Breast MRI Background Parenchymal Enhancement Categorization Using Deep Learning: Outperforming the Radiologist.

BACKGROUND: Background parenchymal enhancement (BPE) is assessed on breast MRI reports as mandated by the Breast Imaging Reporting and Data System (BI-RADS) but is prone to inter and intrareader variation. Semiautomated and fully automated BPE assessment tools have been developed but none has surpassed radiologist BPE designations. PURPOSE: To develop a deep learning model for automated BPE classification and to compare its performance with current standard-of-care radiology report BPE designations. STUDY TYPE: Retrospective. POPULATION: Consecutive high-risk patients (i.e. >20% lifetime risk of breast cancer) who underwent contrast-enhanced screening breast MRI from October 2013 to January 2019. The study included 5224 breast MRIs, divided into 3998 training, 444 validation, and 782 testing exams. On radiology reports, 1286 exams were categorized as high BPE (i.e., marked or moderate) and 3938 as low BPE (i.e., mild or minimal). FIELD STRENGTH/SEQUENCE: A 1.5 T or 3 T system; one precontrast and three postcontrast phases of fat-saturated T1-weighted dynamic contrast-enhanced imaging. ASSESSMENT: Breast MRIs were used to develop two deep learning models (Slab artificial intelligence (AI); maximum intensity projection [MIP] AI) for BPE categorization using radiology report BPE labels. Models were tested on a heldout test sets using radiology report BPE and three-reader averaged consensus as the reference standards. STATISTICAL TESTS: Model performance was assessed using receiver operating characteristic curve analysis. Associations between high BPE and BI-RADS assessments were evaluated using McNemar's chi-square test (α* = 0.025). RESULTS: The Slab AI model significantly outperformed the MIP AI model across the full test set (area under the curve of 0.84 vs. 0.79) using the radiology report reference standard. Using three-reader consensus BPE labels reference standard, our AI model significantly outperformed radiology report BPE labels. Finally, the AI model was significantly more likely than the radiologist to assign "high BPE" to suspicious breast MRIs and significantly less likely than the radiologist to assign "high BPE" to negative breast MRIs. DATA CONCLUSION: Fully automated BPE assessments for breast MRIs could be more accurate than BPE assessments from radiology reports. LEVEL OF EVIDENCE: 4 TECHNICAL EFFICACY STAGE: 3.

Yield of flow cytometry in addition to cytology for lymph node sampling in patients with incidental axillary adenopathy without a concurrent diagnosis of primary breast malignancy.

PURPOSE: Non-specific lymphadenopathy is increasingly being reported especially given the COVID-19 vaccination campaign and is a diagnostic dilemma especially in oncology patients. The purpose of this study was to evaluate the diagnostic accuracy and discordance rate between fine-needle aspiration (FNA) cytology and flow cytometry (FC) immunophenotyping in axillary FNA in patients with morphologically abnormal axillary lymph nodes on imaging and no concurrent diagnosis of primary breast malignancy. METHODS: This retrospective study included 222 patients who underwent screening or diagnostic axillary ultrasound that yielded suspicious lymphadenopathy without concurrent or recent prior diagnosis of breast cancer and who had subsequent image-guided axillary FNA and FC. Diagnostic accuracy, sensitivity, specificity, and positive and negative predictive value (PPV and NPV) were reported for FNA with cytology alone, and FC alone, and in combination. Discordance rate between FNA cytology and FC was assessed. Discordant cases were evaluated with histology or clinical and imaging follow-up. RESULTS: Diagnostic sensitivity, specificity, PPV, NPV, and diagnostic accuracy were 88%, 92%, 77%, 96%, and 91%, for FNA alone, 98%, 98%, 92%, 99%, and 98% for FC alone, and 100%, 92%, 79%, 100%, and 94% when combined. The overall discordance rate between FNA and FC was 7% (16/222). 7/16 (44%) patients with discordant results were diagnosed with lymphoma, while 9/16 (56%) patients with discordant results had benign findings. CONCLUSION: With a diagnostic accuracy of 91%, FNA with cytology is sufficient to screen patients with indeterminate and incidental lymphadenopathy. Flow cytometry could be initially deferred in patients with low pretest probability of lymphoma.

Utility of Targeted Ultrasound to Predict Malignancy Among Lesions Detected on Contrast-Enhanced Digital Mammography.

BACKGROUND. Targeted ultrasound (US) can be performed to characterize and potentially biopsy areas of enhancement detected on contrast-enhanced mammography (CEM). OBJECTIVE. The purpose of this study was to assess the utility of targeted US in predicting malignancy of lesions with indeterminate or suspicious enhancement on CEM. METHODS. One thousand consecutive CEM examinations with same-day targeted breast US at one institution between October 2013 and May 2018 were retrospectively reviewed. All patients with indeterminate or suspicious enhancement detected on CEM that underwent US evaluation were included. Patients with palpable or symptomatic lesions, those with suspicious findings on low-energy mammograms or images obtained with another modality, and those with less than 1 year of follow-up were excluded. Medical records, imaging, and pathology data were reviewed. Histopathologic analysis was used as the reference standard for biopsied lesions, and follow-up imaging was used for unbiopsied lesions. Associations between pathologic diagnosis, presence of a US correlate, and lesion characteristics were assessed by Fisher exact, chi-square, and Wilcox-on rank sum tests. RESULTS. Among 153 enhancing lesions detected on CEM in 144 patients, 47 (31%) had a US correlate. The frequency of a correlate between CEM and US was significantly higher among enhancing masses (28/43 [65%]) than among lesions exhibiting nonmass enhancement (19/110 [17%]) (p < .001). The likelihood of malignancy was significantly greater among lesions with a US correlate (12/47 [26%]) than among those without a US correlate (11/106 [10%]) (p = .03), and among mass lesions (11/43 [26%]) than among nonmass lesions (12/110 [11%]) (p = .04). The PPV of US-guided biopsy after CEM-directed US was 32%. CONCLUSION. Enhancing CEM-detected lesions that have a US correlate are more likely to be malignant and can be evaluated with US-guided biopsy to obviate additional breast MRI. CLINICAL IMPACT. CEM-directed US of enhancing lesions is useful given that lesions with a US correlate are more likely to be malignant and can be used as targets for US-guided biopsy until a CEM biopsy system becomes commercially available.

Palpable Breast Findings in High-risk Patients: Are Self- and Clinical Breast Exams Worthwhile?

OBJECTIVE: To assess breast imaging findings, biopsy rates, and malignancy rates in areas of palpable concern in women at high risk for breast cancer. METHODS: An IRB-approved retrospective review of a tertiary cancer center's breast imaging database was performed. Breast imaging and electronic medical records of high-risk women with palpable findings detected on self- or clinical breast examination from January 1, 2010, to January 1, 2016, were reviewed. Descriptive statistical analyses were conducted. RESULTS: Imaging correlates for 322 palpable findings in 238 high-risk women included 55/203 (27.1%) on mammography, 183/302 (60.6%) on US, and 20/47 (42.6%) on MRI. Biopsies were performed for 104/322 (32.3%) palpable findings: 95/104 (91.3%) under imaging guidance and 9/104 (8.7%) under palpation after negative imaging. Of 322 palpable findings, 16 (5.0%) were malignant in 16/238 (6.7%) women, yielding a positive predictive value of biopsy of 16.8% (95% CI: 9.2%-24%). Women diagnosed with cancer had 16/16 (100%) sonographic, 9/14 (64.3%) mammographic, and 7/7 (100%) MRI correlates. Cancer histopathology included 12 invasive ductal carcinomas, 1 ductal carcinoma in situ, 1 invasive lobular carcinoma, 1 malignant phyllodes tumor, and 1 metastatic carcinoid tumor. Over two years of follow-up imaging in 183/238 (76.9%) women were reviewed; 7/183 (3.8%) were diagnosed with breast cancer at least one year after presenting with a palpable concern in a different location. CONCLUSION: High-risk women with palpable findings exhibit a 6.7% malignancy rate, indicating the value of imaging workup in this population. In our cohort, imaging demonstrated a high negative predictive value.

Primary lymphoma of the breast: A report of two cases.

Primary breast lymphoma (PBL) should be distinguished from secondary breast lymphoma arising in the setting of lymphoma elsewhere in the body. Multimodality imaging is key to diagnosing PBL, and imaging manifestations thereof may indicate PBL and alter the treatment course. Treatment options including chemotherapy, radiation therapy, and/or surgery depend on histology. We report two cases of PBL, illustrating the transformative impact that multimodality imaging may have on clinical management.

Performance of Dual-Energy Contrast-enhanced Digital Mammography for Screening Women at Increased Risk of Breast Cancer.

BackgroundContrast agent-enhanced digital mammography (CEDM) has been shown to be more sensitive and specific than two-dimensional full-field digital mammography in the diagnostic setting. Few studies have reported on its performance in the screening setting.PurposeTo evaluate the performance of CEDM for breast cancer screening.Materials and MethodsThis retrospective study included women who underwent dual-energy CEDM for breast cancer screening from December 2012 through April 2016. Medical records were reviewed for age, risk factors, short-interval follow-up and biopsies recommended, and cancers detected. Sensitivity, specificity, positive predictive value of abnormal findings at screening (PPV1), positive predictive value of biopsy performed (PPV3), and negative predictive value were determined.ResultsIn the study period 904 baseline CEDMs were performed. Mean age was 51.8 years ± 9.4 (standard deviation). Of 904 patients, 700 (77.4%) had dense breasts, 247 (27.3%) had a family history of breast cancer in a first-degree relative age 50 years or younger, and 363 (40.2%) a personal history of breast cancer. The final Breast Imaging Reporting and Data System score was 1 or 2 in 832 of 904 (92.0%) patients, score of 3 in 25 of 904 (2.8%) patients, and score of 4 or 5 in 47 of 904 (5.2%) patients. By using CEDM, 15 cancers were diagnosed in 14 of 904 women (cancer detection rate, 15.5 of 1000). PPV3 was 29.4% (15 of 51). At least 1-year follow up was available in 858 women. There were two interval cancers. Sensitivity was 50.0% (eight of 16; 95% confidence interval [CI]: 24.7%, 75.3%) on the low-energy images compared with 87.5% (14 of 16; 95% CI: 61.7%, 98.4%) for the entire study (low-energy and iodine images; P = .03). Specificity was 93.7% (789 of 842; 95% CI: 91.8%, 95.2%); PPV1 was 20.9% (14 of 67; 95% CI: 11.9%, 32.6%), and negative predictive value was 99.7% (789 of 791; 95% CI: 99.09%, 99.97%).ConclusionContrast-enhanced digital mammography is a promising technique for screening women with higher-than-average risk for breast cancer.© RSNA, 2019.

Stereotactic breast biopsy efficiency: Does a pre-biopsy grid image help?

OBJECTIVE: Prior to stereotactic breast biopsy, some radiologists obtain a mammogram image with an overlying alphanumeric grid to mark the skin overlying the target. Our purpose is to determine if this grid image affects stereotactic biopsy efficiency and accuracy, including total images obtained, procedure time and need for retargeting. MATERIALS AND METHODS: IRB approved, HIPAA compliant retrospective review of prone stereotactic biopsy cases targeting calcifications 9/1/2015 to 9/1/2016 was performed. Images and reports were reviewed for number and type of images obtained, evidence of retargeting and biopsy table time. Attending radiologist, technologist and trainee involvement were recorded. Statistical analysis was performed utilizing SAS statistical software v 9.4 (SAS Institute, Cary, NC). RESULTS: Of 463 women (avg age 58.0 years, range 30-94), 392/463 (84.7%) had grid images obtained pre-biopsy. Grid patients had more images total than non-grid (avg 9.26 versus 8.44 images/patient; p < 0.0001) but spent less time on the biopsy table (avg 15 min 2 s versus 16 min 44 s/procedure; p < 0.0001). Non-grid patients were more likely to undergo initial retargeting (45% non-grid vs 30% of grid patients; p = 0.013); however, later retargeting after needle placement was comparable (p = 0.3). CONCLUSION: Grid imaging increases images obtained but decreases retargeting and biopsy table time at the expense of mammogram room/technologist time to obtain the grid image. The overall result is longer total procedure time (grid time plus table time) for the patient/technologist. A grid image therefore has limited usefulness and should be used judiciously in cases where prone positioning is challenging to patients.

The Impact of Patient-Initiated Subspecialty Review on Patient Care.

PURPOSE: To determine the effect of subspecialty review of breast imaging on patients without a diagnosis of breast cancer who self-referred for a second opinion. METHODS: Institutional review board-approved retrospective review was performed of 415 breast imaging studies submitted to our cancer center for second-opinion review by 245 patients in 2014, excluding cases without follow-up or change in original BI-RADS 0 assessment. One hundred forty-seven patients with 176 lesions were included. Original and second-opinion interpretations and recommendations were compared with histopathology or follow-up imaging. RESULTS: Ninety-six of 176 (55%) lesions were reported as suspicious in the original interpretation. Second-opinion review was discordant with the original interpretation in 82 of 176 (47%) lesions. Second-opinion review downgraded 24 of 96 (25%) lesions originally reported as suspicious to benign or probably benign, averting biopsy of these lesions. Second-opinion review upgraded 31 of 80 (39%) lesions originally reported as benign or probably benign to suspicious. A total of 87 lesions were biopsied yielding malignancy in 23 (26%) biopsies. Twenty-eight of 87 (32%) biopsies performed were recommended after second-opinion review, with 8 of 28 (29%) of these biopsies yielding cancer. Eight of 23 (35%) cancers detected represented malignancy not initially detected in the original interpretation. CONCLUSION: Second-opinion review is valuable in patients pursuing a breast imaging specialist's opinion, even before they are diagnosed with breast cancer. Second-opinion review disagreed with the original interpretation for 47% of lesions, averted 25% of originally recommended biopsies, and detected cancer in 29% of additional biopsies recommended. Thirty-five percent of cancers diagnosed after second-opinion review were not initially detected in the original interpretation.

Second opinion interpretation of breast ultrasound images-Is it worth another look?

PURPOSE: To determine if real-time breast ultrasound (RTUR) after second opinion reinterpretation of submitted static ultrasound images at a comprehensive cancer center impacts clinical management, specifically by detecting additional cancer and preventing unnecessary biopsy. MATERIALS AND METHODS: In this IRB-approved and HIPAA-compliant retrospective study, 209 patients were included who had breast ultrasound studies from outside facilities submitted for second opinion review between January 2013 and May 2014, and who subsequently underwent RTUR at our institution within three months of the outside study. Findings on submitted exams were compared with those on RTUR and disagreements between them were annotated to indicate the presence or absence of suspicious lesions and recommendation for biopsy. Changes in management were defined as any additional biopsies performed or biopsies averted after RTUR and reported as frequencies and percentages using 95% confidence intervals. RESULTS: Following RTUR, 49 additional biopsies were performed in 43/209 patients (20.6%; 95% CI 15.1-26.1%). Additional cancer was found in 12/49 (24.5%) biopsies in 11/209 patients (5.3%; 95% CI 2.2-8.2%). Forty biopsies in 31/209 (14.8%; CI 10.0-19.7%) patients originally recommended were canceled after RTUR. Overall, a change in management after RTUR was observed in 68/209 patients (32.5%; 95% CI 26.1-38.9%), including patients with either additional biopsies performed or biopsies averted. CONCLUSION: RTUR was found to be an important tool in the management of patients at our comprehensive cancer center. Although additional false-positive lesions may be detected on RTUR, a great number of patients will benefit from RTUR in finding additional cancers or avoiding unnecessary biopsies.

Positive predictive value of biopsy of palpable masses following mastectomy.

Determine the positive predictive value (PPV) of biopsy of palpable masses following mastectomy (MX). Determine if there are patient characteristics, tumor, or imaging features more predictive of cancer. IRB-approved retrospective review of 16 396 breast ultrasounds June 2008-December 2015 identified patients with MX presenting with palpable masses. Medical records and imaging studies were reviewed. Statistical analysis was performed using Fisher's exact test. 95% confidence intervals (CI) were calculated. In all, 117 patients presented with palpable masses on the MX side. 101/117 patients who had a palpable mass on physical examination had a true sonographic mass to correlate with the clinical findings. 91/101 (90%) underwent biopsy: 19/91 (21%, 95% CI; 13-31) biopsies were malignant. 72/91 (79%) were benign. All 19 cancers were on the original cancer side. Recurrences ranged from 0.4 to 4.5 cm maximum diameter, mean 1.3 cm. Prophylactic vs therapeutic mastectomy was very statistically significant (P = .01). The use of tamoxifen or an AI was also statistically significant (P = .04). Patient age (P = 1.0), radiation therapy (P = 1.05), chemotherapy (P = .2), immediate breast reconstruction (P = .2), or implant vs flap (P = .2) had no statistically significant association with finding cancer on biopsy. Lesion shape (irregular vs oval/round) was highly statistically significant (P = .0003) as was non-parallel orientation on ultrasound (P = .008). Circumscribed vs non-circumscribed margins was also statistically significant (P = .008). The PPV of biopsy of palpable masses on the side of MX was 21% (95% CI; 13-31). All recurrences were on the original cancer side and this was very statistically significant.

Second-Opinion Review of Breast Imaging at a Cancer Center: Is It Worthwhile?

OBJECTIVE: Second-opinion review of breast imaging studies can be a time-consuming and labor-intensive process. The purpose of this investigation was to determine whether reinterpretation of studies obtained at institutions outside a cancer center influences clinical management, specifically by revealing additional cancer and preventing unnecessary biopsy. MATERIALS AND METHODS: A review was conducted of breast imaging studies of 200 patients who underwent ultrasound and MRI at community facilities and had the images submitted for second opinions at a cancer center between January and April 2014. Each case was evaluated for concordance between the original report and the second-opinion interpretation. Second-opinion review resulting in the recommendation and performance of new biopsies was further subdivided into benign, high-risk, and malignant categories based on the histopathologic results obtained at the cancer center. RESULTS: Second-opinion review of the 200 cases showed a change in interpretation in 55 cases (28%; 95% CI, 21-34%). Overall, 26 recommendations (13%; 95% CI, 9-18%) led to a major change in management. Twenty new biopsies were performed, yielding 10 malignancies (5%; 95% CI, 2-9%) and four high-risk lesions (2%; 95% CI, 1-5%). Surgical management was changed to mastectomy for 6 of 10 patients (60%) with new sites of biopsy-proven malignancy. Eight biopsies were averted (4%; 95% CI, 2-8%) on the basis of benign interpretation of the imaging findings, and no disease was found at 1-year follow-up evaluation. CONCLUSION: Reinterpretation of studies obtained outside a cancer center resulted in a change in interpretation in more than one-fourth of submitted studies. Additional cancer was detected in 5% of patients, and biopsy was averted for 4%. The practice of second-opinion review influences clinical management and adds value to patient care.

Comparison of Background Parenchymal Enhancement at Contrast-enhanced Spectral Mammography and Breast MR Imaging.

Purpose To assess the extent of background parenchymal enhancement (BPE) at contrast material-enhanced (CE) spectral mammography and breast magnetic resonance (MR) imaging, to evaluate interreader agreement in BPE assessment, and to examine the relationships between clinical factors and BPE. Materials and Methods This was a retrospective, institutional review board-approved, HIPAA-compliant study. Two hundred seventy-eight women from 25 to 76 years of age with increased breast cancer risk who underwent CE spectral mammography and MR imaging for screening or staging from 2010 through 2014 were included. Three readers independently rated BPE on CE spectral mammographic and MR images with the ordinal scale: minimal, mild, moderate, or marked. To assess pairwise agreement between BPE levels on CE spectral mammographic and MR images and among readers, weighted κ coefficients with quadratic weights were calculated. For overall agreement, mean κ values and bootstrapped 95% confidence intervals were calculated. The univariate and multivariate associations between BPE and clinical factors were examined by using generalized estimating equations separately for CE spectral mammography and MR imaging. Results Most women had minimal or mild BPE at both CE spectral mammography (68%-76%) and MR imaging (69%-76%). Between CE spectral mammography and MR imaging, the intrareader agreement ranged from moderate to substantial (κ = 0.55-0.67). Overall agreement on BPE levels between CE spectral mammography and MR imaging and among readers was substantial (κ = 0.66; 95% confidence interval: 0.61, 0.70). With both modalities, BPE demonstrated significant association with menopausal status, prior breast radiation therapy, hormonal treatment, breast density on CE spectral mammographic images, and amount of fibroglandular tissue on MR images (P < .001 for all). Conclusion There was substantial agreement between readers for BPE detected on CE spectral mammographic and MR images. © RSNA, 2016.

Cancelled stereotactic biopsy of calcifications not seen using the stereotactic technique: do we still need to biopsy?

OBJECTIVE: To determine the frequency of cancelled stereotactic biopsy due to non-visualisation of calcifications, and assess associated features and outcome data. METHODS: A retrospective review was performed on 1,874 patients scheduled for stereotactic-guided breast biopsy from 2009 to 2011. Medical records and imaging studies were reviewed. RESULTS: Of 1,874 stereotactic biopsies, 76 (4 %) were cancelled because of non-visualisation of calcifications. Prompt histological confirmation was obtained in 42/76 (55 %). In 28/76 (37 %) follow-up mammography was performed, and 7/28 subsequently underwent biopsy. Of 27 without biopsy, 21 (78 %) had follow-up. Nine cancers (9/49, 18 %) were found: 6 ductal carcinoma in situ (DCIS), 3 infiltrating ductal carcinoma (IDC). Of 54 patients with either biopsy or at least 2 years' follow-up, 9 (17 %) had cancer (95 % CI 8-29). Cancer was present in 7/42 (17 %, 95 % CI 7-31 %) lesions that had prompt histological confirmation (DCIS = 5, IDC = 2) and in 2/28 (7 %, 95 % CI 0.8-24 %) lesions referred for follow-up (DCIS = 1, IDC = 1). Neither calcification morphology (P = 0.2), patient age (P = 0.7), breast density (P = 1.0), personal history (P = 1.0) nor family history of breast cancer (P = 0.5) had a significant association with cancer. CONCLUSION: Calcifications not visualised on the stereotactic unit are not definitely benign and require surgical biopsy or follow-up. No patient or morphological features were predictive of cancer. KEY POINTS: • Half of cancelled stereotactic biopsies were due to non-visualisation of calcified foci. • This reflects the improved detection of calcifications by digital mammography. • Calcifications too faint for the stereotactic technique require alternative biopsy or follow-up • 17 % of patients with biopsy or at least 2 years' follow-up had cancer. • No patient/morphological features were found to aid selection for re-biopsy vs. follow-up.