Search

PurposeThe aim of this study was to develop a prioritization strategy for scheduling immediate screening mammographic interpretation and possible diagnostic evaluation.MethodsA population-based cohort with screening mammograms performed from 2012 to 2020 at 126 radiology facilities from 7 Breast Cancer Surveillance Consortium registries was identified. Classification trees identified combinations of clinical history (age, BI-RADS® density, time since prior mammogram, history of false-positive recall or biopsy result), screening modality (digital mammography, digital breast tomosynthesis), and facility characteristics (profit status, location, screening volume, practice type, academic affiliation) that grouped screening mammograms by recall rate, with ≥12/100 considered high and ≥16/100 very high. An efficiency ratio was estimated as the percentage of recalls divided by the percentage of mammograms.ResultsThe study cohort included 2,674,051 screening mammograms in 925,777 women, with 235,569 recalls. The most important predictor of recall was time since prior mammogram, followed by age, history of false-positive recall, breast density, history of benign biopsy, and screening modality. Recall rates were very high for baseline mammograms (21.3/100; 95% confidence interval, 19.7-23.0) and high for women with ≥5 years since prior mammogram (15.1/100; 95% confidence interval, 14.3-16.1). The 9.2% of mammograms in subgroups with very high and high recall rates accounted for 19.2% of recalls, an efficiency ratio of 2.1 compared with a random approach. Adding women

BackgroundEstimating advanced breast cancer risk in women undergoing annual or biennial mammography could identify women who may benefit from less or more intensive screening. We developed an actionable model to predict cumulative 6-year advanced cancer (prognostic pathologic stage II or higher) risk according to screening interval.MethodsWe included 931 186 women aged 40-74 years in the Breast Cancer Surveillance Consortium undergoing 2 542 382 annual (prior mammogram within 11-18 months) or 752 049 biennial (prior within 19-30 months) screening mammograms. The prediction model includes age, race and ethnicity, body mass index, breast density, family history of breast cancer, and prior breast biopsy subdivided by menopausal status and screening interval. We used fivefold cross-validation to internally validate model performance. We defined higher than 95th percentile as high risk (>0.658%), higher than 75th percentile to 95th or less percentile as intermediate risk (0.380%-0.658%), and 75th or less percentile as low to average risk (

The COVID-19 pandemic disrupted breast cancer screening and diagnostic imaging in the United States. We sought to evaluate how medical facilities prioritized breast imaging services during periods of reduced capacity or upon re-opening after closures. In fall 2020, we surveyed 77 breast imaging facilities within the Breast Cancer Surveillance Consortium in the United States. The survey ascertained the pandemic's impact on clinical practices during March-September 2020. Nearly all facilities (97%) reported closing or operating at reduced capacity at some point during this period. All facilities were open by August 2020, though 14% were still operating at reduced capacity in September 2020. During periods of re-opening or reduced capacity, 93% of facilities reported prioritizing diagnostic breast imaging over breast cancer screening. For diagnostic imaging, facilities prioritized based on rescheduling canceled appointments (89%), specific indication for diagnostic imaging (89%), patient demand (84%), individual characteristics and risk factors (77%), and time since last imaging examination (72%). For screening mammography, facilities prioritized based on rescheduled cancelations (96%), patient demand (83%), individual characteristics and risk factors (73%), and time since last mammogram (71%). For biopsy services, more than 90% of facilities reported prioritization based on rescheduling of canceled exams, patient demand, patient characteristics and risk factors and level of suspicion on imaging. The observed patterns from this large and geographically diverse sample of facilities in the United States indicate that multiple factors were commonly used to prioritize breast imaging services during periods of reduced capacity.

BackgroundThe coronavirus disease 2019 (COVID-19) pandemic led to a near-total cessation of mammography services in the United States in mid-March 2020. It is unclear if screening and diagnostic mammography volumes have recovered to prepandemic levels and whether use has varied by women's characteristics.MethodsWe collected data on 461 083 screening mammograms and 112 207 diagnostic mammograms conducted during January 2019 through July 2020 at 62 radiology facilities in the Breast Cancer Surveillance Consortium. We compared monthly screening and diagnostic mammography volumes before and during the pandemic stratified by age, race and ethnicity, breast density, and family history of breast cancer.ResultsScreening and diagnostic mammography volumes in April 2020 were 1.1% (95% confidence interval [CI] = 0.5% to 2.4%) and 21.4% (95% CI = 18.7% to 24.4%) of the April 2019 prepandemic volumes, respectively, but by July 2020 had rebounded to 89.7% (95% CI = 79.6% to 101.1%) and 101.6% (95% CI = 93.8% to 110.1%) of the July 2019 prepandemic volumes, respectively. The year-to-date cumulative volume of screening and diagnostic mammograms performed through July 2020 was 66.2% (95% CI = 60.3% to 72.6%) and 79.9% (95% CI = 75.4% to 84.6%), respectively, of year-to-date volume through July 2019. Screening mammography rebound was similar across age groups and by family history of breast cancer. Monthly screening mammography volume in July 2020 for Black, White, Hispanic, and Asian women reached 96.7% (95% CI = 88.1% to 106.1%), 92.9% (95% CI = 82.9% to 104.0%), 72.7% (95% CI = 56.5% to 93.6%), and 51.3% (95% CI = 39.7% to 66.2%) of the July 2019 prepandemic volume, respectively.ConclusionsDespite a strong overall rebound in mammography volume by July 2020, the rebound lagged among Asian and Hispanic women, and a substantial cumulative deficit in missed mammograms accumulated, which may have important health consequences.

Background Since 2007, digital mammography and digital breast tomosynthesis (DBT) replaced screen-film mammography. Whether these technologic advances have improved diagnostic performance has, to the knowledge of the authors, not yet been established. Purpose To evaluate the performance and outcomes of surveillance mammography (digital mammography and DBT) performed from 2007 to 2016 in women with a personal history of breast cancer and compare with data from 1996 to 2007 and the performance of digital mammography screening benchmarks. Materials and Methods In this observational cohort study, five Breast Cancer Surveillance Consortium registries provided prospectively collected mammography data linked with tumor registry and pathologic outcomes. This study identified asymptomatic women with American Joint Committee on Cancer anatomic stages 0-III primary breast cancer who underwent surveillance mammography from 2007 to 2016. The primary outcome was a second breast cancer diagnosis within 1 year of mammography. Performance measures included the recall rate, cancer detection rate, interval cancer rate, positive predictive value of biopsy recommendation, sensitivity, and specificity. Results Among 32 331 women who underwent 117 971 surveillance mammographic examinations (112 269 digital mammographic examinations and 5702 DBT examinations), the mean age at initial diagnosis was 59 years ± 12 (standard deviation). Of 1418 second breast cancers diagnosed, 998 were surveillance-detected cancers and 420 were interval cancers. The recall rate was 8.8% (10 365 of 117 971; 95% CI: 8.6%, 9.0%), the cancer detection rate was 8.5 per 1000 examinations (998 of 117 971; 95% CI: 8.0, 9.0), the interval cancer rate was 3.6 per 1000 examinations (420 of 117 971; 95% CI: 3.2, 3.9), the positive predictive value of biopsy recommendation was 31.0% (998 of 3220; 95% CI: 29.4%, 32.7%), the sensitivity was 70.4% (998 of 1418; 95% CI: 67.9%, 72.7%), and the specificity was 98.1% (114 331 of 116 553; 95% CI: 98.0%, 98.2%). Compared with previously published studies, interval cancer rate was comparable with rates from 1996 to 2007 in women with a personal history of breast cancer and was higher than the published digital mammography screening benchmarks. Conclusion In transitioning from screen-film to digital mammography and digital breast tomosynthesis, surveillance mammography performance demonstrated minimal improvement over time and remained inferior to the performance of screening mammography benchmarks. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Moy and Gao in this issue.

BackgroundAdvanced breast cancer is an outcome used to evaluate screening effectiveness. The advanced cancer definition resulting in the best discrimination of breast cancer death has not been studied in a breast imaging population.MethodsA total of 52 496 women aged 40-79 years participating in the Breast Cancer Surveillance Consortium diagnosed with invasive cancer were staged using the 8th edition of American Joint Committee on Cancer (AJCC) anatomic and prognostic pathologic systems and Tomosynthesis Mammographic Imaging Screening Trial (TMIST) tumor categories. We calculated the area under the receiver operating characteristic curve for predicting 5-year breast cancer death and the sensitivity and specificity for predicting 5-year breast cancer death for 3 advanced cancer classifications: anatomic stage IIB or higher, prognostic pathologic stage IIA or higher, and TMIST advanced cancer.ResultsThe area under the receiver operating characteristic curves for predicting 5-year breast cancer death for AJCC anatomic stage, AJCC prognostic pathologic stage, and TMIST tumor categories were 0.826 (95% confidence interval [CI] = 0.817 to 0.835), 0.856 (95% CI = 0.846 to 0.866), and 0.789 (95% CI = 0.780 to 0.797), respectively. AJCC prognostic pathologic stage had statistically significantly better discrimination than AJCC anatomic stage (difference = 0.030, bootstrap 95% CI = 0.024 to 0.037) and TMIST tumor categories (difference = 0.067, bootstrap 95% CI = 0.059 to 0.075). The sensitivity and specificity for predicting 5-year breast cancer death for AJCC anatomic stage IIB or higher, AJCC prognostic pathologic stage IIA or higher, and TMIST advanced cancer were 72.6%, 76.7%, and 96.1%; and 78.9%, 81.6%, and 41.1%, respectively.ConclusionsDefining advanced cancer as AJCC prognostic pathologic stage IIA or higher most accurately predicts breast cancer death. Use of this definition by investigators will facilitate comparing breast cancer screening effectiveness studies.

ObjectiveWe sought to identify and characterize examinations in women with a personal history of breast cancer likely performed for asymptomatic surveillance.MethodsWe included surveillance mammograms (1997-2017) in asymptomatic women with a personal history of breast cancer diagnosed at age ≥18 years (1996-2016) from 103 Breast Cancer Surveillance Consortium facilities. We examined facility-level variability in examination indication. We modeled the relative risk (RR) and 95% confidence intervals (CIs) at the examination level of a (1) nonscreening indication and (2) surveillance interval ≤9 months using Poisson regression with fixed effects for facility, stage, diagnosis age, surgery, examination year, and time since diagnosis.ResultsAmong 244,855 surveillance mammograms, 69.5% were coded with a screening indication, 12.7% short-interval follow-up, and 15.3% as evaluation of a breast problem. Within a facility, the proportion of examinations with a screening indication ranged from 6% to 100% (median 86%, interquartile range 79%-92%). Facilities varied the most for examinations in the first 5 years after diagnosis, with 39.4% of surveillance mammograms having a nonscreening indication. Within a facility, breast conserving surgery compared with mastectomy (RR = 1.64; 95% CI = 1.60-1.68) and less time since diagnosis (1 year versus 5 years; RR = 1.69; 95% CI = 1.66-1.72; 3 years versus 5 years = 1.20; 95% CI = 1.18-1.23) were strongly associated with a nonscreening indication with similar results for ≤9-month surveillance interval. Screening indication and >9-month surveillance intervals were more common in more recent years.ConclusionVariability in surveillance indications across facilities in the United States supports including indications beyond screening in studies evaluating surveillance mammography effectiveness and demonstrates the need for standardization.

BackgroundNational patterns of breast imaging in women with a personal history of breast cancer (PHBC) are unknown making evaluation of annual surveillance recommendations a challenge.ObjectiveTo describe variation in use of mammography and breast magnetic resonance imaging (MRI) examinations beginning 6 months after diagnosis among women with PHBC in US community practice. We report on the breast imaging indication, imaging intervals, and time since breast cancer diagnosis by examination type.DesignLongitudinal study using cross-sectional data.SettingBreast Cancer Surveillance Consortium breast imaging facilities.Participants19,955 women diagnosed between 2005 and 2012 with AJCC stage 0-III incident breast cancer who had 69,386 mammograms and 3,553 breast MRI examinations from January 2005 to September 2013; median follow-up of 37.6 months (interquartile range, 22.1-60.7).Main measuresBreast imaging indication, imaging intervals, and time since breast cancer diagnosis by examination type.Key resultsAmong women with a PHBC who received breast imaging, 89.4% underwent mammography alone, 0.8% MRI alone, and 10.3% had both mammography and MRI. About half of mammograms and MRIs were indicated for surveillance vs. diagnostic, with an increase in the proportion of surveillance exams as time from diagnosis increased (mammograms, 45.7% at 1 year to 72.2% after 5 years; MRIs, 54.8% at 1 year to 78.6% after 5 years). In the first post-diagnosis period, 32.8% of women had > 2 breast imaging examinations and of these, 65.8% were less than 6 months apart. During the first 5-year post-diagnosis, the frequency of examinations per year decreased and the interval between examinations shifted towards annual examinations.ConclusionIn women with a PHBC who received post-diagnosis imaging, a third underwent multiple breast imaging examinations per year during the first 2-year post-diagnosis despite recommendations for annual exams. As time since diagnosis increases, imaging indication shifts from diagnostic to surveillance.

Background There is lack of consensus regarding the use of breast MRI for routine surveillance for second breast cancer events in women with a personal history of breast cancer. Purpose To compare performance of surveillance mammography with breast MRI. Materials and Methods This observational cohort study used prospectively collected data and included 13 266 women age 18 years and older (mean age, 60 years ± 13) with stage 0-III breast cancer who underwent 33 938 mammographic examinations and 2506 breast MRI examinations from 2005 to 2012 in the Breast Cancer Surveillance Consortium. Women were categorized into two groups: mammography alone (n = 11 745) or breast MRI (n = 1521). Performance measures were calculated by using end-of-day assessment and occurrence of second breast cancer events within 1 year of imaging. Logistic regression was used to compare performance for breast MRI versus mammography alone, adjusting for women, examination, and primary breast cancer characteristics. Analysis was conducted on a per-examination basis. Results Breast MRI was associated with younger age at diagnosis, chemotherapy, and higher education and income. Raw performance measures for breast MRI versus mammography were as follows, respectively: cancer detection rates, 10.8 (95% confidence interval [CI]: 6.7, 14.8) versus 8.2 (95% CI: 7.3, 9.2) per 1000 examinations; sensitivity, 61.4% (27 of 44; 95% CI: 46.5%, 76.2%) versus 70.3% (279 of 397; 95% CI: 65.8%, 74.8%); and biopsy rate, 10.1% (253 of 2506; 95% CI: 8.9%, 11.3%) versus 4.0% (1343 of 33 938; 95% CI: 3.7%, 4.2%). In multivariable models, breast MRI was associated with higher biopsy rate (odds ratio [OR], 2.2; 95% CI: 1.9, 2.7; P < .001) and cancer detection rate (OR, 1.7; 95% CI: 1.1, 2.7; P = .03) than mammography alone. However, there were no differences in sensitivity (OR, 1.1; 95% CI: 0.4, 2.9; P = .84) or interval cancer rate (OR, 1.1; 95% CI: 0.6, 2.2; P = .70). Conclusion Comparison of the performance of surveillance breast MRI with mammography must account for patient characteristics. Whereas breast MRI leads to higher biopsy and cancer detection rates, there were no significant differences in sensitivity or interval cancers compared with mammography. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Newell in this issue.

PurposeTo evaluate comparative associations of breast magnetic resonance imaging (MRI) background parenchymal enhancement (BPE) and mammographic breast density with subsequent breast cancer risk.Patients and methodsWe examined women undergoing breast MRI in the Breast Cancer Surveillance Consortium from 2005 to 2015 (with one exam in 2000) using qualitative BPE assessments of minimal, mild, moderate, or marked. Breast density was assessed on mammography performed within 5 years of MRI. Among women diagnosed with breast cancer, the first BPE assessment was included if it was more than 3 months before their first diagnosis. Breast cancer risk associated with BPE was estimated using Cox proportional hazards regression.ResultsAmong 4,247 women, 176 developed breast cancer (invasive, n = 129; ductal carcinoma in situ,n = 47) over a median follow-up time of 2.8 years. More women with cancer had mild, moderate, or marked BPE than women without cancer (80% v 66%, respectively). Compared with minimal BPE, increasing BPE levels were associated with significantly increased cancer risk (mild: hazard ratio [HR], 1.80; 95% CI, 1.12 to 2.87; moderate: HR, 2.42; 95% CI, 1.51 to 3.86; and marked: HR, 3.41; 95% CI, 2.05 to 5.66). Compared with women with minimal BPE and almost entirely fatty or scattered fibroglandular breast density, women with mild, moderate, or marked BPE demonstrated elevated cancer risk if they had almost entirely fatty or scattered fibroglandular breast density (HR, 2.30; 95% CI, 1.19 to 4.46) or heterogeneous or extremely dense breasts (HR, 2.61; 95% CI, 1.44 to 4.72), with no significant interaction (P = .82). Combined mild, moderate, and marked BPE demonstrated significantly increased risk of invasive cancer (HR, 2.73; 95% CI, 1.66 to 4.49) but not ductal carcinoma in situ (HR, 1.48; 95% CI, 0.72 to 3.05).ConclusionBPE is associated with future invasive breast cancer risk independent of breast density. BPE should be considered for risk prediction models for women undergoing breast MRI.

Purpose The aim of the current study was to characterize the risk of interval invasive second breast cancers within 5 years of primary breast cancer treatment. Methods We examined 65,084 surveillance mammograms from 18,366 women with a primary breast cancer diagnosis of unilateral ductal carcinoma in situ or stage I to III invasive breast carcinoma performed from 1996 to 2012 in the Breast Cancer Surveillance Consortium. Interval invasive breast cancer was defined as ipsilateral or contralateral cancer diagnosed within 1 year after a negative surveillance mammogram. Discrete-time survival models-adjusted for all covariates-were used to estimate the probability of interval invasive cancer, given the risk factors for each surveillance round, and aggregated across rounds to estimate the 5-year cumulative probability of interval invasive cancer. Results We observed 474 surveillance-detected cancers-334 invasive and 140 ductal carcinoma in situ-and 186 interval invasive cancers which yielded a cancer detection rate of 7.3 per 1,000 examinations (95% CI, 6.6 to 8.0) and an interval invasive cancer rate of 2.9 per 1,000 examinations (95% CI, 2.5 to 3.3). Median cumulative 5-year interval cancer risk was 1.4% (interquartile range, 0.8% to 2.3%; 10th to 90th percentile range, 0.5% to 3.7%), and 15% of women had ≥ 3% 5-year interval invasive cancer risk. Cumulative 5-year interval cancer risk was highest for women with estrogen receptor- and progesterone receptor-negative primary breast cancer (2.6%; 95% CI, 1.7% to 3.5%), interval cancer presentation at primary diagnosis (2.2%; 95% CI, 1.5% to 2.9%), and breast conservation without radiation (1.8%; 95% CI, 1.1% to 2.4%). Conclusion Risk of interval invasive second breast cancer varies across women and is influenced by characteristics that can be measured at initial diagnosis, treatment, and imaging. Risk prediction models that evaluate the risk of cancers not detected by surveillance mammography should be developed to inform discussions of tailored surveillance.

Purpose To compare screening magnetic resonance (MR) imaging performance in the Breast Cancer Surveillance Consortium (BCSC) with Breast Imaging Reporting and Data System (BI-RADS) benchmarks. Materials and Methods This study was approved by the institutional review board and compliant with HIPAA and included BCSC screening MR examinations collected between 2005 and 2013 from 5343 women (8387 MR examinations) linked to regional Surveillance, Epidemiology, and End Results program registries, state tumor registries, and pathologic information databases that identified breast cancer cases and tumor characteristics. Clinical, demographic, and imaging characteristics were assessed. Performance measures were calculated according to BI-RADS fifth edition and included cancer detection rate (CDR), positive predictive value of biopsy recommendation (PPV2), sensitivity, and specificity. Results The median patient age was 52 years; 52% of MR examinations were performed in women with a first-degree family history of breast cancer, 46% in women with a personal history of breast cancer, and 15% in women with both risk factors. Screening MR imaging depicted 146 cancers, and 35 interval cancers were identified (181 total-54 in situ, 125 invasive, and two status unknown). The CDR was 17 per 1000 screening examinations (95% confidence interval [CI]: 15, 20 per 1000 screening examinations; BI-RADS benchmark, 20-30 per 1000 screening examinations). PPV2 was 19% (95% CI: 16%, 22%; benchmark, 15%). Sensitivity was 81% (95% CI: 75%, 86%; benchmark, >80%), and specificity was 83% (95% CI: 82%, 84%; benchmark, 85%-90%). The median tumor size of invasive cancers was 10 mm; 88% were node negative. Conclusion The interpretative performance of screening MR imaging in the BCSC meets most BI-RADS benchmarks and approaches benchmark levels for remaining measures. Clinical practice performance data can inform ongoing benchmark development and help identify areas for quality improvement. © RSNA, 2017.

Purpose To establish performance benchmarks for modern screening digital mammography and assess performance trends over time in U.S. community practice. Materials and Methods This HIPAA-compliant, institutional review board-approved study measured the performance of digital screening mammography interpreted by 359 radiologists across 95 facilities in six Breast Cancer Surveillance Consortium (BCSC) registries. The study included 1 682 504 digital screening mammograms performed between 2007 and 2013 in 792 808 women. Performance measures were calculated according to the American College of Radiology Breast Imaging Reporting and Data System, 5th edition, and were compared with published benchmarks by the BCSC, the National Mammography Database, and performance recommendations by expert opinion. Benchmarks were derived from the distribution of performance metrics across radiologists and were presented as 50th (median), 10th, 25th, 75th, and 90th percentiles, with graphic presentations using smoothed curves. Results Mean screening performance measures were as follows: abnormal interpretation rate (AIR), 11.6 (95% confidence interval [CI]: 11.5, 11.6); cancers detected per 1000 screens, or cancer detection rate (CDR), 5.1 (95% CI: 5.0, 5.2); sensitivity, 86.9% (95% CI: 86.3%, 87.6%); specificity, 88.9% (95% CI: 88.8%, 88.9%); false-negative rate per 1000 screens, 0.8 (95% CI: 0.7, 0.8); positive predictive value (PPV) 1, 4.4% (95% CI: 4.3%, 4.5%); PPV2, 25.6% (95% CI: 25.1%, 26.1%); PPV3, 28.6% (95% CI: 28.0%, 29.3%); cancers stage 0 or 1, 76.9%; minimal cancers, 57.7%; and node-negative invasive cancers, 79.4%. Recommended CDRs were achieved by 92.1% of radiologists in community practice, and 97.1% achieved recommended ranges for sensitivity. Only 59.0% of radiologists achieved recommended AIRs, and only 63.0% achieved recommended levels of specificity. Conclusion The majority of radiologists in the BCSC surpass cancer detection recommendations for screening mammography; however, AIRs continue to be higher than the recommended rate for almost half of radiologists interpreting screening mammograms. © RSNA, 2016 Online supplemental material is available for this article.

Purpose To establish contemporary performance benchmarks for diagnostic digital mammography with use of recent data from the Breast Cancer Surveillance Consortium (BCSC). Materials and Methods Institutional review board approval was obtained for active or passive consenting processes or to obtain a waiver of consent to enroll participants, link data, and perform analyses. Data were obtained from six BCSC registries (418 radiologists, 92 radiology facilities). Mammogram indication and assessments were prospectively collected for women undergoing diagnostic digital mammography and linked with cancer diagnoses from state cancer registries. The study included 401 548 examinations conducted from 2007 to 2013 in 265 360 women. Results Overall diagnostic performance measures were as follows: cancer detection rate, 34.7 per 1000 (95% confidence interval [CI]: 34.1, 35.2); abnormal interpretation rate, 12.6% (95% CI: 12.5%, 12.7%); positive predictive value (PPV) of a biopsy recommendation (PPV2), 27.5% (95% CI: 27.1%, 27.9%); PPV of biopsies performed (PPV3), 30.4% (95% CI: 29.9%, 30.9%); false-negative rate, 4.8 per 1000 (95% CI: 4.6, 5.0); sensitivity, 87.8% (95% CI: 87.3%, 88.4%); and specificity, 90.5% (95% CI: 90.4%, 90.6%). Among cancers detected, 63.4% were stage 0 or 1 cancers, 45.6% were minimal cancers, the mean size of invasive cancers was 21.2 mm, and 69.6% of invasive cancers were node negative. Performance metrics varied widely across diagnostic indications, with cancer detection rate (64.5 per 1000) and abnormal interpretation rate (18.7%) highest for diagnostic mammograms obtained to evaluate a breast problem with a lump. Compared with performance during the screen-film mammography era, diagnostic digital performance showed increased abnormal interpretation and cancer detection rates and decreasing PPVs, with less than 70% of radiologists within acceptable ranges for PPV2 and PPV3. Conclusion These performance measures can serve as national benchmarks that may help transform the marked variation in radiologists' diagnostic performance into targeted quality improvement efforts. © RSNA, 2017 Online supplemental material is available for this article.

ObjectiveThe purpose of this study was to evaluate concordance between BI-RADS assessments and management recommendations for breast MRI in community practice.Materials and methodsBreast MRI data were collected from four regional Breast Cancer Surveillance Consortium registries from 2005 to 2011 for women who were 18-79 years old. Assessments and recommendations were compared to determine concordance according to BI-RADS guidelines. Concordance was compared by assessment category as well as by year of examination and clinical indication.ResultsIn all, 8283 MRI examinations were included in the analysis. Concordance was highest (93% [2475/2657]) in examinations with a BI-RADS category 2 (benign) assessment. Concordance was also high in examinations with category 1 (negative) (87% [1669/1909]), category 0 (incomplete) (83% [348/417]), category 5 (highly suggestive of malignancy) (83% [208/252]), and category 4 (suspicious) (74% [734/993]) assessments. Examinations with categories 3 (probably benign) and 6 (known biopsy-proven malignancy) assessments had the lowest concordance rates (36% [302/837] and 56% [676/1218], respectively). The most frequent discordant recommendation for a category 3 assessment was routine follow-up. The most frequent discordant recommendation for a category 6 assessment was biopsy. Concordance of assessments and management recommendations differed across clinical indications (p < 0.0001), with the lowest concordance in examinations to assess disease extent.ConclusionBreast MRI BI-RADS management recommendations were most concordant for assessments of negative, incomplete, suspicious, and highly suggestive of malignancy. Lower concordance for assessments of probably benign and known biopsy-proven malignancy and for examinations performed to assess disease extent highlight areas for interventions to improve breast MRI reporting.

Background: Machine learning (ML) approaches facilitate risk prediction model development using high-dimensional predictors and higher-order interactions at the cost of model interpretability and transparency. We compared the relative predictive performance of statistical and ML models to guide modeling strategy selection for surveillance mammography outcomes in women with a personal history of breast cancer (PHBC). Methods: We cross-validated seven risk prediction models for two surveillance outcomes, failure (breast cancer within 12 months of a negative surveillance mammogram) and benefit (surveillance-detected breast cancer). We included 9,447 mammograms (495 failures, 1,414 benefits, and 7,538 nonevents) from years 1996 to 2017 using a 1:4 matched case-control samples of women with PHBC in the Breast Cancer Surveillance Consortium. We assessed model performance of conventional regression, regularized regressions (LASSO and elastic-net), and ML methods (random forests and gradient boosting machines) by evaluating their calibration and, among well-calibrated models, comparing the area under the receiver operating characteristic curve (AUC) and 95% confidence intervals (CI). Results: LASSO and elastic-net consistently provided well-calibrated predicted risks for surveillance failure and benefit. The AUCs of LASSO and elastic-net were both 0.63 (95% CI, 0.60-0.66) for surveillance failure and 0.66 (95% CI, 0.64-0.68) for surveillance benefit, the highest among well-calibrated models. Conclusions: For predicting breast cancer surveillance mammography outcomes, regularized regression outperformed other modeling approaches and balanced the trade-off between model flexibility and interpretability. Impact: Regularized regression may be preferred for developing risk prediction models in other contexts with rare outcomes, similar training sample sizes, and low-dimensional features. [ABSTRACT FROM AUTHOR]

Background: National patterns of breast imaging in women with a personal history of breast cancer (PHBC) are unknown making evaluation of annual surveillance recommendations a challenge.Objective: To describe variation in use of mammography and breast magnetic resonance imaging (MRI) examinations beginning 6 months after diagnosis among women with PHBC in US community practice. We report on the breast imaging indication, imaging intervals, and time since breast cancer diagnosis by examination type.Design: Longitudinal study using cross-sectional data.Setting: Breast Cancer Surveillance Consortium breast imaging facilities.Participants: 19,955 women diagnosed between 2005 and 2012 with AJCC stage 0-III incident breast cancer who had 69,386 mammograms and 3,553 breast MRI examinations from January 2005 to September 2013; median follow-up of 37.6 months (interquartile range, 22.1-60.7).Main Measures: Breast imaging indication, imaging intervals, and time since breast cancer diagnosis by examination type.Key Results: Among women with a PHBC who received breast imaging, 89.4% underwent mammography alone, 0.8% MRI alone, and 10.3% had both mammography and MRI. About half of mammograms and MRIs were indicated for surveillance vs. diagnostic, with an increase in the proportion of surveillance exams as time from diagnosis increased (mammograms, 45.7% at 1 year to 72.2% after 5 years; MRIs, 54.8% at 1 year to 78.6% after 5 years). In the first post-diagnosis period, 32.8% of women had > 2 breast imaging examinations and of these, 65.8% were less than 6 months apart. During the first 5-year post-diagnosis, the frequency of examinations per year decreased and the interval between examinations shifted towards annual examinations.Conclusion: In women with a PHBC who received post-diagnosis imaging, a third underwent multiple breast imaging examinations per year during the first 2-year post-diagnosis despite recommendations for annual exams. As time since diagnosis increases, imaging indication shifts from diagnostic to surveillance. [ABSTRACT FROM AUTHOR]

Background: Earlier detection of second breast cancers after primary breast cancer (PBC) treatment improves survival, yet mammography is less accurate in women with prior breast cancer. The purpose of this study was to examine women presenting clinically with second breast cancers after negative surveillance mammography (interval cancers), and to estimate the five-year risk of interval-invasive second cancers for women with varying risk profiles.Methods: We evaluated a prospective cohort of 15 114 women with 47 717 surveillance mammograms diagnosed with stage 0-II unilateral PBC from 1996 through 2008 at facilities in the Breast Cancer Surveillance Consortium. We used discrete time survival models to estimate the association between odds of an interval-invasive second breast cancer and candidate predictors, including demographic, PBC, and imaging characteristics. All statistical tests were two-sided.Results: The cumulative incidence of second breast cancers after five years was 54.4 per 1000 women, with 325 surveillance-detected and 138 interval-invasive second breast cancers. The five-year risk of interval-invasive second cancer for women with referent category characteristics was 0.60%. For women with the most and least favorable profiles, the five-year risk ranged from 0.07% to 6.11%. Multivariable modeling identified grade II PBC (odds ratio [OR] = 1.95, 95% confidence interval [CI] = 1.15 to 3.31), treatment with lumpectomy without radiation (OR = 3.27, 95% CI = 1.91 to 5.62), interval PBC presentation (OR = 2.01, 95% CI 1.28 to 3.16), and heterogeneously dense breasts on mammography (OR = 1.54, 95% CI = 1.01 to 2.36) as independent predictors of interval-invasive second breast cancers.Conclusions: PBC diagnosis and treatment characteristics contribute to variation in subsequent-interval second breast cancer risk. Consideration of these factors may be useful in developing tailored post-treatment imaging surveillance plans. [ABSTRACT FROM AUTHOR]