Radial Scars on Screening Digital Breast Tomosynthesis: Upstaging Rates and Management Strategies.

BACKGROUND. Radial scars are more commonly identified on digital breast tomosynthesis (DBT) than on digital mammography (DM). Nonetheless, universal guidelines for radial scar management in the current era of DBT are lacking. OBJECTIVE. The purpose of this study was to determine the upstaging rates of screening DBT-detected radial scars with and without atypia and to identify features related to upstaging risk. METHODS. This retrospective study included patients who underwent core needle biopsy (CNB) showing a radial scar after screening DBT and DM from January 1, 2013, to December 31, 2020. Patients without surgical excision or at least 2 years of imaging follow-up after CNB were excluded. Rates of upstaging to breast cancer (ductal carcinoma in situ [DCIS] or invasive disease) were compared between radial scars with and without atypia at CNB. Associations of upstaging with patient, imaging, and pathologic variables were explored using standard statistical tests. RESULTS. Of 165 women with 171 radial scars, the final study sample included 153 women (mean age, 56 years; range, 33-83 years) with 159 radial scars that underwent surgical excision (80.5%, 128/159) or at least 2 years of imaging follow-up (19.5%, 31/159). Seven radial scars were upstaged to DCIS and one to invasive disease. Therefore, the up-staging rate of radial scars to cancer was 5.0% (8/159). The upstaging rate of radial scars without atypia at CNB was 1.6% (2/129) and that of radial scars with atypia was 20.0% (6/30) (p < .001). On multivariable analysis, features associated with higher upstaging risk included a prior breast cancer diagnosis (62.5% vs 4.8%; p = .01) and the presence of atypia at CNB (75.0% vs 15.9%; p = .02). The upstaging rate according to mammographic finding type was 7.1% (1/14) for asymmetries, 12.5% (2/16) for masses, 5.3% (5/95) for architectural distortion, and 0.0% (0/34) for calcifications. CONCLUSION. Screening-detected radial scars without atypia at CNB have a low upstaging rate to breast cancer of 1.6%. CLINICAL IMPACT. Imaging surveillance rather than surgery is a reasonable approach for radial scars without atypia, particularly for those presenting as calcifications.

Performance Metrics of Screening Digital Breast Tomosynthesis Based on Years Since a Prior Breast Cancer Diagnosis.

BACKGROUND. Mammography surveillance protocols after breast cancer treatment vary widely. Some practices recommend performing diagnostic mammography for a certain number of years or indefinitely, whereas others recommend returning immediately to screening. OBJECTIVE. This study's objective was to determine performance metrics of screening digital breast tomosynthesis (DBT) in patients who resume screening mammography immediately after breast cancer treatment, based on the number of years since the breast cancer diagnosis. METHODS. This retrospective study included screening DBT examinations performed from January 2013 to June 2019 in patients who resumed screening mammography immediately after a prior breast cancer diagnosis. Multivariable logistic regression models with generalized estimating equations were used to evaluate associations between screening performance metrics and years since the prior breast cancer diagnosis, controlling for age, race and ethnicity, breast density, presence of a prior screening mammogram, and interpreting radiologist. RESULTS. The study included 8090 patients (mean age, 65 ± 11 [SD] years) with a prior breast cancer diagnosis who underwent 30,812 screening DBT examinations during the study period. The cancer detection rate (CDR) was 8.6 per 1000 examinations (265/30,812), abnormal interpretation rate (AIR) was 5.7% (1750/30,812), PPV1 was 15.1% (265/1750), sensitivity was 80.3% (265/330), specificity was 95.1% (28,997/30,482), and false-negative rate was 2.1 per 1000 examinations (65/30,812). CDR showed a significant independent positive association with years since breast cancer diagnosis (adjusted OR, 1.03; 95% CI, 1.01-1.05; p < .001), being lowest more than 2 to up to 3 years after diagnosis (4.9 per 1000 examinations) and highest more than 8 to up to 9 years after diagnosis (11.2 per 1000 examinations). AIR showed a significant independent negative association with years since breast cancer diagnosis (adjusted OR, 0.99; 95% CI, 0.98-1.00; p = .01), being highest 1 year or less after diagnosis (7.5%) and lowest more than 5 to up to 6 years after diagnosis (5.0%). CONCLUSION. Among 8090 patients with a prior breast cancer diagnosis, even though the AIR was higher during the year after diagnosis compared with subsequent years, the AIR remained acceptably low (< 10%) in all years. CLINICAL IMPACT. These results support the study institution's mammographic surveillance protocol for patients with a prior breast cancer diagnosis of returning immediately to DBT screening.

Artificial Intelligence for Breast Ultrasound: AJR Expert Panel Narrative Review.

Breast ultrasound is used in a wide variety of clinical scenarios, including both diagnostic and screening applications. Limitations of ultrasound, however, include its low specificity and, for automated breast ultrasound screening, the time necessary to review whole-breast ultrasound images. As of this writing, four AI tools that are approved or cleared by the FDA address these limitations. Current tools, which are intended to provide decision support for lesion classification and/or detection, have been shown to increase specificity among non-specialists and to decrease interpretation times. Potential future applications include triage of patients with palpable masses in low-resource settings, preoperative prediction of axillary lymph node metastasis, and preoperative prediction of neoadjuvant chemotherapy response. Challenges in the development and clinical deployment of AI for ultrasound include: the limited availability of curated training datasets compared to mammography; the high variability in ultrasound image acquisition due to equipment- and operator-related factors (which may limit algorithm generalizability); and the lack of post-implementation evaluation studies. Furthermore, current AI tools for lesion classification were developed based on 2D data, but diagnostic accuracy could potentially be improved if multimodal ultrasound data were used, such as color Doppler, elastography, cine clips, and 3D imaging.

Ipsilateral tumor recurrence risk in women with ductal carcinoma in situ: application of the Van Nuys Prognostic Index and the Memorial Sloan Kettering Cancer Center nomogram.

PURPOSE: To apply the Van Nuys Prognostic Index (VNPI) and the Memorial Sloan Kettering Cancer Center (MSKCC) ductal carcinoma in situ (DCIS) nomogram to DCIS patients with known long-term outcomes. METHODS: A retrospective review was performed of consecutive patients diagnosed with DCIS from 2007 to 2014. Included patients underwent breast-conserving surgery (BCS) and were followed with imaging for at least five years. For each patient, the VNPI and MSKCC nomogram risk estimates were determined. In addition, variables used in both models were compared between women with and without recurrences using the Wilcoxon signed-rank test and the Pearson's chi-squared test. RESULTS: Over the eight-year period, 456 women (average age 57 years, range 30-87) underwent BCS for DCIS. Thirty-one (6.8%) experienced an ipsilateral recurrence. The average VNPI scores were 7 (range 5-9) and 7 (range 4-10) for women with and without a recurrence (p = 0.14), respectively, with 4-6, 7-9, and 10-12 being the low, moderate, and high-risk groups, respectively. Per the MSKCC nomogram, the average five-year recurrence risks were 5% (range 1-12%) and 4% (range 1-38%) for women with and without a recurrence (p = 0.09), respectively. The recurrence risk-related variables were younger patient age, need for one or more re-excision surgeries, and use of endocrine therapy for 0 to less than five years after surgery. CONCLUSION: Ipsilateral tumor recurrence risk estimates based on the VNPI and MSKCC nomogram are similar between women with DCIS who did and did not have a recurrence, suggesting that more robust prognostic models are needed.

Evidence-Based Pragmatic Approach to the Management of Borderline or High-Risk Breast Lesions.

The management of borderline or high-risk breast lesions (HRLs) remains controversial. We propose a pragmatic evidence-based approach based on lesion type. For lobular carcinoma in situ with pleomorphism, papilloma with atypia, atypical ductal hyperplasia, and fibroepithelial lesions, surgical consultation and excision are recommended. Patients with other borderline or HRLs are referred for surgical consultation to discuss excision versus surveillance. Our recommendations align with American Society of Breast Surgeons guidelines, which aim to reduce unnecessary surgery and ensure appropriate follow-up.

Management of Architectural Distortion on Digital Breast Tomosynthesis With Nonmalignant Pathology at Biopsy.

BACKGROUND. Digital breast tomosynthesis (DBT) has led to increased detection and biopsy of architectural distortion, which may yield malignancy, radial scar, or other benign pathologies. Management of nonmalignant architectural distortion on DBT remains controversial. OBJECTIVE. The purpose of this study was to determine upgrade rates of architectural distortion on DBT from nonmalignant pathology at biopsy to malignancy at surgery. METHODS. This retrospective study included cases of mammographically detected architectural distortion from July 1, 2016, to June 30, 2019, that were nonmalignant at image-guided needle biopsy and underwent surgical excision. Mammographic examinations included digital 2D mammography and DBT. Imaging data were extracted from radiology reports. Upgrade rates were summarized using descriptive statistics. Features of upgraded and nonupgraded cases were compared using Pearson chi-square test and Wilcoxon signed rank test. RESULTS. The study included 129 cases of architectural distortion with nonmalignant pathology at biopsy that underwent excision in 125 women (mean age, 54 years; range, 23-90 years). At biopsy, 92 (71.3%) were radial scars and 37 (28.7%) were other nonmalignant pathologies. Of 66 radial scars without atypia at biopsy, one (1.5%) was upgraded to ductal carcinoma in situ (DCIS) at surgery and none to invasive cancer. Of 24 benign pathologies without atypia at biopsy, one was considered discordant. Of the 23 remaining concordant cases, one (4.3%) was upgraded to DCIS at surgery and none to invasive cancer. The overall upgrade rate to cancer of architectural distortion with concordant nonmalignant pathology at biopsy was 10.2% (13/128). The upgrade rate to cancer of architectural distortion without atypia was 2.2% (2/89) and with atypia was 28.2% (11/39). Explored features (age, personal or family breast cancer history, presentation by screening vs diagnostic mammography, breast density, associated mammographic findings, presence and size of ultrasound correlate, biopsy modality) showed no signifi-cant associations with upgrade risk (p > .05). CONCLUSION. Architectural distortion on DBT with concordant nonmalignant pathology at biopsy has an overall upgrade rate to malignancy at surgery of 10.2%. Architectural distortion without atypia has a low upgrade rate of 2.2%. CLINICAL IMPACT. Imaging surveillance can be considered for architectural distortion on DBT yielding radial scar without atypia or other concordant benign pathologies without atypia at biopsy.

Artificial Intelligence (AI) for Screening Mammography, From the AJR Special Series on AI Applications.

Artificial intelligence (AI) applications for screening mammography are being marketed for clinical use in the interpretative domains of lesion detection and diagnosis, triage, and breast density assessment and in the noninterpretive domains of breast cancer risk assessment, image quality control, image acquisition, and dose reduction. Evidence in support of these nascent applications, particularly for lesion detection and diagnosis, is largely based on multireader studies with cancer-enriched datasets rather than rigorous clinical evaluation aligned with the application's specific intended clinical use. This article reviews commercial AI algorithms for screening mammography that are currently available for clinical practice, their use, and evidence supporting their performance. Clinical implementation considerations, such as workflow integration, governance, and ethical issues, are also described. In addition, the future of AI for screening mammography is discussed, including the development of interpretive and noninterpretive AI applications and strategic priorities for research and development.

Risk factors for an advanced breast cancer diagnosis within 2 years of a negative mammogram.

BACKGROUND: Identifying women at risk for advanced interval cancers would allow better targeting of mammography and supplemental screening. The authors assessed risk factors for advanced breast cancer within 2 years of a negative mammogram. METHODS: The authors included 293,520 negative mammograms performed from 2006 to 2015 among 74,736 women. Breast cancers were defined as advanced if they were >2 cm, were >1 cm and triple-negative or human epidermal growth factor receptor 2-positive, had positive lymph nodes, or were metastatic. Cox proportional hazards modeling was used to evaluate associations of age, breast density, menopause, mammogram type, prior breast biopsy, body mass index (BMI), and a family history of breast cancer with a cancer diagnosis within 2 years of a negative mammogram. Models were stratified by year since a negative mammogram. RESULTS: Among 1345 breast cancers, 357 were advanced (26.5%), and 988 (73.5%) were at an early stage. Breast density, prior biopsy, and family history were associated with an increased risk of both advanced and early-stage cancers. Overweight and obese women had a 40% higher risk of early-stage cancer only in year 2 (overweight hazard ratio [HR], 1.41; 95% confidence interval [CI], 1.19-1.67; P < .001; obese HR, 1.41; 95% CI, 1.17-1.70; P < .001). Obese women had a 90% increased risk of advanced cancer in year 1 (HR, 1.90; 95% CI, 1.14-3.18; P = .014), and both overweight and obese women had a 40% or greater increased risk in year 2 (overweight HR, 1.55; 95% CI, 1.14-2.07; P = .005; obese HR, 1.42; 95% CI, 1.00-2.01; P = .051). CONCLUSIONS: A higher BMI was associated with an advanced breast cancer diagnosis within 2 years of a negative mammogram. These results have important implications for risk assessment, screening intervals, and use of supplemental screening.

Imaging Surveillance of Breast Cancer Survivors with Digital Mammography versus Digital Breast Tomosynthesis.

Background Among breast cancer survivors, detecting a breast cancer when it is asymptomatic (rather than symptomatic) improves survival; thus, imaging surveillance in these patients is warranted. Digital breast tomosynthesis (DBT) is used for screening, but data on DBT for surveillance in this high-risk population are limited. Purpose To determine whether DBT leads to improved screening performance metrics when compared with two-dimensional digital mammography among breast cancer survivors. Materials and Methods In this study, screening mammograms obtained in breast cancer survivors before and after DBT implementation were retrospectively reviewed (March 2008-February 2011 for the digital mammography group; January 2013-December 2017 for the DBT group). Mammograms were interpreted by breast imaging radiologists with the assistance of computer-aided detection. Performance metrics and tumor characteristics between the groups were compared using multivariable logistic regression models. Results The digital mammography and DBT groups were composed of 9019 and 22 887 mammographic examinations, respectively, in 8170 women (mean age, 62 years ± 12 [standard deviation]). In the DBT group, the abnormal interpretation rate was lower (5.8% [1331 of 22 887 examinations] vs 6.2% [563 of 9019 examinations]; odds ratio [OR], 0.80; 95% CI: 0.71, 0.91; P = .001) and specificity was higher (95.0% [21 502 of 22 644 examinations] vs 94.7% [8424 of 8891 examinations]; OR, 1.23; 95% CI: 1.07, 1.41; P = .003) than in the digital mammography group. The cancer detection rates did not differ (8.3 per 1000 examinations with DBT vs 10.6 with digital mammography; OR, 0.76; 95% CI: 0.57, 1.02; P = .07). The proportions of screening-detected invasive cancers, versus in situ cancers, were similar (74% [140 of 189 cancers] in the DBT group vs 72% [69 of 96 cancers] in the digital mammography group; P = .69). Of 86 interval cancers, 58% (50 of 86 cancers) manifested with symptoms, and 33% (28 of 86 cancers) were detected at screening MRI. Conclusion Among breast cancer survivors, screening with digital breast tomosynthesis led to fewer false-positive results and higher specificity but did not affect cancer detection. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Hooley and Butler in this issue.

Predictors of Reexcision following Breast-Conserving Surgery for Ductal Carcinoma In Situ.

BACKGROUND: Reexcision following breast-conserving surgery (BCS) in women with ductal carcinoma in situ (DCIS) results in adjuvant treatment delays, higher health care costs, and undesirable cosmetic outcomes. The purpose of this study is to determine patient, imaging, pathological, and surgical predictors of reexcision following BCS for DCIS. PATIENTS AND METHODS: A retrospective review of women with DCIS who had BCS from 2007 to 2016 was conducted. Patient, imaging, pathological, and surgical features, in addition to surgical outcomes, were collected from medical records. Standard statistical tests were used to compare features between patients who did and did not undergo at least one reexcision. A multivariable logistic regression model was fit to assess features associated with reexcision. RESULTS: A total of 547 women (mean age 59 years; range 30-88 years) diagnosed with DCIS at core needle biopsy underwent BCS. Of all women, 31.6% (173/547) had at least one reexcision. With multivariable analysis, features associated with reexcision included younger patient age (adjusted odds ratio [aOR] 0.98, 95% confidence interval [CI] 0.97-1.0, p = 0.049), African-American race (aOR 2.66, 95% CI 1.13-6.26, p = 0.03), biopsy modality of ultrasound (aOR 2.35, 95% CI 1.22-4.53, p = 0.01), and earlier year of surgery (aOR 0.92, 95% CI 0.86-0.98, p = 0.01). No pathological features of DCIS were associated with reexcision risk. CONCLUSIONS: In our cohort of nearly 550 women with DCIS who underwent BCS, 31.6% had at least one reexcision. Features associated with reexcision include younger patient age, African-American race, biopsy modality of ultrasound, and earlier year of surgery.

Breast Cancer Screening with Digital Breast Tomosynthesis: Are Initial Benefits Sustained?

Background Performance metrics with digital breast tomosynthesis (DBT) are based on early experiences. There is limited research on whether the benefits of DBT are sustained. Purpose To determine whether improved screening performance metrics with DBT are sustained over time at the population level and after the first screening round at the individual level. Materials and Methods A retrospective review was conducted of screening mammograms that had been obtained before DBT implementation (March 2008 to February 2011, two-dimensional digital mammography [DM] group) and for 5 years after implementation (January 2013 to December 2017, DBT1-DBT5 groups, respectively). Patients who underwent DBT were also categorized according to the number of previous DBT examinations they had undergone. Performance metrics were compared between DM and DBT groups and between patients with no previous DBT examinations and those with at least one prior DBT examination by using multivariable logistic regression models. Results The DM group consisted of 99 582 DM examinations in 55 086 women (mean age, 57.3 years ± 11.6 [standard deviation]). The DBT group consisted of 205 048 examinations in 76 276 women (mean age, 58.2 years ± 11.2). There were no differences in the cancer detection rate (CDR) between DM and DBT groups (4.6-5.8 per 1000 examinations, P = .08 to P = .95). The highest CDR was observed with a woman's first DBT examination (6.1 per 1000 examinations vs 4.4-5.7 per 1000 examinations with at least one prior DBT examination, P = .001 to P = .054). Compared with the DM group, the DBT1 group had a lower abnormal interpretation rate (AIR) (adjusted odds ratio [AOR], 0.85; P < .001), which remained reduced in the DBT2, DBT3, and DBT5 groups (P < .001 to P = .02). The reduction in AIR was also sustained after the first examination (P < .001 to P = .002). Compared with the DM group, the DBT1 group had a higher specificity (AOR, 1.20; P < .001), which remained increased in DBT2, DBT3, and DBT5 groups (P < .001 to P = .004). The increase in specificity was also sustained after the first examination (P < .001 to P = .01). Conclusion The benefits of reduced false-positive examinations and higher specificity with screening tomosynthesis were sustained after the first screening round at the individual level. © RSNA, 2020 See also the editorial by Taourel in this issue.

Do Eligibility Criteria for Ductal Carcinoma In Situ (DCIS) Active Surveillance Trials Identify Patients at Low Risk for Upgrade to Invasive Carcinoma?

BACKGROUND: Clinical trials are currently ongoing to determine the safety and efficacy of active surveillance (AS) versus usual care (surgical and radiation treatment) for women with ductal carcinoma in situ (DCIS). This study aimed to determine upgrade rates of DCIS at needle biopsy to invasive carcinoma at surgery among women who meet the eligibility criteria for AS trials. METHODS: A retrospective review was performed of consecutive women at an academic medical center with a diagnosis of DCIS at needle biopsy from 2007 to 2016. Medical records were reviewed for mode of presentation, imaging findings, biopsy pathology results, and surgical outcomes. Each patient with DCIS was evaluated for AS trial eligibility based on published criteria for the COMET, LORD, and LORIS trials. RESULTS: During a 10-year period, DCIS was diagnosed in 858 women (mean age 58 years; range 28-89 years). Of the 858 women, 498 (58%) were eligible for the COMET trial, 101 (11.8%) for the LORD trial, and 343 (40%) for the LORIS trial. The rates of upgrade to invasive carcinoma were 12% (60/498) for the COMET trial, 5% (5/101) for the LORD trial, and 11.1% (38/343) for the LORIS trial. The invasive carcinomas ranged from 0.2 to 20 mm, and all were node-negative. CONCLUSIONS: Women who meet the eligibility criteria for DCIS AS trials remain at risk for occult invasive carcinoma at presentation, with upgrade rates ranging from 5 to 12%. These findings suggest that more precise criteria are needed to ensure that women with invasive carcinoma are excluded from AS trials.

Ductal carcinoma in situ on digital mammography versus digital breast tomosynthesis: rates and predictors of pathologic upgrade.

OBJECTIVES: To compare upgrade rates of ductal carcinoma in situ (DCIS) on digital mammography (DM) versus digital breast tomosynthesis (DBT) and identify patient, imaging, and pathological features associated with upgrade risk. METHODS: A retrospective review was performed of 318 women (mean 59 years, range 37-89) with screening-detected DCIS from 2007 to 2011 (DM group) and from 2013 to 2016 (DBT group). Comparisons made between DM and DBT groups using the unpaired t test and chi-square test include detection rates of DCIS, upgrade rates to invasive cancer, and pathological features of DCIS and upgraded cases. Patient, imaging, and pathological features associated with upgrade were also determined. P values < 0.05 were considered significant. RESULTS: There was no significant difference in detection rates of DCIS between DM and DBT groups (0.9 versus 1.0 per 1000 examinations, p = 0.45). Upgrade rates of DCIS to invasive cancer in DM and DBT groups were similar (17.3% versus 16.8%, p = 0.90), despite significant differences in pathological features of DCIS between DM and DBT groups (including nuclear grade, comedonecrosis, and progesterone receptor status [p ≤ 0.01]). Among upgraded cases, a higher proportion were high-grade invasive cancers with DBT (36.7% versus 9.5%, p = 0.03). In both groups, ultrasound-guided (versus stereotactic) biopsy was associated with higher upgrade risk (p ≤ 0.03). CONCLUSIONS: There was no significant difference in detection rates or upgrade rates of DCIS on DM versus DBT; however, upgraded cases were more likely to be high grade with DBT, suggesting possible differences in tumor biology between cancers with DM and DBT. In both DM and DBT groups, biopsy modality was associated with upgrade risk. KEY POINTS: • Detection rates and upgrade rates of ductal carcinoma in situ (DCIS) on digital mammography (DM) versus digital breast tomosynthesis (DBT) are similar. • A higher proportion of upgraded cases were high-grade invasive cancers with DBT than DM, suggesting possible differences in tumor biology between cancers that are detected with DM and DBT. • With both DM and DBT, ultrasound-guided biopsy (versus stereotactic biopsy) was associated with a higher risk of upgrade.

Digital 2D versus Tomosynthesis Screening Mammography among Women Aged 65 and Older in the United States.

Background Although breast cancer incidence and mortality rates increase with advancing age, there are limited data on the benefits and risks of screening mammography in older women and on the performance of two-dimensional digital mammography (DM) and digital breast tomosynthesis (DBT) in older women. Purpose To compare performance metrics of DM and DBT among women aged 65 years and older. Materials and Methods For this retrospective study, consecutive screening mammograms in patients aged 65 years and older from March 2008 to February 2011 (DM group) and from January 2013 to December 2015 (DBT group) were reviewed. Cancer detection rate, abnormal interpretation rate, positive predictive values, sensitivity, and specificity were calculated. Multivariable logistic regression models were fit to compare performance metrics in the DM versus DBT groups. Results The DM group had 15 019 women (mean age ± standard deviation, 72.7 years ± 6.3), and the DBT group had 20 646 women (mean age, 72.1 years ± 5.9). After adjusting for multiple variables, there was no difference in cancer detection rate between the DM and DBT groups (6.9 vs 8.2 per 1000 examinations; adjusted odds ratio [AOR], 1.13; P = .23). Compared with the DM group, the DBT group had a lower abnormal interpretation rate (5.7% vs 5.8%; AOR, 0.88; P < .001), higher positive predictive value 1 (14.5% vs 11.9%; AOR, 1.26; P = .03), and higher specificity (95.1% vs 94.8%; AOR, 1.18; P < .001). The DBT group had a higher proportion of invasive cancers relative to in situ cancers (81.1% vs 74.4%; P = .06) and fewer node-positive cancers (10.2% vs 16.6%; P = .054) than did the DM group. Conclusion In women aged 65 years and older, integration of digital breast tomosynthesis led to improved performance metrics, with a lower abnormal interpretation rate, higher positive predictive value 1, and higher specificity. © RSNA, 2019 See also the editorial by Philpotts and Durand in this issue.

Comparison of Upright Digital Breast Tomosynthesis-guided versus Prone Stereotactic Vacuum-assisted Breast Biopsy.

Purpose To compare the performance of upright digital breast tomosynthesis (DBT)-guided vacuum-assisted breast biopsy (VABB) with that of prone stereotactic (PS) VABB. Materials and Methods This retrospective review involved consecutive patients who underwent PS VABB from August 2014 to December 2015 and upright DBT-guided VABB from February 2016 to June 2017. Tissue sampling methods were the same for PS and DBT-guided biopsies. Wilcoxon and Pearson χ2 tests were used to compare the groups. Results During the study period, 439 PS VABBs in 408 patients (mean age, 56.5 years; age range, 32-84 years) and 706 DBT-guided VABBs in 682 patients (mean age, 57.9 years; age range, 23-90 years) were recommended. Technical success was achieved for more lesions with DBT-guided VABB versus PS VABB (99.3% [695 of 700] vs 95.1% [410 of 431], respectively; P < .001). Mean procedure time was shorter with DBT-guided VABB versus PS VABB (12 vs 27 minutes, respectively; P < .001), and fewer exposures were acquired with DBT-guided VABB versus PS VABB (three vs 12, respectively; P < .001). A higher percentage of lesions for which DBT-guided VABB was performed were noncalcified lesions (eg, architectural distortion, asymmetry, and mass) than for PS VABB (29.2% [203 of 695] vs 3.4% [14 of 410], respectively; P < .001). There were no differences in the distribution of histologic results (P = .42). No major complications were observed in either group. Conclusion Upright digital breast tomosynthesis-guided vacuum-assisted breast biopsy has a higher rate of technical success than does prone stereotactic vacuum-assisted biopsy and can be performed in less than half the time and with one-fourth of the exposures. In addition, more architectural distortions and asymmetries are amenable to biopsy with digital breast tomosynthesis-guided vacuum-assisted breast biopsy. © RSNA, 2018.

Mammographic Breast Density Assessment Using Deep Learning: Clinical Implementation.

Purpose To develop a deep learning (DL) algorithm to assess mammographic breast density. Materials and Methods In this retrospective study, a deep convolutional neural network was trained to assess Breast Imaging Reporting and Data System (BI-RADS) breast density based on the original interpretation by an experienced radiologist of 41 479 digital screening mammograms obtained in 27 684 women from January 2009 to May 2011. The resulting algorithm was tested on a held-out test set of 8677 mammograms in 5741 women. In addition, five radiologists performed a reader study on 500 mammograms randomly selected from the test set. Finally, the algorithm was implemented in routine clinical practice, where eight radiologists reviewed 10 763 consecutive mammograms assessed with the model. Agreement on BI-RADS category for the DL model and for three sets of readings-(a) radiologists in the test set, (b) radiologists working in consensus in the reader study set, and (c) radiologists in the clinical implementation set-were estimated with linear-weighted κ statistics and were compared across 5000 bootstrap samples to assess significance. Results The DL model showed good agreement with radiologists in the test set (κ = 0.67; 95% confidence interval [CI]: 0.66, 0.68) and with radiologists in consensus in the reader study set (κ = 0.78; 95% CI: 0.73, 0.82). There was very good agreement (κ = 0.85; 95% CI: 0.84, 0.86) with radiologists in the clinical implementation set; for binary categorization of dense or nondense breasts, 10 149 of 10 763 (94%; 95% CI: 94%, 95%) DL assessments were accepted by the interpreting radiologist. Conclusion This DL model can be used to assess mammographic breast density at the level of an experienced mammographer. © RSNA, 2018 Online supplemental material is available for this article . See also the editorial by Chan and Helvie in this issue.

Comparison of performance metrics with digital 2D versus tomosynthesis mammography in the diagnostic setting.

OBJECTIVES: To compare performance metrics between digital 2D mammography (DM) and digital breast tomosynthesis (DBT) in the diagnostic setting. METHODS: Consecutive diagnostic examinations from August 2008 to February 2011 (DM group) and from January 2013 to July 2015 (DM/DBT group) were reviewed. Core biopsy and surgical pathology results within 365 days after the mammogram were collected. Performance metrics, including cancer detection rate (CDR), abnormal interpretation rate (AIR), positive predictive value (PPV) 2, PPV3, sensitivity, and specificity were calculated. Multivariable logistic regression models were fit to compare performance metrics in the DM and DM/DBT groups while adjusting for clinical covariates. RESULTS: A total of 22,883 mammograms were performed before DBT integration (DM group), and 22,824 mammograms were performed after complete DBT integration (DM/DBT group). After adjusting for multiple variables, the CDR was similar in both groups (38.2 per 1,000 examinations in the DM/DBT group versus 31.3 per 1,000 examinations in the DM group, p = 0.14); however, a higher proportion of cancers were invasive rather than in situ in the DM/DBT group [83.7% (731/873) versus 72.3% (518/716), p < 0.01]. The AIR was lower in the DM/DBT group (p < 0.01), and PPV2, PPV3, and specificity were higher in the DM/DBT group (all p = 0.01 or p < 0.01). CONCLUSIONS: Complete integration of DBT into the diagnostic setting is associated with improved diagnostic performance. Increased utilization of DBT may thus result in better patient outcomes and lead to a shift in the benchmarks that have been established for DM. KEY POINTS: • Integration of tomosynthesis into the diagnostic setting is associated with improved performance. • A higher proportion of cancers are invasive rather than in situ with digital breast tomosynthesis. • Increased utilization of tomosynthesis may lead to a shift in established benchmarks.

Use of ChatGPT to Assign BI-RADS Assessment Categories to Breast Imaging Reports.

The large language model GPT-4 showed limited utility in generating BI-RADS assessment categories for factitious breast imaging reports containing findings and impression sections, with frequent incorrect BI-RADS category assignments and poor reproducibility in assigned BI-RADS categories across independent tests for the same report using the same prompt.