Frequency and Cancer Yield of BI-RADS Category 3 Lesions Detected at High-Risk Screening Breast MRI.

OBJECTIVE. The purpose of this study was to evaluate the frequency and cancer yield of BI-RADS category 3 lesions in baseline versus nonbaseline (those with at least one prior) MRI screening examinations. MATERIALS AND METHODS. Consecutive MRI screening examinations performed from 2011 through 2015 were reviewed. Pearson and Wilcoxon tests were used to examine differences in age, breast density, screening indication, background parenchymal enhancement, and cancer yield between baseline and nonbaseline MRI BI-RADS category 3 assessments. Multivariate logistic regression models based on generalized estimating equations were used to assess the odds of receiving a BI-RADS 3 assessment as a function of the variables. RESULTS. Of 6672 MRI screening examinations of 3214 patients, 202 examinations (3%) were assessed BI-RADS category 3. Among baseline examinations, 8% (82/983) were assessed BI-RADS 3, compared with 2% (120/5689) of nonbaseline examinations (p < 0.001). Among the total BI-RADS 3 examinations, 6% (13/202) yielded malignancy of the lesion that had been assessed BI-RADS 3; 12 of 13 cancers were stage 0 or I at diagnosis. The cancer yield of BI-RADS 3 at baseline examinations was 2% (2/82), compared with 9% (11/120) for nonbaseline examinations (p = 0.056). Ten of 13 examinations were upgraded at or before 6-month follow-up MRI. CONCLUSION. Baseline screening breast MRI examinations are associated with a significantly higher rate of BI-RADS category 3 assessments than are nonbaseline examinations. Most cancers diagnosed at follow-up of BI-RADS 3 lesions are in an early stage and are diagnosed at or before the 6-month follow-up examination. When used judiciously, short-interval follow-up MRI is an appropriate method for identifying early-stage breast cancer while avoiding unnecessary biopsies with benign findings.

Performance of Screening Breast MRI across Women with Different Elevated Breast Cancer Risk Indications.

Background Screening breast MRI is recommended for women with BRCA mutation or a history of chest radiation, but guidelines are equivocal for MRI screening of women with a personal history of breast cancer or high-risk lesion. Purpose To evaluate screening breast MRI performance across women with different elevated breast cancer risk indications. Materials and Methods All screening breast MRI examinations performed between 2011 and 2014 underwent retrospective medical record review. Indications for screening were as follows: BRCA mutation carrier or history of chest radiation (BRCA/RT group), family history of breast cancer (FH group), personal history of breast cancer (PH group), and history of high-risk lesion (HRL group). Screening performance metrics were calculated and compared among indications by using logistic regression adjusted for age, available prior MRI, mammographic density, examination year, and multiple risk factors. Results There were 5170 screening examinations in 2637 women (mean age, 52 years; range, 23-86 years); 67 breast cancers were detected. The cancer detection rate (CDR) was highest in the BRCA/RT group (26 per 1000 examinations; 95% confidence interval [CI]: 16, 43 per 1000 examinations), intermediate for those in the PH and HRL groups (12 per 1000 examinations [95% CI: 9, 17 per 1000 examinations] and 15 per 1000 examinations [95% CI: 7, 32 per 1000 examinations], respectively), and lowest for those in the FH group (8 per 1000 examinations; 95% CI: 4, 14 per 1000 examinations). No difference in CDR was evident for the PH or HRL group compared with the BRCA/RT group (P = .14 and .18, respectively). The CDR was lower for the FH group compared with the BRCA/RT group (P = .02). No difference was evident in positive predictive value for biopsies performed (PPV3) for the BRCA/RT group (41%; 95% CI: 26%, 56%) compared with the PH (41%; 95% CI: 31%, 52%; P = .63) or HRL (36%, 95% CI: 17%, 60%; P = .37) groups. PPV3 was lower for the FH group (14%; 95% CI: 8%, 25%; P = .048). Conclusion Screening breast MRI should be considered for women with a personal history of breast cancer or high-risk lesion. Worse screening MRI performance in patients with a family history of breast cancer suggests that better risk assessment strategies may benefit these women. © RSNA, 2019.

Effect of Background Parenchymal Enhancement on Cancer Risk Across Different High-Risk Patient Populations Undergoing Screening Breast MRI.

OBJECTIVE: The purpose of this study is to evaluate the effect of background parenchymal enhancement (BPE) on breast cancer risk across different high-risk patient populations undergoing screening breast MRI. MATERIALS AND METHODS: Consecutive screening breast MRI examinations performed between 2011 and 2014 were reviewed. Multivariate logistic regression using generalized estimating equations was used to assess the association of the following variables with cancer risk: age, qualitative BPE prospective clinical assessment (minimal or mild vs moderate or marked), mammographic breast density (dense vs not dense), and screening indication (prioritized in the following order: BRCA carrier or history of thoracic radiation, breast cancer personal history, high-risk lesion, and breast cancer family history). Cancer diagnosis was defined as a tissue diagnosis of invasive or in situ carcinoma within 12 months of the screening MRI. RESULTS: The study cohort included 4686 screening MRI examinations performed in 2446 women, grouped by BPE as minimal or mild (3975/4686; 85%) versus moderate or marked (711/4686; 15%) and by screening indication as BRCA carrier or history of thoracic radiation (548/4686; 12%), breast cancer personal history (2541/4686; 54%), high-risk lesion (362/4686; 8%), and breast cancer family history (1235/4686; 26%). After adjustment for confounding variables, only BPE and screening indication were independent predictors of cancer diagnosis (p = 0.02 and p < 0.01, respectively). The odds ratio for developing cancer in the moderate or marked BPE group compared with the minimal or mild group was 2.1 (95% CI, 1.1-4.0), after adjusting for age, breast density, and screening indication. CONCLUSION: Increased BPE level is an independent predictor of breast cancer in women undergoing screening MRI for different high-risk indications.

Proposing New RadLex Terms by Analyzing Free-Text Mammography Reports.

After years of development, the RadLex terminology contains a large set of controlled terms for the radiology domain, but gaps still exist. We developed a data-driven approach to discover new terms for RadLex by mining a large corpus of radiology reports using natural language processing (NLP) methods. Our system, developed for mammography, discovers new candidate terms by analyzing noun phrases in free-text reports to extend the mammography part of RadLex. Our NLP system extracts noun phrases from free-text mammography reports and classifies these noun phrases as "Has Candidate RadLex Term" or "Does Not Have Candidate RadLex Term." We tested the performance of our algorithm using 100 free-text mammography reports. An expert radiologist determined the true positive and true negative RadLex candidate terms. We calculated precision/positive predictive value and recall/sensitivity metrics to judge the system's performance. Finally, to identify new candidate terms for enhancing RadLex, we applied our NLP method to 270,540 free-text mammography reports obtained from three academic institutions. Our method demonstrated precision/positive predictive value of 0.77 (159/206 terms) and a recall/sensitivity of 0.94 (159/170 terms). The overall accuracy of the system is 0.80 (235/293 terms). When we ran our system on the set of 270,540 reports, it found 31,800 unique noun phrases that are potential candidates for RadLex. Our data-driven approach to mining radiology reports can identify new candidate terms for expanding the breast imaging lexicon portion of RadLex and may be a useful approach for discovering new candidate terms from other radiology domains.

Clinical Utility of Breast MRI in the Diagnosis of Malignancy After Inconclusive or Equivocal Mammographic Diagnostic Evaluation.

OBJECTIVE: The purpose of this study was to determine the clinical utility of breast MRI for diagnosing malignancy in women with equivocal mammographic findings but no symptoms. MATERIALS AND METHODS: Retrospective review of an institutional MRI database of 7332 contrast-enhanced breast MRI examinations from January 1, 2009, through December 31, 2012, yielded the records of 296 (4.0%) examinations of 294 women without symptoms who underwent MRI for mammographic findings uncertain at diagnostic evaluation. Imaging findings, histopathologic results, and patient demographics were obtained from the electronic medical record. RESULTS: The mean patient age was 55 years (range, 29-83 years). Mammographic lesion type (n = 294) included 89 focal asymmetries, 76 asymmetries, 64 masses, 44 architectural distortions, 17 surgical scar versus lesion, and four miscellaneous lesions. Diagnostic ultrasound, performed on 286 of 294 (97.3%) lesions at mammographic evaluation, showed an ultrasound correlate in 37 (12.9%) lesions, equivocal correlate in 48 (16.8%), and no ultrasound correlate in 201 (70.3%). MRI examination of 294 index lesions showed a correlate in 133 (45.2%) and no correlate in 161 (54.8%). Forty of 294 (13.6%) index lesions were malignant, 37 (92.5%) with an MRI correlate and three (7.5%) without an MRI correlate. Among 250 patients who underwent biopsy or had 2 or more years of imaging stability, the sensitivity, specificity, negative predictive value, and positive predictive value of breast MRI for malignancy were 92.5%, 62.4%, 97.8%, and 31.9%. Forty-four of 294 (15.0%) patients had lesions incidentally found at MRI; 7 of 41 (17.1%) lesions that were biopsied or were stable for at least 1 year were malignant. CONCLUSION: Problem-solving breast MRI for inconclusive mammographic findings helps identify malignancies with high sensitivity and a high negative predictive value.

Characteristics, Malignancy Rate, and Follow-up of BI-RADS Category 3 Lesions Identified at Breast MR Imaging: Implications for MR Image Interpretation and Management.

Purpose To (a) evaluate the frequency of Breast Imaging Reporting and Data System (BI-RADS) category 3 assessment in screening and diagnostic breast magnetic resonance (MR) imaging, (b) review findings considered indicative of BI-RADS category 3, and (c) determine outcomes of BI-RADS category 3 lesions, including upgrades, downgrades, and malignancy rates. Materials and Methods This retrospective study was approved by the institutional review board and compliant with HIPAA. The authors retrospectively reviewed the breast MR imaging database (2009-2011) to identify breast MR images classified as showing BI-RADS category 3 lesions. There were 9216 BI-RADS assessments in 5778 examinations (3360 women). Of the 9216 assessments, 567 (6%) in 483 women (average age, 47.2 years; median age, 47.0 years) were assigned BI-RADS category 3. In women with more than one BI-RADS category 3 lesion, the first lesion reported in the impression was used for data analysis. Outcomes data were available for 435 of the 483 women (90.1%). These women comprised the study cohort. Medical records from January 1, 2009, to May 31, 2015, were reviewed to obtain demographic characteristics and outcomes. χ(2) statistics and 95% exact confidence intervals (CIs) were constructed. Results MR imaging was performed for high-risk screening in 240 of the 435 patients (55.2%) and for diagnostic purposes in 195 (44.8%). Findings included mass (n = 125, 28.7%), focus (n = 111, 25.5%), nonmass enhancement (n = 80, 18.3%), moderate or marked background parenchymal enhancement (BPE) (n = 91, 20.9%), posttreatment changes (n = 16, 3.8%), and other findings (n = 12, 2.8%). Outcomes were as follows: 339 of the 435 patients (78%) did not have evidence of malignancy at more than 24 months, 28 (6.4%) underwent mastectomy (all benign), and 68 (15.6%) had lesion upgrades, with 11 cancers (2.5%). All 11 cancers were diagnosed in women with a genetic mutation or a personal history of breast cancer. No cancer was detected in cases of moderate or marked BPE. Conclusion Six percent of all breast MR imaging assessments were categorized as BI-RADS category 3, with a cancer rate of 2.5% (95% CI: 1.3%, 4.5%). All cancers were in women with a genetic mutation or personal history of breast cancer. Marked BPE does not necessitate a BI-RADS 3 assessment. (©) RSNA, 2016.

Breast MR Imaging for Equivocal Mammographic Findings: Help or Hindrance?

Breast magnetic resonance (MR) imaging, because of its extremely high sensitivity in detecting invasive breast cancers, is sometimes used as a diagnostic tool to evaluate equivocal mammographic findings. However, breast MR imaging should never substitute for a complete diagnostic evaluation or for biopsy of suspected, localizable suspicious mammographic lesions, whenever possible. The modality's high cost, in addition to only moderate specificity, mandate that radiologists use it sparingly and with discrimination for problematic mammographic findings. It is rare that the reality or significance of a noncalcified mammographic finding remains equivocal or problematic at diagnostic mammography evaluation, which usually includes targeted ultrasonography (US). There are several reasons for this infrequent occurrence: (a) an asymmetry may persist on diagnostic views but be visible only on craniocaudal or mediolateral oblique projections, precluding three-dimensional localization for US or biopsy, or a lesion may persist on some diagnostic spot views but dissipate or efface on others; (b) uncertainty may exist as to whether apparent change is clinically important or owing to technical factors such as compression or positioning differences; or (c) a lesion may be suspected but biopsy options are limited owing to lack of a US correlate and lesion inaccessibility for stereotactic biopsy, or biopsy of a vague or questionably real lesion has been attempted unsuccessfully. This article will discuss the indications for problem-solving MR imaging for equivocal mammographic findings, present cases illustrating appropriate and inappropriate uses of problem-solving MR imaging, and present false-positive and false-negative cases affecting the specificity of breast MR imaging. (©)RSNA, 2016.

Automated extraction of BI-RADS final assessment categories from radiology reports with natural language processing.

The objective of this study is to evaluate a natural language processing (NLP) algorithm that determines American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) final assessment categories from radiology reports. This HIPAA-compliant study was granted institutional review board approval with waiver of informed consent. This cross-sectional study involved 1,165 breast imaging reports in the electronic medical record (EMR) from a tertiary care academic breast imaging center from 2009. Reports included screening mammography, diagnostic mammography, breast ultrasound, combined diagnostic mammography and breast ultrasound, and breast magnetic resonance imaging studies. Over 220 reports were included from each study type. The recall (sensitivity) and precision (positive predictive value) of a NLP algorithm to collect BI-RADS final assessment categories stated in the report final text was evaluated against a manual human review standard reference. For all breast imaging reports, the NLP algorithm demonstrated a recall of 100.0 % (95 % confidence interval (CI), 99.7, 100.0 %) and a precision of 96.6 % (95 % CI, 95.4, 97.5 %) for correct identification of BI-RADS final assessment categories. The NLP algorithm demonstrated high recall and precision for extraction of BI-RADS final assessment categories from the free text of breast imaging reports. NLP may provide an accurate, scalable data extraction mechanism from reports within EMRs to create databases to track breast imaging performance measures and facilitate optimal breast cancer population management strategies.

The Adoption and Impact on Performance of an Automated Outcomes Feedback Application for Tomosynthesis Screening Mammography.

OBJECTIVE: To evaluate a tomosynthesis screening mammography automated outcomes feedback application's adoption and impact on performance. METHODS: This prospective intervention study evaluated a feedback application that provided mammographers subsequent imaging and pathology results for patients that radiologists had personally recalled from screening. Deployed to 13 academic and 5 private practice attending radiologists, adoption was studied from March 29, 2018, to March 20, 2019. Radiologists indicated if reviewed feedback would influence future clinical decisions. For a subset of eight academic radiologists consistently interpreting screening mammograms during the study, performance metrics were compared pre-intervention (January 1, 2016, to September 30, 2017) and post-intervention (October 1, 2017 to June 30, 2018). Abnormal interpretation rate, positive predictive value of biopsies performed, sensitivity, specificity, and cancer detection rate were compared using Pearson's χ2 test. Logistic regression models were fit, adjusting for age, race, breast density, prior comparison, breast cancer history, and radiologist. RESULTS: The 18 radiologists reviewed 68.5% (1,398 of 2,042) of available feedback cases and indicated that 17.4% of cases (243 of 1,398) could influence future decisions. For the eight academic radiologist subset, after multivariable adjustment with comparison to pre-intervention, average abnormal interpretation rate decreased (from 7.5% to 6.7%, adjusted odds ratio [aOR] 0.86, P < .01), positive predictive value of biopsies performed increased (from 40.6% to 51.3%, aOR 1.48, P = .011), and specificity increased (from 93.0% to 93.9%, aOR 1.17, P < .01) post-intervention. There was no difference in cancer detection rate per 1,000 examinations (from 5.8 to 6.1, aOR 1.01, P = .91) or sensitivity (from 81.2% to 78.7%, aOR 0.84, P = .47). CONCLUSIONS: Radiologists used a screening mammography automated outcomes feedback application. Its use decreased false-positive examinations, without evidence of reduced cancer detection.

Impact of Background Parenchymal Enhancement on Diagnostic Performance in Screening Breast MRI.

RATIONALE AND OBJECTIVES: To evaluate the impact of background parenchymal enhancement (BPE) on diagnostic performance in screening breast magnetic resonance imaging (MRI). MATERIALS AND METHODS: Consecutive screening breast MRIs performed at our institution from 2011 to 2014 were reviewed in a HIPAA-compliant manner with institutional review board approval. BPE was extracted from radiology reports and examinations grouped into minimal/mild (lower) or moderate/marked (higher) BPE. Performance measures were compared between the two groups with Pearson's χ2 test and with logistic regression to adjust for possible confounders of age, screening indication, mammographic density, available prior MRI, and examination year, using lower BPE as the reference group. RESULTS: For 4686 screening MRIs performed in 2446 women, BPE was reported as minimal or mild for 3975 (85%) examinations and moderate or marked for 711(15%). Following logistic regression to adjust for multiple confounders, abnormal interpretation rate (AIR) significantly differed between the two BPE groups. AIR was 13% (89/711) in the higher BPE group versus 7% (295/3975) in the lower BPE group with an adjusted odds ratio of 1.37 (95% confidence interval: 1.03, 1.82). After adjustment, all other performance metrics, including cancer detection rate, positive predictive value, sensitivity, and specificity did not significantly differ between the two BPE groups (P > 0.05). CONCLUSION: Higher BPE on screening MRI is associated with higher abnormal interpretation rate, with no impact on cancer detection rate, sensitivity, or specificity.

The Effect of Prior Comparison MRI on Interpretive Performance of Screening Breast MRI.

OBJECTIVE: To evaluate the effect of prior comparison MRI on interpretive performance of screening breast MRI. METHODS: After institutional review board approval, all screening breast MRI examinations performed from January 2011 through December 2014 were retrospectively reviewed. Screening performance metrics were estimated and compared for exams with and without a prior comparison MRI, using logistic regression models to adjust for age and screening indication (BRCA mutation or thoracic radiation versus breast cancer history versus high-risk lesion history versus breast cancer family history). RESULTS: Most exams, 4509 (87%), had a prior comparison MRI (incidence round), while 661 (13%) did not (prevalence round). Abnormal interpretation rate (6% vs 20%, P < 0.01), biopsy rate (3% vs 9%, P < 0.01), and false-positive biopsy recommendation rate per 1000 exams (21 vs 71, P < 0.01) were significantly lower in the incidence rounds compared to the prevalence rounds, while specificity was significantly higher (95% vs 81%, P < 0.01). There was no difference in cancer detection rate (CDR) per 1000 exams (12 vs 20, P = 0.1), positive predictive value of biopsies performed (PPV3) (35% vs 23%, P = 0.1), or sensitivity (86% vs 76%, P = 0.4). CONCLUSION: Presence of a prior comparison significantly improves incidence round screening breast MRI examination performance compared with prevalence round screening. Consideration should be given to updating the BI-RADS breast MRI screening benchmarks and auditing prevalence and incidence round examinations separately.

Performance of Breast Cancer Risk-Assessment Models in a Large Mammography Cohort.

BACKGROUND: Several breast cancer risk-assessment models exist. Few studies have evaluated predictive accuracy of multiple models in large screening populations. METHODS: We evaluated the performance of the BRCAPRO, Gail, Claus, Breast Cancer Surveillance Consortium (BCSC), and Tyrer-Cuzick models in predicting risk of breast cancer over 6 years among 35 921 women aged 40-84 years who underwent mammography screening at Newton-Wellesley Hospital from 2007 to 2009. We assessed model discrimination using the area under the receiver operating characteristic curve (AUC) and assessed calibration by comparing the ratio of observed-to-expected (O/E) cases. We calculated the square root of the Brier score and positive and negative predictive values of each model. RESULTS: Our results confirmed the good calibration and comparable moderate discrimination of the BRCAPRO, Gail, Tyrer-Cuzick, and BCSC models. The Gail model had slightly better O/E ratio and AUC (O/E = 0.98, 95% confidence interval [CI] = 0.91 to 1.06, AUC = 0.64, 95% CI = 0.61 to 0.65) compared with BRCAPRO (O/E = 0.94, 95% CI = 0.88 to 1.02, AUC = 0.61, 95% CI = 0.59 to 0.63) and Tyrer-Cuzick (version 8, O/E = 0.84, 95% CI = 0.79 to 0.91, AUC = 0.62, 95% 0.60 to 0.64) in the full study population, and the BCSC model had the highest AUC among women with available breast density information (O/E = 0.97, 95% CI = 0.89 to 1.05, AUC = 0.64, 95% CI = 0.62 to 0.66). All models had poorer predictive accuracy for human epidermal growth factor receptor 2 positive and triple-negative breast cancers than hormone receptor positive human epidermal growth factor receptor 2 negative breast cancers. CONCLUSIONS: In a large cohort of patients undergoing mammography screening, existing risk prediction models had similar, moderate predictive accuracy and good calibration overall. Models that incorporate additional genetic and nongenetic risk factors and estimate risk of tumor subtypes may further improve breast cancer risk prediction.

Metastatic Disease to the Breast From Extramammary Malignancies: A Multimodality Pictorial Review.

This pictorial review demonstrates imaging features of extramammary malignancies metastatic to the breast seen with multiple modalities, including mammography, ultrasound, computed tomography (CT), positron emission tomography, and magnetic resonance imaging. Although rare, metastases to the breast may have a distinct imaging appearance from the appearance of primary breast cancers. They are important to identify because they can mimic benign breast disease and their treatment differs from that of primary breast cancer. Metastatic disease to the breast most commonly appears as a single round or oval mass with circumscribed margins. Sonographically it is usually hypoechoic, and with CT or magnetic resonance imaging it usually enhances. In contrast with primary breast cancer, breast metastases do not demonstrate spiculated margins and rarely have associated calcifications. A variety of clinical presentations of breast metastases are reviewed, including presentation with a palpable mass, detection at screening mammography, and detection with CT or positron emission tomography.

Screening Breast MRI in Patients Previously Treated for Breast Cancer: Diagnostic Yield for Cancer and Abnormal Interpretation Rate.

RATIONALE AND OBJECTIVES: To determine the cancer detection rate and abnormal interpretation rate of screening breast magnetic resonance imaging (MRI) in previously treated breast cancer patients. MATERIALS AND METHODS: Institutional review board-approved retrospective review of the breast MRI database from 2009 to 2011 identified a total of 3297 screening examinations. After excluding genetic mutation carriers, untested first-degree relatives of known mutation carriers, and patients with a history of chest irradiation, there were 1194 (36.2%) examinations in 691 patients previously treated for breast cancer. MRI reports were reviewed to determine MRI findings and breast imaging reporting and data system (BI-RADS) assessments. The longitudinal medical record was reviewed to determine patient demographics and outcomes of imaging surveillance and biopsy. RESULTS: Mean patient age at initial cancer diagnosis was 46.1 years, and mean patient age during the study interval was 52 years. Cancer detection rate was 10 per 1000 (1%; 95% confidence interval [CI], 0.5%-1.8%]; 12 of 1194 examinations). Overall 10.7% (128 of 1194) of examinations received an abnormal interpretation, including 5.4% (65 of 1194) BI-RADS 4 or 5 and 5.3% (63 of 1194) BI-RADS 3 assessments with a 9.4% positive predictive value (PPV1; 12 of 128 examinations) and a 17.9% PPV3 (12 malignancies per 67 biopsies). CONCLUSIONS: Screening breast MRI in women previously treated for breast cancer detected cancer in 1.0% of examinations, with a 10.7% abnormal interpretation rate, and a PPV for malignancy of 17.9%.