ScholarCrossref
33.
Skaane  P.  Breast cancer screening with digital breast tomosynthesis.  Breast Cancer. 2017;24(1):32-41.PubMedGoogle ScholarCrossref
Original Investigation
April 2018

Breast Biopsy Intensity and Findings Following Breast Cancer Screening in Women With and Without a Personal History of Breast Cancer

Author Affiliations
  • 1Kaiser Permanente Washington Health Research Institute, Seattle
  • 2Department of Radiology, University of Washington School of Medicine, Seattle
  • 3Department of Radiology, Massachusetts General Hospital, Boston
  • 4Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, Massachusetts
  • 5Department of Radiology, University of North Carolina, Chapel Hill
  • 6Department of Internal Medicine, University of New Mexico, Albuquerque
  • 7Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts
  • 8Dartmouth Institute for Health Policy and Clinical Practice, Department of Medicine, and Norris Cotton Cancer Center, Geisel School of Medicine at Dartmouth, Lebanon, New Hampshire
  • 9Departments of Medicine and Epidemiology and Biostatistics, General Internal Medicine Section, Department of Veterans Affairs, University of California, San Francisco
  • 10Department of Biomedical Data Science, Norris Cotton Cancer Center, Geisel School of Medicine at Dartmouth, Lebanon, New Hampshire
JAMA Intern Med. 2018;178(4):458-468. doi:10.1001/jamainternmed.2017.8549
Key Points

Question  What are biopsy rates and yield in the 90 days following screening among women with and without a personal history of breast cancer (PHBC)?

Findings  In this population-based cohort including 812 164 women undergoing screening (mammography vs magnetic resonance imaging [MRI] with or without mammography), there were 2-fold higher and 5-fold higher core and surgical biopsy rates following MRI compared with mammography among women with and without a PHBC, respectively, resulting in lower invasive cancer and ductal carcinoma in situ yield for both groups.

Meaning  Women with and without PHBC who undergo screening MRI experience higher biopsy rates coupled with significantly lower cancer yield compared with mammography alone.

Abstract

Importance  There is little evidence on population-based harms and benefits of screening breast magnetic resonance imaging (MRI) in women with and without a personal history of breast cancer (PHBC).

Objective  To evaluate biopsy rates and yield in the 90 days following screening (mammography vs magnetic resonance imaging with or without mammography) among women with and without a PHBC.

Design, Setting, and Participants  Observational cohort study of 6 Breast Cancer Surveillance Consortium (BCSC) registries. Population-based sample of 812 164 women undergoing screening, 2003 through 2013.

Exposures  A total of 2 048 994 digital mammography and/or breast MRI screening episodes (mammogram alone vs MRI with or without screening mammogram within 30 days).

Main Outcomes and Measures  Biopsy intensity (surgical greater than core greater than fine-needle aspiration) and yield (invasive cancer greater than ductal carcinoma in situ greater than high-risk benign greater than benign) within 90 days of a screening episode. We computed age-adjusted rates of biopsy intensity (per 1000 screening episodes) and biopsy yield (per 1000 screening episodes with biopsies). Outcomes were stratified by PHBC and by BCSC 5-year breast cancer risk among women without PHBC.

Results  We included 101 103 and 1 939 455 mammogram screening episodes in women with and without PHBC, respectively; MRI screening episodes included 3763 with PHBC and 4673 without PHBC. Age-adjusted core and surgical biopsy rates (per 1000 episodes) doubled (57.1; 95% CI, 50.3-65.1) following MRI compared with mammography (23.6; 95% CI, 22.4-24.8) in women with PHBC. Differences (per 1000 episodes) were even larger in women without PHBC: 84.7 (95% CI, 75.9-94.9) following MRI and 14.9 (95% CI, 14.7-15.0) following mammography episodes. Ductal carcinoma in situ and invasive biopsy yield (per 1000 episodes) was significantly higher following mammography compared with MRI episodes in women with PHBC (mammography, 404.6; 95% CI, 381.2-428.8; MRI, 267.6; 95% CI, 208.0-337.8) and nonsignificantly higher, but in the same direction, in women without PHBC (mammography, 279.3; 95% CI, 274.2-284.4; MRI, 214.6; 95% CI, 158.7-280.8). High-risk benign lesions were more commonly identified following MRI regardless of PHBC. Higher biopsy rates and lower cancer yield following MRI were not explained by increasing age or higher 5-year breast cancer risk.

Conclusions and Relevance  Women with and without PHBC who undergo screening MRI experience higher biopsy rates coupled with significantly lower cancer yield findings following biopsy compared with screening mammography alone. Further work is needed to identify women who will benefit from screening MRI to ensure an acceptable benefit-to-harm ratio.

Introduction

Although mammography is the only test with evidence demonstrating breast cancer mortality reduction, its utility is challenged by false-positive recalls leading to additional imaging and invasive benign biopsies, interval invasive cancers, and overdiagnosis.1-5 Balancing the benefit-to-harm ratio of screening strategies has generated substantial debate, particularly over what constitutes a harm for whom, and what weight harms should be given relative to the benefit of a potential life saved.

Screening breast magnetic resonance imaging (MRI) is recommended to augment screening mammography for women at high breast cancer risk (eg, >20% lifetime risk)6,7; use among average- or low-risk women is not recommended. Screening breast MRI has increased in community practice, but its use still remains low even among the more than 2.8 million US women with a personal history of breast cancer (PHBC).8-10 Current guidelines recommend annual mammography for all women with treated breast cancer11 and recommend against routine MRI screening except for women who meet high-risk criteria for increased breast cancer surveillance.6,11 Screening mammography in women with a PHBC has limitations, with approximately 35% of second breast cancers presenting as interval cancers within 1 year of a negative result on a surveillance mammogram,12,13 with 5-year risk of interval invasive second breast cancer varying widely across women.14,15 This presents a clear opportunity to improve clinical care women with a PHBC. New imaging technologies offer promise for screening, but biopsy rates and yield have not been well established, particularly in community settings.

We previously reported findings from community-based Breast Cancer Surveillance Consortium (BCSC) studies on screening mammography and MRI performance and breast cancer outcomes in the 12 months following screening.16,17 This study does not incorporate radiologists’ examination interpretation or follow each examination for 12 months; instead, we examine biopsy intensity and yield in the 90 days following a screening examination (screening mammography alone vs screening MRI with or without screening mammography) in community settings by women’s PHBC and by breast cancer risk in women without PHBC.

Methods

This study included data from 6 BCSC registries (Carolina Mammography Registry [North Carolina], Kaiser Permanente Washington Registry [Washington State], Metro Chicago Breast Cancer Registry, New Hampshire Mammography Network, San Francisco Mammography Registry, and Vermont Breast Cancer Surveillance System). The BCSC registries collect information on examinations performed at participating facilities in their defined catchment areas and link this information to local pathology databases and state tumor registries or regional Surveillance, Epidemiology, and End Results programs to obtain population-based cancer data.18,19 Planned analyses required complete capture of benign and malignant biopsies and their results; therefore, we limited our sample to examinations from BCSC facilities with complete biopsy capture. Demographic and breast cancer risk factor data including age, first-degree family history, and time since last mammogram were collected using a self-reported questionnaire completed at each screening examination.

We included women with at least 1 screening digital mammogram or screening breast MRI from 2003 to 2013. We only included examinations with a radiologist’s indication of screening. For women without a PHBC, a screening examination (mammogram or MRI) was defined as a bilateral examination without the same type of imaging in the prior 9 months. For women with a PHBC, screening examinations included examinations in women without the following: prior bilateral mastectomy, imaging of the same type (mammogram or MRI) within the prior 60 days, or a breast cancer diagnosis within the prior 6 months.20

Analyses were conducted at the level of a screening episode (Figure 1). Each examination was followed for 90 days, unless there was another screening examination in the 90-day follow-up. The first examination within the episode is referred to as the index examination. In cases with 2 screening examinations within 30 days, we combined these into 1 episode followed for 90 days from the index examination. For screening examinations that occurred 31 to 90 days apart, the index examination was followed until the next screening examination more than 30 days later; the second screening examination started another screening episode with 90 days’ follow-up. Magnetic resonance imaging episodes included screening MRI with or without a screening mammogram because MRI alone really means MRI with adjunct mammography outside the 30-day window.8,16

All biopsies were linked with an individual screening episode to ensure that biopsy intensity was based on the most invasive biopsy performed (surgical biopsy greater than core biopsy greater than fine-needle aspiration) within the 90-day follow-up period. For 571 episodes (476 mammography and 95 MRI episodes), the follow-up period for biopsy ascertainment was truncated by another screening episode occurring within 31 to 90 days following the index examination. This ensured that a biopsy was only linked to 1 episode. Biopsy result was based on the most invasive finding found from any biopsy (invasive greater than ductal carcinoma in situ [DCIS] greater than high-risk benign greater than benign). High-risk benign diagnoses included lobular carcinoma in situ and atypical hyperplasia; benign findings included usual ductal hyperplasia, fibroadenoma, cystosarcoma phyllodes, and calcifications.21-23 We separated high-risk benign diagnoses from other benign results because the clinical course of action may differ.24

We calculated the BCSC 5-year risk score25 to evaluate imaging limited to women without a PHBC because the risk score is not designed for women with a PHBC. Risk scores were calculated for each episode using age, race, first-degree family history, breast biopsy history, and Breast Imaging Reporting and Data System (BI-RADS) breast density, and were categorized into low (<1.00%), average (1.00%-1.66%), intermediate (1.67%-2.49%), and high/very high (≥2.50%) risk.25 We used BI-RADS breast density interpretation from the index examination. Women’s addresses at each episode were geocoded using census block group and linked to median household income. Facility characteristics included profit status, academic affiliation, hospital-based location, and on-site MRI biopsy capability.

The BCSC registries and the Statistical Coordinating Center received institutional review board approval for active or passive consenting processes to enroll participants, link data, and perform analytic studies and a Federal Certificate of Confidentiality and other protections for the identities of participating women, physicians, and facilities.

Statistical Analysis

Frequency distributions of biopsy intensity, biopsy result within biopsy intensity, demographic characteristics (age at examination, race/ethnicity, income, rurality, BCSC risk, family history), mammography (examination year, density, time since last mammogram), and facility characteristics were examined and stratified by screening episode type and PHBC.

Unadjusted rates of each biopsy intensity (per 1000 episodes) were computed using the total number of screening episodes having a specific biopsy intensity divided by the total number of episodes. Unadjusted rates of biopsy yield (high-risk benign/DCIS/invasive) were computed within each biopsy intensity category (rate = total number of screening episodes linked to a specific biopsy intensity and result divided by the total number of screening episodes linked to a specific biopsy intensity). For age-adjusted rates, logistic regression models were fit for each outcome of interest (eg, core biopsy) including age as a covariate. Parameter estimates from each model were used to compute predicted probabilities based on age values in our sample. Finally, predicted probabilities were combined using weights based on the age distribution of the overall sample. We used the same approach to examine rates by BCSC risk category (low, average, intermediate, high/very high) for episodes in women without a PHBC. We combined core and surgical biopsy categories, and DCIS and invasive categories. Rates were stratified by screening modality and PHBC.

We conducted a propensity-matched sensitivity analysis to tightly control for potential confounding between women receiving different episode types. Predicted probability of MRI episodes was computed using a logistic regression model with BCSC registry, age, examination year, family history, and breast density. A SAS macro26 was used to perform 2 mammogram:1 MRI matching (without replacement) based on a 0.2–standard deviation caliper width of the propensity score logit. Unadjusted and age-adjusted biopsy intensity and finding rates were computed (on the subsets of matched episodes) separately by PHBC. Analyses were performed using SAS 9.4 (SAS Institute).

Results

We examined 2 048 994 imaging episodes from 812 164 women; the median number of episodes per woman was 2 for mammography and 1 for MRI (interquartile range, 1-3 overall). Mammogram episodes included 101 103 episodes in 36 318 women with PHBC and 1 939 455 episodes in 780 373 women without a PHBC; MRI episodes included 3763 episodes in 2323 women with PHBC and 4673 episodes in 3149 women without PHBC. Biopsy rate was higher following MRI than mammography: 6.3% (n = 236) vs 2.2% (n = 2231) in PHBC episodes and 10.5% (n = 489) vs 1.6% (n = 30 757) in episodes without a PHBC (Table 1). There was little difference in biopsy intensity by imaging modality regardless of PHBC status; most episodes had core biopsies as the most intensive biopsy.

Compared with MRI, mammography episodes with core and surgical biopsies had a higher yield of DCIS and invasive breast cancer and lower benign biopsy rates regardless of PHBC. Among episodes with a PHBC, 531 (34.7%) of core biopsies associated with a mammography episode resulted in a DCIS or invasive cancer diagnosis compared with 32 (19.5%) associated with an MRI episode; for surgical biopsies the percent was higher but had the same comparative relationship. For episodes without a PHBC, overall DCIS and invasive cancer rates were lower than among episodes with a PHBC, but we observed the same comparative relationship.

Compared with mammography episodes, MRI episodes (Table 2) were more likely to occur in women with a first-degree family history of breast cancer, particularly among episodes without a PHBC (72.3% vs 16.4% mammography only). Higher breast density (heterogeneously or extremely dense) was more likely in women with MRI vs mammography episodes. The BCSC 5-year breast cancer risk was higher in MRI episodes.

The same patterns were observed regardless of PHBC, with higher biopsy rates and intensity for MRI vs mammography episodes (Table 3). Age-adjusted core and surgical biopsy rates (per 1000 episodes) were nearly double (57.1; 95% CI, 50.3-65.1) following MRI compared with mammography (23.6; 95% CI, 22.4-24.8) in women with a PHBC, with larger differences in women without a PHBC. High-risk benign lesions were more commonly identified following MRI compared with mammography regardless of PHBC; confidence intervals overlapped in women with a PHBC but not in women without a PHBC. In contrast, DCIS and invasive yield (per 1000 episodes) was significantly higher following mammography compared with MRI episodes in women with a PHBC (mammography, 404.6; 95% CI, 381.2-428.8; MRI, 267.6; 95% CI, 208.0-337.8) and nonsignificantly higher, but in the same direction, in women without a PHBC (mammography, 279.3; 95% CI, 274.2-284.4; MRI, 214.6; 95% CI, 158.7-280.8).

Propensity matching yielded comparable populations (eTable 1 in the Supplement) to the main sample with 2 notable differences: a much larger proportion of mammogram episodes having a family history in women without a PHBC, and higher density distributions in mammogram episodes in women with and without a PHBC. There were few differences in the unadjusted and age-adjusted propensity-matched results compared with the main analysis. Continued higher DCIS and invasive cancers yield was observed in mammogram episodes with larger differences in episodes without PHBC (eTable 2 in the Supplement). However, high-risk benign findings were no longer higher in MRI episodes in women with a PHBC, which may be due to propensity matching leading to a higher density distribution in a smaller sample of mammogram episodes.

Among women without PHBC, age-adjusted core and surgical biopsy rates following mammography and MRI episodes were stratified by BCSC 5-year breast cancer risk (Figure 2). Among mammography episodes, age-adjusted biopsy rates increased with increasing risk. Biopsy rates were approximately 5-fold higher for MRI compared with mammography episodes across BCSC risk groups. Age-adjusted biopsy rates in MRI episodes increased by risk category, but with overlapping confidence intervals.

Discussion

We evaluated biopsy rates by intensity and yield within 90 days following screening mammography and MRI episodes among women with and without a PHBC. After accounting for age, we observed more than 2-fold higher core and surgical biopsy rates following MRI among women with a PHBC and more than 5-fold higher core and surgical biopsy rates following MRI in women without, compared with mammography. Regardless of PHBC status, rates of DCIS and invasive cancer diagnosed within 90 days were higher following mammography vs MRI, with significantly higher rates following mammography among women with a PHBC. In contrast, higher rates of high-risk benign lesions were detected following MRI compared with mammography regardless of PHBC; this was not observed in our propensity-matched sensitivity analysis. Differences in core and surgical biopsy rates by imaging modality were not explained by breast cancer risk because higher biopsy rates were observed across all risk groups following MRI compared with mammography. We expected and observed higher DCIS and invasive cancer rates with increasing biopsy intensity. We also expected higher biopsy rates following MRI because of its higher sensitivity and lower specificity compared with mammography. While positive findings on mammography may be resolved with subsequent diagnostic mammography and/or ultrasound, positive screening MRI results are usually resolved with short-interval follow-up MRI or biopsy.

Many individuals believe that reports of the rapid increase in use of MRI8,9 mean that MRI is overused. The reality is that most women with a PHBC, as well as high-risk women without a PHBC, do not receive breast MRI in community settings.8 This is consistent with the recommendation that women not receive routine MRI screening except those meeting high-risk criteria for increased breast cancer surveillance.6,11 The use of MRI may facilitate identifying high-risk women eligible for genetic counseling and/or primary prevention by identifying high-risk benign lesions that would increase 5-year risk to greater than 3%,25 but guidelines have not been established to support this practice. Our findings support the need to define and identify the appropriate selection of women for MRI screening to minimize the harms in low-risk women. Better understanding clinical pathway patterns of multimodality screening and their associated outcomes could help to simultaneously improve the patient care experience and the health of populations while reducing per capita costs for achieving high-quality outcomes for high-risk populations and among the growing population of women with a PHBC.

Our results should not be compared with traditional screening performance measures, which take into account examination interpretation with 12 months’ follow-up for cancer outcomes.16,17 We included any biopsy occurring within a 90-day period following a screening examination regardless of examination interpretation; thus, these results are not comparable to cancer detection rates, which include interpretation and 12-month follow-up. Multimodality screening with mammography and MRI is recommended for women with at least 20% lifetime breast cancer risk6,7 and some women with a PHBC.11 Nearly all women who receive screening MRI receive it as an adjunct to mammography. However, there is substantial variability as to how adjunct imaging occurs—from same day with MRI or mammography first to a 6-month alternating interval. Differences in timing between screening examinations are driven by multiple factors including clinician preference and availability/access and were major drivers in our intentional analytic strategy to focus on screening episodes with a limited biopsy follow-up period. We defined our MRI episodes to include MRI in combination with screening mammography within 30 days of one another and also included MRI alone with no adjunct screening, in part due to small sample size, but also due to the underlying clinical relevance because MRI alone really means MRI with adjunct mammography outside the 30-day window.8,16

Despite reports demonstrating rapid increases in the uptake of screening MRI,8,9 we did not see big changes in MRI episodes by year regardless of PHBC status; only a small percentage of imaging episodes in women with a PHBC (<4%) or among high-risk women (<1%) were MRIs. We found important differences regardless of PHBC in who received MRI, including younger age and other demographic characteristics.8,10 Age adjustment was critical to help account for differences in biopsy rates and findings by imaging modality and will be important for other evaluations comparing imaging strategies and performance. Interestingly, our propensity-matched analyses yielded results similar to those of our full sample. We were able to evaluate whether BCSC 5-year risk explained differences in biopsy intensity and findings, but only among women without a PHBC. Genetic mutation data are not routinely available in the BCSC, which would likely improve our ability to characterize the underlying risk of women and appropriateness of screening MRI. It remains possible that our analyses have residual confounding by risk, but it is unlikely to account for the magnitude of difference observed by imaging modality; however, we are reassured by the fact that our propensity-matched sensitivity analyses yielded similar results.

Important strengths include the inclusion of biopsy intensity and findings from 2 040 558 mammogram episodes and 8436 MRI episodes from 136 BCSC community and academic radiology facilities linked to pathology databases, as well as state and regional tumor registries. Our study includes a geographically and racially representative US population sample, likely reflecting US clinical radiology practice.

Limitations

We acknowledge that there may be misattribution of the most significant pathological findings with the most invasive biopsy finding, but we believe that it would be unlikely for women to undergo a more intensive biopsy after identifying DCIS or invasive cancer. There were a few DCIS and invasive findings associated with fine-needle aspiration, suggesting that we may have missed some additional biopsies in this small number of cases. We were also unable to evaluate biopsy guidance, which we acknowledge as clinically important.

There is hope that tomosynthesis will decrease the rate of false-positive results and improve biopsy yield. However, in addition to increased radiation dose,27 acquisition and interpretation time,13,28,29 and cost,30,31 tomosynthesis has been shown in nonrandomized studies to improve cancer detection and recall rates, but potentially increase benign biopsy rates.28,32,33 We were unable to evaluate tomosynthesis in this study.

Conclusions

In women with and without a PHBC, we observed clinically and statistically higher biopsy rates in the 90 days following screening MRI compared with mammography episodes, which resulted in lower rates of DCIS and invasive cancer findings following MRI. Higher biopsy rates and lower cancer yield following MRI were not explained after accounting for age or 5-year breast cancer risk or propensity-matched analyses. Further work is needed to identify women who will benefit from screening MRI to ensure an acceptable benefit-to-harm ratio. Women who undergo screening MRI should also be notified that their likelihood of undergoing a core or surgical breast biopsy is significantly higher than for women undergoing mammography alone, with a lower likelihood of clinically actionable findings.

Back to top
Article Information

Accepted for Publication: December 18, 2017.

Corresponding Author: Diana S.M. Buist, PhD, MPH, Kaiser Permanente Washington Health Research Institute, 1730 Minor Ave, Ste 1600, Seattle, WA 98101 (buist.d@ghc.org).

Published Online: February 12, 2018. doi:10.1001/jamainternmed.2017.8549

Author Contributions: Dr Buist and Ms Abraham had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Buist, C.I. Lee, J.M. Lee, Lehman, Stout, Tosteson, Kerlikowske, Onega.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Buist, Abraham, C.I. Lee, O’Meara.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Abraham, Henderson, Onega.

Obtained funding: Buist, C.I. Lee, Lehman, Henderson, Hill, Tosteson, Kerlikowske, Onega.

Administrative, technical, or material support: Buist, C.I. Lee, Kerlikowske.

Study supervision: Buist, Lehman.

Conflict of Interest Disclosures: None reported.

Funding/Support: This research was supported by the National Cancer Institute’s (NCI’s) Breast Cancer Surveillance Consortium P01 CA154292. Data collection was supported by the NCI (P01 CA154292, U54CA163303, and HHSN261201100031C). Cancer data collection was supported in part by several state public health departments and cancer registries throughout the United States. A full description of these sources is available at http://www.bcsc-research.org/work/acknowledgement.html. Dr C.I. Lee’s effort was supported in part by the American Cancer Society (126947-MRSG-1416001CPHPS). The collection of cancer incidence and vital status data used in this study was supported, in part, by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the NCI’s Surveillance, Epidemiology, and End Results (SEER) Program under contract HHSN261201000140C awarded to the Cancer Prevention Institute of California, contract HHSN261201000035C awarded to the University of Southern California, and contract HHSN261201000034C awarded to the Public Health Institute; and the Centers for Disease Control and Prevention’s (CDC’s) National Program of Cancer Registries, under agreement U58-DP003862-01 awarded to the California Department of Public Health; Vermont Cancer Registry, Vermont Department of Health; Cancer Surveillance System of the Fred Hutchinson Cancer Research Center, which is funded by Contracts N01-CN-005230, N01-CN-67009, N01-PC-35142, HHSN261201000029C, and HHSN261201300012I from the SEER Program of the NCI with additional support from the Fred Hutchinson Cancer Research Center and the State of Washington; New Hampshire State Cancer Registry supported in part by cooperative agreement U55/CCU-121912 awarded to the New Hampshire Department of Health and Human Services, Division of Public Health Services, Bureau of Disease Control and Health Statistics, Health Statistics and Data Management Section; North Carolina Central Cancer Registry, which is partially supported by the CDC under cooperative agreement DP12-120503CONT14; Colorado Central Cancer Registry, which is partially supported by the Colorado State General Fund and the CDC (National Program of Cancer Registries) under Cooperative Agreement U58000848; New Mexico Tumor Registry supported, in part, by NCI Contract NO1-PC-35138 and by the University of New Mexico Cancer Center, a recipient of NCI Cancer Support Grant P30-CA118100. Manuscripts including the Metro Chicago Breast Cancer Registry Data were supported in part by the Illinois Department of Public Health, Illinois State Cancer Registry, which is partially supported by the CDC under cooperative agreement DP12-120504CONT15.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Breast Cancer Surveillance Consortium Group Information: Linn Abraham, MS, Kaiser Permanente Washington; Andrew Avins, MD, University of California, San Francisco, Kaiser Permanente Northern California; Thad Benefield, MS, University of North Carolina, Chapel Hill; Erin Aiello Bowles, MPH, Kaiser Permanente Washington; Mark Bowman, University of Vermont; Diana Buist, PhD, MPH, Kaiser Permanente Washington; David Burian, BA, University of California, San Francisco; Mike Butler, BS, University of Vermont; Elyse Chiapello, BASc, University of California, San Francisco; Rachael Chicoine, BS, University of Vermont; Firas Dabbous, PhD, Advocate Health Care; Evan de Bie, BS, University of California, Davis; Tammy Dodd, Kaiser Permanente Washington; Therese Dolecek, PhD, MS, University of Illinois at Chicago; Scottie Eliassen, MS, Dartmouth College; Kevin Filocamo, MA, Kaiser Permanente Washington; Pete Frawley, BA, Kaiser Permanente Washington; Hongyuan Gao, MS, Kaiser Permanente Washington; Charlotte Gard, PhD, MS, New Mexico State University; Berta Geller, PhD, University of Vermont; Martha Goodrich, MS, Dartmouth College; Mikael Anne Greenwood-Hickman, MPH, Kaiser Permanente Washington; Cindy Groseclose, University of Vermont; Louise Henderson, PhD, MSPH, University of North Carolina, Chapel Hill; Sally Herschorn, MD, University of Vermont; Deirdre Hill, PhD, University of New Mexico; Michael Hofmann, MS, University of California, San Francisco; Alejandro Hughes, MPH, University of Illinois at Chicago; Rebecca Hubbard, PhD, University of Pennsylvania; Tiffany Hoots, BA, University of North Carolina, Chapel Hill; Kathleen Howe, AA, University of Vermont; Laura Ichikawa, MS, Kaiser Permanente Washington; Bonnie Joe, MD, University of California, San Francisco; Doug Kane, MS, Kaiser Permanente Washington; Karla Kerlikowske, MD, University of California, San Francisco; Gabe Knop, BS, University of North Carolina, Chapel Hill; Casey Luce, MSPH, Kaiser Permanente Washington; Lin Ma, MS, University of California, San Francisco; Terry Macarol, RT(R), Advocate Health Care; John Mace, PhD, University of Vermont; Jennifer Maeser, MS, University of Washington; Kathy Malvin, BA, University of California, San Francisco; Katie Marsh, MPH, University of North Carolina, Chapel Hill; Diana Miglioretti, PhD, University of California, Davis; Anne Marie Murphy, PhD, Metropolitan Chicago Breast Cancer Task Force; Ellen O’Meara, PhD, Kaiser Permanente Washington; Tracy Onega, PhD, MA, MS, Dartmouth College; Tiffany Pelkey, BA, University of Vermont; Dusty Quick, University of Vermont; Melissa Rabelhofer, AB, Kaiser Permanente Washington; Garth Rauscher, PhD, University of Illinois at Chicago; KatieRose Richmire, BA, Kaiser Permanente Washington; Scott Savioli, MA, Dartmouth College; Deborah Seger, BA, Kaiser Permanente Washington; Brian Sprague, PhD, University of Vermont; Katherine Tossas-Milligan, MS, University of Illinois at Chicago; Anna Tosteson, ScD, Dartmouth College; Lisa Vu, MPH, University of California, San Francisco; Donald Weaver, MD, University of Vermont; Julie Weiss, MS, Dartmouth College; Karen Wernli, PhD, Kaiser Permanente Washington; Bonnie Yankaskas, PhD, University of North Carolina, Chapel Hill; Weiwei Zhu, MS, Kaiser Permanente Washington.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health. The ideas and opinions expressed herein are those of the authors and endorsement by the State of California, the California Department of Public Health; Illinois Department of Public Health; New Hampshire Department of Health and Human Services; the National Cancer Institute, and the Centers for Disease Control and Prevention or their contractors and subcontractors is not intended nor should be inferred.

Additional Contributions: We thank the participating women, mammography facilities, and radiologists for the data they have provided for this study. We thank R. Yates Coley, PhD, Kaiser Permanente Washington Health Research Institute, for consulting with our propensity-matched analyses. No compensation was received by Dr Coley.

Additional Information: Procedures for requesting BCSC data for research purposes are detailed at http://www.bcsc-research.org.

References
1.
Esserman  L, Shieh  Y, Thompson  I.  Rethinking screening for breast cancer and prostate cancer.  JAMA. 2009;302(15):1685-1692.PubMedGoogle ScholarCrossref
2.
Bleyer  A, Welch  HG.  Effect of three decades of screening mammography on breast-cancer incidence.  N Engl J Med. 2012;367(21):1998-2005.PubMedGoogle ScholarCrossref
3.
Myers  ER, Moorman  P, Gierisch  JM,  et al.  Benefits and harms of breast cancer screening: a systematic review.  JAMA. 2015;314(15):1615-1634.PubMedGoogle ScholarCrossref
4.
Nelson  HD, Pappas  M, Cantor  A, Griffin  J, Daeges  M, Humphrey  L.  Harms of breast cancer screening: systematic review to update the 2009 US Preventive Services Task Force recommendation.  Ann Intern Med. 2016;164(4):256-267.PubMedGoogle ScholarCrossref
5.
Marmot  MG.  Sorting through the arguments on breast screening.  JAMA. 2013;309(24):2553-2554.PubMedGoogle ScholarCrossref
6.
Saslow  D, Boetes  C, Burke  W,  et al; American Cancer Society Breast Cancer Advisory Group.  American Cancer Society guidelines for breast screening with MRI as an adjunct to mammography.  CA Cancer J Clin. 2007;57(2):75-89.PubMedGoogle ScholarCrossref
7.
National Comprehensive Cancer Network.  Breast Cancer. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines). Vol 1. Fort Washington, PA: National Comprehensive Cancer Network; 2015.
8.
Wernli  KJ, DeMartini  WB, Ichikawa  L,  et al; Breast Cancer Surveillance Consortium.  Patterns of breast magnetic resonance imaging use in community practice.  JAMA Intern Med. 2014;174(1):125-132.PubMedGoogle ScholarCrossref
9.
Stout  NK, Nekhlyudov  L, Li  L,  et al.  Rapid increase in breast magnetic resonance imaging use: trends from 2000 to 2011.  JAMA Intern Med. 2014;174(1):114-121.PubMedGoogle ScholarCrossref
10.
Haas  JS, Hill  DA, Wellman  RD,  et al.  Disparities in the use of screening magnetic resonance imaging of the breast in community practice by race, ethnicity, and socioeconomic status.  Cancer. 2016;122(4):611-617.PubMedGoogle ScholarCrossref
11.
Runowicz  CD, Leach  CR, Henry  NL,  et al.  American Cancer Society/American Society of Clinical Oncology Breast Cancer Survivorship Care Guideline.  J Clin Oncol. 2016;34(6):611-635.PubMedGoogle ScholarCrossref
12.
Buist  DS, Abraham  LA, Barlow  WE,  et al; Breast Cancer Surveillance Consortium.  Diagnosis of second breast cancer events after initial diagnosis of early stage breast cancer.  Breast Cancer Res Treat. 2010;124(3):863-873.PubMedGoogle ScholarCrossref
13.
Houssami  N, Abraham  LA, Miglioretti  DL,  et al.  Accuracy and outcomes of screening mammography in women with a personal history of early-stage breast cancer.  JAMA. 2011;305(8):790-799.PubMedGoogle ScholarCrossref
14.
Houssami  N, Abraham  LA, Kerlikowske  K,  et al.  Risk factors for second screen-detected or interval breast cancers in women with a personal history of breast cancer participating in mammography screening.  Cancer Epidemiol Biomarkers Prev. 2013;22(5):946-961.PubMedGoogle ScholarCrossref
15.
Lee  JM, Buist  DS, Houssami  N,  et al.  Five-year risk of interval-invasive second breast cancer.  J Natl Cancer Inst. 2015;107(7):djv109.PubMedGoogle ScholarCrossref
16.
Lee  JM, Ichikawa  L, Valencia  E,  et al.  Performance benchmarks for screening breast MR imaging in community practice.  Radiology. 2017;285(1):44-52.PubMedGoogle ScholarCrossref
17.
Lehman  CD, Arao  RF, Sprague  BL,  et al.  National performance benchmarks for modern screening digital mammography: update from the Breast Cancer Surveillance Consortium.  Radiology. 2017;283(1):49-58.PubMedGoogle ScholarCrossref
18.
Breast Cancer Surveillance Consortium website. 2016. http://www.bcsc-research.org/. Accessed June 11, 2017.
19.
Ballard-Barbash  R, Taplin  SH, Yankaskas  BC,  et al.  Breast Cancer Surveillance Consortium: a national mammography screening and outcomes database.  AJR Am J Roentgenol. 1997;169(4):1001-1008.PubMedGoogle ScholarCrossref
20.
BCSC Glossary of Terms, Version 2. 2017. http://www.bcsc-research.org/data/bcsc_data_definitions_version2_final__2017.pdf. Accessed July 31, 2017.
21.
Menes  TS, Kerlikowske  K, Lange  J, Jaffer  S, Rosenberg  R, Miglioretti  DL.  Subsequent breast cancer risk following diagnosis of atypical ductal hyperplasia on needle biopsy.  JAMA Oncol. 2017;3(1):36-41.PubMedGoogle ScholarCrossref
22.
Menes  TS, Rosenberg  R, Balch  S, Jaffer  S, Kerlikowske  K, Miglioretti  DL.  Upgrade of high-risk breast lesions detected on mammography in the Breast Cancer Surveillance Consortium.  Am J Surg. 2014;207(1):24-31.PubMedGoogle ScholarCrossref
23.
Tice  JA, O’Meara  ES, Weaver  DL, Vachon  C, Ballard-Barbash  R, Kerlikowske  K.  Benign breast disease, mammographic breast density, and the risk of breast cancer.  J Natl Cancer Inst. 2013;105(14):1043-1049.PubMedGoogle ScholarCrossref
24.
Kiluk  JV, Acs  G, Hoover  SJ.  High-risk benign breast lesions: current strategies in management.  Cancer Control. 2007;14(4):321-329.PubMedGoogle ScholarCrossref
25.
Tice  JA, Miglioretti  DL, Li  CS, Vachon  CM, Gard  CC, Kerlikowske  K.  Breast density and benign breast disease: risk assessment to identify women at high risk of breast cancer.  J Clin Oncol. 2015;33(28):3137-3143.PubMedGoogle ScholarCrossref
26.
Coca-Perraillon  M. Local and global optimal propensity score matching. Paper presented at: SAS Global Forum; April 16-19, 2007; Orlando, Florida. Paper 185-2007.
27.
Svahn  TM, Houssami  N, Sechopoulos  I, Mattsson  S.  Review of radiation dose estimates in digital breast tomosynthesis relative to those in two-view full-field digital mammography.  Breast. 2015;24(2):93-99.PubMedGoogle ScholarCrossref
28.
Gilbert  FJ, Tucker  L, Gillan  MG,  et al.  The TOMMY trial: a comparison of tomosynthesis with digital mammography in the UK NHS Breast Screening Programme—a multicentre retrospective reading study comparing the diagnostic performance of digital breast tomosynthesis and digital mammography with digital mammography alone.  Health Technol Assess. 2015;19(4):i-xxv, 1-136.PubMedGoogle ScholarCrossref
29.
Zuley  ML, Guo  B, Catullo  VJ,  et al.  Comparison of two-dimensional synthesized mammograms versus original digital mammograms alone and in combination with tomosynthesis images.  Radiology. 2014;271(3):664-671.PubMedGoogle ScholarCrossref
30.
Hardesty  LA.  Issues to consider before implementing digital breast tomosynthesis into a breast imaging practice.  AJR Am J Roentgenol. 2015;204(3):681-684.PubMedGoogle ScholarCrossref
31.
Lee  CI, Lehman  CD.  Digital breast tomosynthesis and the challenges of implementing an emerging breast cancer screening technology into clinical practice.  J Am Coll Radiol. 2013;10(12):913-917.PubMedGoogle ScholarCrossref
32.
Friedewald  SM, Rafferty  EA, Rose  SL,  et al.  Breast cancer screening using tomosynthesis in combination with digital mammography.  JAMA. 2014;311(24):2499-2507.PubMedGoogle ScholarCrossref
33.
Skaane  P.  Breast cancer screening with digital breast tomosynthesis.  Breast Cancer. 2017;24(1):32-41.PubMedGoogle ScholarCrossref
×