Ruth Walsh

Breast self-examination: defining a cohort still in need

BACKGROUND The value of breast self-examination (BSE) to detect early breast cancer is controversial. METHODS Within an institutional review board-approved prospective study, 147 high-risk women were enrolled from 2004 to 2007. Yearly clinical examination, BSE teaching, and mammography were performed simultaneously followed by interval breast magnetic resonance imaging (MRI). Women underwent additional BSE teaching at 6 months. Women reporting a mass on BSE underwent clinical evaluation. RESULTS Fourteen breast cancers were detected in 12 women. BSE detected 6/14 breast cancers versus 6/14 detected by MRI and 2/14 by mammography. Of 24 masses detected by BSE, 6/24 were malignant. The sensitivity, specificity, and predictive value of BSE to detect breast cancer were 58.3%, 87.4%, and 29.2%, respectively. The sensitivity, specificity, and predictive value of a Breast Image Reporting and Data System (BI-RADS) score of >or=4 on MRI were 66.7%, 88.9%, and 34.8%, respectively. CONCLUSIONS BSE detects new breast cancers in high-risk women undergoing screening mammogram, CBE, and yearly breast MRI.

Comparative performance of multiview stereoscopic and mammographic display modalities for breast lesion detection

PURPOSE Mammography is known to be one of the most difficult radiographic exams to interpret. Mammography has important limitations, including the superposition of normal tissue that can obscure a mass, chance alignment of normal tissue to mimic a true lesion and the inability to derive volumetric information. It has been shown that stereomammography can overcome these deficiencies by showing that layers of normal tissue lay at different depths. If standard stereomammography (i.e., a single stereoscopic pair consisting of two projection images) can significantly improve lesion detection, how will multiview stereoscopy (MVS), where many projection images are used, compare to mammography? The aim of this study was to assess the relative performance of MVS compared to mammography for breast mass detection. METHODS The MVS image sets consisted of the 25 raw projection images acquired over an arc of approximately 45 degrees using a Siemens prototype breast tomosynthesis system. The mammograms were acquired using a commercial Siemens FFDM system. The raw data were taken from both of these systems for 27 cases and realistic simulated mass lesions were added to duplicates of the 27 images at the same local contrast. The images with lesions (27 mammography and 27 MVS) and the images without lesions (27 mammography and 27 MVS) were then postprocessed to provide comparable and representative image appearance across the two modalities. All 108 image sets were shown to five full-time breast imaging radiologists in random order on a state-of-the-art stereoscopic display. The observers were asked to give a confidence rating for each image (0 for lesion definitely not present, 100 for lesion definitely present). The ratings were then compiled and processed using ROC and variance analysis. RESULTS The mean AUC for the five observers was 0.614 +/- 0.055 for mammography and 0.778 +/- 0.052 for multiview stereoscopy. The difference of 0.164 +/- 0.065 was statistically significant with a p-value of 0.0148. CONCLUSIONS The differences in the AUCs and the p-value suggest that multiview stereoscopy has a statistically significant advantage over mammography in the detection of simulated breast masses. This highlights the dominance of anatomical noise compared to quantum noise for breast mass detection. It also shows that significant lesion detection can be achieved with MVS without any of the artifacts associated with tomosynthesis.

Interobserver Variability Between Breast Imagers Using the Fifth Edition of the BI-RADS MRI Lexicon.

OBJECTIVE The purpose of this study was to assess the interobserver variability of users of the MRI lexicon in the fifth edition of the BI-RADS atlas. MATERIALS AND METHODS Three breast imaging specialists reviewed 280 routine clinical breast MRI findings reported as BI-RADS category 3. Lesions reported as BI-RADS 3 were chosen because variability in the use of BI-RADS descriptors may influence which lesions are classified as probably benign. Each blinded reader reviewed every study and recorded breast features (background parenchymal enhancement) and lesion features (lesion morphology, mass shape, mass margin, mass internal enhancement, nonmass enhancement distribution, nonmass enhancement internal enhancement, enhancement kinetics) according to the fifth edition of the BI-RADS lexicon and provided a final BI-RADS assessment. Interobserver variability was calculated for each breast and lesion feature and for the final BI-RADS assessment. RESULTS Interobserver variability for background parenchymal enhancement was fair (ĸ = 0.28). There was moderate agreement on lesion morphology (ĸ = 0.53). For masses, there was substantial agreement on shape (ĸ = 0.72), margin (ĸ = 0.78), and internal enhancement (ĸ = 0.69). For nonmass enhancement, there was substantial agreement on distribution (ĸ = 0.69) and internal enhancement (ĸ = 0.62). There was slight agreement on lesion kinetics (ĸ = 0.19) and final BI-RADS assessment (ĸ = 0.11). CONCLUSION There is moderate to substantial agreement on most MRI BI-RADS lesion morphology descriptors, particularly mass and nonmass enhancement features, which are important predictors of malignancy. Considerable disagreement remains, however, among experienced readers whether to follow particular findings.

Streaming detection for evaluation of indeterminate sonographic breast masses : A pilot study

OBJECTIVE Streaming detection is a novel sonography technique that uses ultrasonic energy to induce movement in cyst fluid that is detected on Doppler sonography. This pilot study evaluates the utility of streaming detection for differentiating cysts from solid masses in breast lesions that are indeterminate on sonography. SUBJECTS AND METHODS Thirty-nine lesions-11 simple cysts and seven solid masses (control group) and 21 masses with indeterminate findings for the diagnosis of a cyst versus a solid lesion (study group)-in 34 patients were evaluated using streaming detection. All lesions underwent cyst aspiration or biopsy (n = 35) or were diagnosed simple cysts (n = 4) on sonography. Lesion size and depth were recorded. Streaming detection software was placed on conventional sonography units. Acoustic pulses were focused on the lesion, and if fluid movement was generated, it was seen on the spectral Doppler display as velocity away from the transducer. Lesions were then aspirated or underwent biopsy, and the viscosity of the aspirated fluid was recorded. The sensitivity and specificity of the technique and the effect of cyst size, cyst depth, and fluid viscosity in diagnosing fluid-filled cysts were assessed. RESULTS Overall, 31 cysts and eight solid masses (seven benign, one carcinoma) were diagnosed in the study and control groups. Aspiration of indeterminate lesions resulted in 20 cysts and one solid mass. Lesions ranged in size from 4 to 47 mm and in depth from 4 to 29 mm. In the control group, streaming detection correctly showed nine of the 11 simple cysts (sensitivity, 82%; positive predictive value, 100%), and acoustic streaming was absent in all seven solid masses (specificity, 100%; negative predictive value, 78%). Of the indeterminate lesions, streaming detection allowed correct identification of 10 of 20 cysts (sensitivity, 50%; positive predictive value, 100%). Acoustic streaming was not detected in the one solid study group lesion. Neither cyst size or depth nor fluid viscosity had a significant effect on the ability to detect fluid. CONCLUSION The streaming detection technique improved differentiation of cysts from solid masses in indeterminate lesions and has potential for reducing the number of recommended cyst aspirations for the diagnosis of indeterminate breast masses.

Frequency of Malignancy and Imaging Characteristics of Probably Benign Lesions Seen at Breast MRI

OBJECTIVE The purposes of this study were to evaluate the frequency, follow-up compliance, and cancer rate of MRI BI-RADS category 3 lesions and to determine the cancer rate for individual BI-RADS descriptors. MATERIALS AND METHODS A retrospective review was conducted of breast MRI examinations with an assessment of probably benign (BI-RADS category 3) from among 4279 consecutive breast MRI examinations performed from January 2005 through December 2009. The review revealed 282 (6.6%) examinations with 332 lesions defined as BI-RADS 3. Pathologic results, 2 years of follow-up imaging findings, or both were reviewed. The frequency of BI-RADS 3 assessments, follow-up imaging compliance, and cancer yield were calculated. Three fellowship-trained breast imagers reevaluated all lesions and recorded descriptors from the MRI lexicon of the fifth edition of the BI-RADS atlas. The distribution and likelihood of malignancy for each descriptor were calculated. RESULTS The follow-up compliance rate was 84.3% (280/332), and the malignancy rate was 4.3% (12/280). There were 50 (17.9%) individual foci, 61 (21.8%) multiple foci, 74 (26.4%) masses, and 95 (33.9%) nonmass enhancement lesions. Masses were most commonly oval (59.5% [44/74]), circumscribed (75.7% [56/74]), and homogeneously enhancing (43.2% [32/74]). Nonmass enhancement was most commonly focal (57.9% [55/95]) with heterogeneous enhancement (53.7% [51/95]) Most of the lesions had persistent kinetics (74.3% [208/280]). The background parenchymal enhancement was most commonly mild (51.1% [143/280]). CONCLUSION MRI BI-RADS category 3 is not frequently used, and the levels of patient compliance with follow-up imaging are acceptable. The cancer yield for probably benign lesions is greater for MRI-detected than for mammographically detected lesions, especially for specific BI-RADS descriptors.

Periprosthetic mycobacterial infection. CT and mammographic findings.

Organisms of the Mycobacterium fortuitum complex are an uncommon but important cause of periprosthetic infection following augmentation mammoplasty or other breast surgery. This etiological agent must be considered in the particular case of periprosthetic infection, because special handling of the fluid is crucial to enhance recovery of the organism. We describe the computed tomography (CT) and mammographic findings in such an abscess with respect to the clinical context and subsequent management. To our knowledge, CT findings associated with any periprosthetic breast infection have not been described.

Does Breast Imaging Experience During Residency Translate Into Improved Initial Performance in Digital Breast Tomosynthesis

PURPOSE To determine the initial digital breast tomosynthesis (DBT) performance of radiology trainees with varying degrees of breast imaging experience. METHODS To test trainee performance with DBT, we performed a reader study, after obtaining IRB approval. Two medical students, 20 radiology residents, 4 nonbreast imaging fellows, 3 breast imaging fellows, and 3 fellowship-trained breast imagers reviewed 60 unilateral DBT studies (craniocaudal and medio-lateral oblique views). Trainees had no DBT experience. Each reader recorded a final BI-RADS assessment for each case. The consensus interpretations from fellowship-trained breast imagers were used to establish the ground truth. Area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were calculated. For analysis, first- through third-year residents were classified as junior trainees, and fourth-year residents plus nonbreast imaging fellows were classified as senior trainees. RESULTS The AUCs were .569 for medical students, .721 for junior trainees, .701 for senior trainees, and .792 for breast imaging fellows. The junior and senior trainee AUCs were equivalent (P < .01) using a two one-sided test for equivalence, with a significance threshold of 0.1. The sensitivities and specificities were highest for breast imaging fellows (.778 and .815 respectively), but similar for junior (.631 and .714, respectively) and senior trainees (.678 and .661, respectively). CONCLUSIONS Initial performance with DBT among radiology residents and nonbreast imaging fellows is independent of years of training. Radiology educators should consider these findings when developing educational materials.

Using the BI-RADS Lexicon in a Restrictive Form of Double Reading as a Strategy for Minimizing Screening Mammography Recall Rates

OBJECTIVE The purpose of this article is to determine the potential reduction in screening recall rates by strictly following standardized BI-RADS lexicon for lesions seen on screening mammography. MATERIALS AND METHODS Of 3084 consecutive mammograms performed at our screening facilities, 345 women with 437 lesions were recalled for additional imaging and constituted our study population. Three radiologists retrospectively classified lesions using the standard BI-RADS lexicon and assigned each to one of four groups: group A, the finding met criteria for recall by the BI-RADS lexicon; group B, the finding did not meet strict BI-RADS criteria for recall but was sufficiently indeterminate to warrant recall by the majority of the study panel; group C, the finding was classifiable by the BI-RADS lexicon but was not recalled because it was benign or stable; and group D, the questioned finding was not considered an abnormality by our study panel. Recall rates and the cancer detection rate were determined. The adjusted recall rate was calculated for lesions considered appropriate for recall (group A), and the reduction in the recall rate was determined. RESULTS Nineteen malignancies were detected in our recalled population, for a cancer detection rate of 0.65%. All 19 malignancies were lesions considered appropriate for recall (group A). If only group A lesions had been recalled, the recall rate would have decreased from 11.4% to 6.2%, representing a 46% reduction in recalls without affecting the cancer detection rate. CONCLUSION Using the BI-RADS lexicon as a decision-making aid may help adjust thresholds for recalling indeterminate or suspicious lesions and reduce recall rates from screening mammography.

Predicting false negative errors in digital breast tomosynthesis among radiology trainees using a computer vision-based approach

We developed a model to predict whether a lesion will be missed by a trainee.The user model can be used to select the most challenging cases for each trainee.Our model improved the status quo of case presentation to trainee in tomosynthesis. PurposeDigital breast tomosynthesis (DBT) can improve lesion visibility in comparison to mammography by eliminating breast tissue superimposition. While the benefits of DBT in breast cancer screening rely on well trained radiologists, the optimal training regimen in DBT is unknown. We propose a computer-aided educational system that individually selects the optimal training cases for each trainee. The first step towards this goal is to capture the individual weaknesses of each trainee. In this study, we present and evaluate a computer algorithm for this purpose with particular focus on false negative errors. MethodsWe developed an algorithm (a user model) that predicted the likelihood of a trainee missing an abnormal location. An individual model is applied for each trainee. The algorithm consists of three steps. First, the lesions on DBT images are segmented by a 3D active contour method with a level set algorithm. Then, 16 features are extracted automatically for the segmented lesions. Finally a multivariate logistic regression classifier predicts the likelihood of error based on the extracted features. The classifier is trained using the previous interpretation data of the trainee. We evaluated the individual predictive algorithms experimentally using data from a reader study in which 29 trainees and 3 expert breast radiologists read 60 DBT cases. Receiver operating characteristic (ROC) analysis, along with a repeated holdout approach, was used to evaluate the predictive performance of our algorithm. ResultsThe average area under the ROC curve (AUC) of the algorithms which predicted which lesions will be detected and which will be missed by a specific trainee was 0.627 (95% CI: 0.579-0.675). The average performance was statistically significantly better than chance (p<0.001). Under the status quo, training involves no specific strategy for case presentation, and this random behavior corresponds to AUC of 0.5. Therefore, the proposed algorithm may provide a significant improvement in distinguishing abnormal locations that will be detected by a trainee from those that will be missed. ConclusionsOur algorithm was able to distinguish abnormal locations that will be detected by a trainee from those that will be missed. This could be used to enrich the training set with cases that are likely to prompt error for the individual trainee while still maintaining a range of cases necessary for comprehensive education.

Association of high proliferation marker Ki-67 expression with DCEMR imaging features of breast: a large scale evaluation.

One of the methods widely used to measure the proliferative activity of cells in breast cancer patients is the immunohistochemical (IHC) measurement of the percentage of cells stained for nuclear antigen Ki-67. Use of Ki-67 expression as a prognostic marker is still under investigation. However, numerous clinical studies have reported an association between a high Ki-67 and overall survival (OS) and disease free survival (DFS). On the other hand, to offer non-invasive alternative in determining Ki-67 expression, researchers have made recent attempts to study the association of Ki-67 expression with magnetic resonance (MR) imaging features of breast cancer in small cohorts (<30). Here, we present a large scale evaluation of the relationship between imaging features and Ki-67 score as: (a) we used a set of 450 invasive breast cancer patients, (b) we extracted a set of 529 imaging features of shape and enhancement from breast, tumor and fibroglandular tissue of the patients, (c) used a subset of patients as the training set to select features and trained a multivariate logistic regression model to predict high versus low Ki-67 values, and (d) we validated the performance of the trained model in an independent test set using the area-under the receiver operating characteristics (ROC) curve (AUC) of the values predicted. Our model was able to predict high versus low Ki-67 in the test set with an AUC of 0.67 (95% CI: 0.58-0.75, p<1.1e-04). Thus, a moderate strength of association of Ki-67 values and MRextracted imaging features was demonstrated in our experiments.