Min Sun Bae

Breast Cancer Detected with Screening US: Reasons for Nondetection at Mammography

PURPOSE To retrospectively review the mammograms of women with breast cancers detected at screening ultrasonography (US) to determine the reasons for nondetection at mammography. MATERIALS AND METHODS This study received institutional review board approval, and informed consent was waived. Between 2003 and 2011, a retrospective database review revealed 335 US-depicted cancers in 329 women (median age, 47 years; age range, 29-69 years) with Breast Imaging Reporting and Data System breast density type 2-4. Five blinded radiologists independently reviewed the mammograms to determine whether the findings on negative mammograms should be recalled. Three unblinded radiologists re-reviewed the mammograms to determine the reasons for nondetection by using the reference location of the cancer on mammograms obtained after US-guided wire localization or breast magnetic resonance imaging. The number of cancers recalled by the blinded radiologists were compared with the reasons for nondetection determined by the unblinded radiologists. RESULTS Of the 335 US-depicted cancers, 63 (19%) were recalled by three or more of the five blinded radiologists, and 272 (81%) showed no mammographic findings that required immediate action. In the unblinded repeat review, 263 (78%) cancers were obscured by overlapping dense breast tissue, and nine (3%) were not included at mammography owing to difficult anatomic location or poor positioning. Sixty-three (19%) cancers were considered interpretive errors. Of these, 52 (82%) were seen as subtle findings (46 asymmetries, six calcifications) and 11 (18%) were evident (six focal asymmetries, one distortion, four calcifications). CONCLUSION Most breast cancers (81%) detected at screening US were not seen at mammography, even in retrospect. In addition, 19% had subtle or evident findings missed at mammography.

Computer-Aided Tumor Detection Based on Multi-Scale Blob Detection Algorithm in Automated Breast Ultrasound Images

Automated whole breast ultrasound (ABUS) is an emerging screening tool for detecting breast abnormalities. In this study, a computer-aided detection (CADe) system based on multi-scale blob detection was developed for analyzing ABUS images. The performance of the proposed CADe system was tested using a database composed of 136 breast lesions (58 benign lesions and 78 malignant lesions) and 37 normal cases. After speckle noise reduction, Hessian analysis with multi-scale blob detection was applied for the detection of tumors. This method detected every tumor, but some nontumors were also detected. The tumor likelihoods for the remaining candidates were estimated using a logistic regression model based on blobness, internal echo, and morphology features. The tumor candidates with tumor likelihoods higher than a specific threshold (0.4) were considered tumors. By using the combination of blobness, internal echo, and morphology features with 10-fold cross-validation, the proposed CAD system showed sensitivities of 100%, 90%, and 70% with false positives per pass of 17.4, 8.8, and 2.7, respectively. Our results suggest that CADe systems based on multi-scale blob detection can be used to detect breast tumors in ABUS images.

Added Value of Shear-Wave Elastography for Evaluation of Breast Masses Detected with Screening US Imaging

PURPOSE To evaluate the additional value of shear-wave elastography (SWE) to B-mode ultrasonography (US) and to determine an appropriate guideline for the combined assessment of screening US-detected breast masses. MATERIALS AND METHODS This study was conducted with institutional review board approval, and written informed consent was obtained. From March 2010 to February 2012, B-mode US and SWE were performed in 159 US-detected breast masses before biopsy. For each lesion, Breast Imaging Reporting and Data System (BI-RADS) category on B-mode US images and the maximum stiffness color and elasticity values on SWE images were assessed. A guideline for adding SWE data to B-mode US was developed with the retrospective cohort to improve diagnostic performance in sensitivity and specificity and was validated in a distinct prospective cohort of 207 women prior to biopsy. RESULTS Twenty-one of 159 masses in the development cohort and 12 of 207 breast masses in the validation cohort were malignant. In the development cohort, when BI-RADS category 4a masses showing a dark blue color or a maximum elasticity value of 30 kPa or less on SWE images were downgraded to category 3, specificity increased from 9.4% (13 of 138) to 59.4% (82 of 138) and 57.2% (79 of 138) (P < .001), respectively, without loss in sensitivity (100% [21 of 21]). In the validation cohort, specificity increased from 17.4% (34 of 195) to 62.1% (121 of 195) and 53.3% (104 of 195) (P < .001) respectively, without loss in sensitivity (91.7% [11 of 12]). CONCLUSION The addition of SWE to B-mode US improved diagnostic performance with increased specificity for screening US-detected breast masses. BI-RADS category 4a masses detected at US screening that showed a dark blue color or a maximum elasticity value of 30 kPa or less on SWE images can be safely followed up instead of performing biopsy.

Robust Texture Analysis Using Multi-Resolution Gray-Scale Invariant Features for Breast Sonographic Tumor Diagnosis

Computer-aided diagnosis (CAD) systems in gray-scale breast ultrasound images have the potential to reduce unnecessary biopsy of breast masses. The purpose of our study is to develop a robust CAD system based on the texture analysis. First, gray-scale invariant features are extracted from ultrasound images via multi-resolution ranklet transform. Thus, one can apply linear support vector machines (SVMs) on the resulting gray-level co-occurrence matrix (GLCM)-based texture features for discriminating the benign and malignant masses. To verify the effectiveness and robustness of the proposed texture analysis, breast ultrasound images obtained from three different platforms are evaluated based on cross-platform training/testing and leave-one-out cross-validation (LOO-CV) schemes. We compare our proposed features with those extracted by wavelet transform in terms of receiver operating characteristic (ROC) analysis. The AUC values derived from the area under the curve for the three databases via ranklet transform are 0.918 (95% confidence interval [CI], 0.848 to 0.961), 0.943 (95% CI, 0.906 to 0.968), and 0.934 (95% CI, 0.883 to 0.961), respectively, while those via wavelet transform are 0.847 (95% CI, 0.762 to 0.910), 0.922 (95% CI, 0.878 to 0.958), and 0.867 (95% CI, 0.798 to 0.914), respectively. Experiments with cross-platform training/testing scheme between each database reveal that the diagnostic performance of our texture analysis using ranklet transform is less sensitive to the sonographic ultrasound platforms. Also, we adopt several co-occurrence statistics in terms of quantization levels and orientations (i.e., descriptor settings) for computing the co-occurrence matrices with 0.632+ bootstrap estimators to verify the use of the proposed texture analysis. These experiments suggest that the texture analysis using multi-resolution gray-scale invariant features via ranklet transform is useful for designing a robust CAD system.

Background Parenchymal Signal Enhancement Ratio at Preoperative MR Imaging: Association with Subsequent Local Recurrence in Patients with Ductal Carcinoma in Situ after Breast Conservation Surgery

PURPOSE To retrospectively investigate whether the background parenchymal features around a tumor at preoperative dynamic contrast material-enhanced magnetic resonance (MR) imaging are associated with ipsilateral breast tumor recurrence (IBTR)-free survival in patients with ductal carcinoma in situ (DCIS) after breast conservation surgery. MATERIALS AND METHODS The institutional review board approved this study, and the requirement for informed consent was waived. Between 2004 and 2009, 215 consecutive women with pure DCIS who had undergone preoperative dynamic contrast-enhanced MR imaging and curative breast conservation surgery were identified. Clinical-pathologic features (age, menopausal status, presentation of clinical findings, biopsy method, tumor size, nuclear grade, hormonal receptor status, margin status, and adjuvant therapy) and MR imaging features (lesion size, background parenchymal enhancement grade, fibroglandular density, parenchymal signal enhancement ratio [SER] around the tumor, lesion type, and lesion kinetics) were analyzed. A Cox proportional hazards model was used to determine the association between MR imaging variables and IBTR-free survival after controlling for clinical-pathologic variables. Reproducibility of SER measurements was evaluated by using the intraclass correlation coefficient. RESULTS There were 15 of 215 (7.0%) IBTR cases (nine DCIS cases and six invasive cases) at a median of 36 months (range, 11-61 months). Multivariate analysis showed that higher parenchymal SER (hazard ratio [HR] = 2.028, P < .001 for reader 1; HR = 1.652, P < .001 for reader 2) and larger histologic tumor size (HR = 1.360, P = .009 for reader 1; HR = 1.402, P = .006 for reader 2) were independent factors associated with worse IBTR-free survival. The intraclass correlation coefficient of SER measurements between two readers was 0.852 (95% confidence interval: 0.811, 0.885). CONCLUSION Higher parenchymal SER around the tumor at preoperative dynamic contrast-enhanced MR imaging and larger histologic tumor size were independent factors associated with worse IBTR-free survival in patients with DCIS after breast conservation surgery.

Two-View versus Single-View Shear-Wave Elastography: Comparison of Observer Performance in Differentiating Benign from Malignant Breast Masses

PURPOSE To determine whether two-view shear-wave elastography (SWE) improves the performance of radiologists in differentiating benign from malignant breast masses compared with single-view SWE. MATERIALS AND METHODS This prospective study was conducted with institutional review board approval, and written informed consent was obtained. B-mode ultrasonographic (US) and orthogonal SWE images were obtained for 219 breast masses (136 benign and 83 malignant; mean size, 14.8 mm) in 219 consecutive women (mean age, 47.9 years; range, 20-78 years). Five blinded radiologists independently assessed the likelihood of malignancy for three data sets: B-mode US alone, B-mode US and single-view SWE, and B-mode US and two-view SWE. Interobserver agreement regarding Breast Imaging Reporting and Data System (BI-RADS) category and the area under the receiver operating characteristic curve (AUC) of each data set were compared. RESULTS Interobserver agreement was moderate (κ = 0.560 ± 0.015 [standard error of the mean]) for BI-RADS category assessment with B-mode US alone. When SWE was added to B-mode US, five readers showed substantial interobserver agreement (κ = 0.629 ± 0.017 for single-view SWE; κ = 0.651 ± 0.014 for two-view SWE). The mean AUC of B-mode US was 0.870 (range, 0.855-0.884). The AUC of B-mode US and two-view SWE (average, 0.928; range, 0.904-0.941) was higher than that of B-mode US and single-view SWE (average, 0.900; range, 0.890-0.920), with statistically significant differences for three readers (P ≤ .003). CONCLUSION The performance of radiologists in differentiating benign from malignant breast masses was improved when B-mode US was combined with two-view SWE compared with that when B-mode US was combined with single-view SWE.

Quantitative MRI morphology of invasive breast cancer: correlation with immunohistochemical biomarkers and subtypes

Background Breast cancer is a heterogeneous disease with intrinsic molecular subtypes. The different biology and histology of breast cancer exhibit different tumor morphology at breast magnetic resonance imaging (MRI). However, few studies have examined the quantitative relationship between the MRI morphological and immunohistochemical features in breast cancer. Purpose To investigate the correlations between tumor roundness, as quantitatively assessed with MRI and biomarkers or subtypes of breast cancer. Material and Methods A total of 280 women (mean age, 51 years; range, 28–79 years) with 282 invasive breast cancers (<5 cm) were included. The associations between the tumor roundness (1–100%), as measured using MRI software, and immunohistochemical (e.g. estrogen receptor [ER], progesterone receptor [PR], human epidermal growth factor receptor 2 [HER2], and Ki67) features were evaluated using Pearson’s or Spearman’s rank correlation coefficients and multiple linear regression analysis. Results An inverse correlation was observed between the ER (r = –0.408, P < 0.001) or PR (r = –0.248, P < 0.001) scores and tumor roundness, whereas a positive correlation was observed between the Ki67 index and tumor roundness (r = 0.354, P < 0.001). In multiple linear regression, the ER score (P < 0.001) and Ki67 index (P = 0.003) were independent factors determining tumor roundness. Triple-negative tumors (ER, PR, and HER2 negative) showed the highest mean roundness scores compared with the other subtypes (e.g. 67.3% for triple-negative, relative to 55.9% for HER2-enriched, 53.8% for luminal B, and 51.7% for luminal A, P < 0.001). Conclusion Our results suggest that breast tumors with lower ER expression and higher cellular proliferation or biologically aggressive triple-negative tumors are likely to manifest with relatively benign morphologic features.

Comparative evaluation of average glandular dose and breast cancer detection between single-view digital breast tomosynthesis (DBT) plus single-view digital mammography (DM) and two-view DM: correlation with breast thickness and density

AbstractObjectivesTo compare the average glandular dose (AGD) and diagnostic performance of mediolateral oblique (MLO) digital breast tomosynthesis (DBT) plus cranio-caudal (CC) digital mammography (DM) with two-view DM, and to evaluate the correlation of AGD with breast thickness and density.MethodsMLO and CC DM and DBT images of both breasts were obtained in 149 subjects. AGDs of DBT and DM per exposure were recorded, and their correlation with breast thickness and density were evaluated. Paired data of MLO DBT plus CC DM and two-view DM were reviewed for presence of malignancy in a jack-knife alternative free-response ROC (JAFROC) method.ResultsThe AGDs of both DBT and DM, and differences in AGD between DBT and DM (ΔAGD), were correlated with breast thickness and density. The average JAFROC figure of merit (FOM) was significantly higher on the combined technique than two-view DM (P = 0.005). In dense breasts, the FOM and sensitivity of the combined technique was higher than that of two-view DM (P = 0.003) with small ΔAGD.ConclusionsMLO DBT plus CC DM provided higher diagnostic performance than two-view DM in dense breasts with a small increase in AGD.Key Points• DBT has higher diagnostic performance and potential to overcome limitations of DM. • Dose differences (DBT-DM, ΔAGD) were inversely correlated with breast thickness and density. • Figure of merit of MLO-DBT/CC-DM was higher than that of two-view DM. • In dense breasts, MLO-DBT/CC-DM provides better diagnostic performance with a small AGD increase.

Early Stage Triple-Negative Breast Cancer: Imaging and Clinical-Pathologic Factors Associated with Recurrence

PURPOSE To determine the imaging and clinical-pathologic factors associated with recurrence in patients with early stage triple-negative breast cancer. MATERIALS AND METHODS This study was approved by the institutional review board, and the requirement to obtain informed consent was waived. The authors evaluated 398 patients with stage I or II triple-negative breast cancer (median age, 48 years; range, 21-81 years) who were treated between January 2003 and December 2008. Data collected included preoperative breast magnetic resonance (MR) images, mammographic density, patient age, symptoms, family history of breast cancer, histologic tumor characteristics, tumor grade, tumor size, lymphovascular invasion, lymph node involvement, surgery type, margin status, and adjuvant treatment received. Multivariate analysis was performed by using a Cox proportional hazards model, and recurrence-free survival was estimated with the adjusted Kaplan-Meier method. RESULTS Of the 398 patients, 63 (15.8%) had recurrent disease after a median follow-up of 6.1 years. The absence of preoperative MR imaging (hazard ratio [HR] with multivariate analysis = 2.66; 95% confidence interval = 1.49, 4.75; P < .001), dense breast tissue (HR = 2.77; 95% confidence interval = 1.39, 5.51; P = .004), family history of breast cancer (HR = 2.32; 95% confidence interval = 1.10, 4.90; P = .028), and lymphovascular invasion (HR = 1.83; 95% confidence interval = 1.11, 3.03; P = .019) were found to be independently associated with recurrence. These same factors were also found to be associated with recurrence-free survival. CONCLUSION The absence of preoperative MR imaging and the presence of dense breast tissue at mammography were associated with an increased risk of recurrence in patients with triple-negative breast cancer.

Breast Cancer Screening With Mammography Plus Ultrasonography or Magnetic Resonance Imaging in Women 50 Years or Younger at Diagnosis and Treated With Breast Conservation Therapy

Importance Younger women (aged ⩽50 years) who underwent breast conservation therapy may benefit from breast magnetic resonance imaging (MRI) screening as an adjunct to mammography. Objective To prospectively determine the cancer yield and tumor characteristics of combined mammography with MRI or ultrasonography screening in women who underwent breast conservation therapy for breast cancers and who were 50 years or younger at initial diagnosis. Design, Setting, and Participants This multicenter, prospective, nonrandomized study was conducted from December 1, 2010, to January 31, 2016, at 6 academic institutions. Seven hundred fifty-four women who were 50 years or younger at initial diagnosis and who had undergone breast conservation therapy for breast cancer were recruited to participate in the study. Reference standard was defined as a combination of pathology and 12-month follow-up. Interventions Participants underwent 3 annual MRI screenings of the conserved and contralateral breasts in addition to mammography and ultrasonography, with independent readings. Main Outcomes and Measures Cancer detection rate, sensitivity, specificity, interval cancer rate, and characteristics of detected cancers. Results A total of 754 women underwent 2065 mammograms, ultrasonography, and MRI screenings. Seventeen cancers were diagnosed, and most of the detected cancers (13 of 17 [76%]) were stage 0 or stage 1. Overall cancer detection rate (8.2 vs 4.4 per 1000; P = .003) or sensitivity (100% vs 53%; P = .01) of mammography with MRI was higher than that of mammography alone. After the addition of ultrasonography, the cancer detection rate was higher than that by mammography alone (6.8 vs 4.4 per 1000; P = .03). The specificity of mammography with MRI or ultrasonography was lower than that by mammography alone (87% or 88% vs 96%; P < .001). No interval cancer was found. Conclusions and Relevance After breast conservation therapy in women 50 years or younger, the addition of MRI to annual mammography screening improves detection of early-stage but biologically aggressive breast cancers at acceptable specificity. Results from this study can inform patient decision making on screening methods after breast conservation therapy.