| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
ABSTRACT |
Tuesday, May 2, 11:20 AM–12:30 PM
Abstracts 108–113
Moderators: Jessica W.T. Leung, MD and Linda J. Warren-Burhenne, MD
11:20 AM
108. Comparison of Image Quality and Average Glandular Dose on Four FDA-approved Full-field Digital Mammography Systems
Berns E.A.*; Hendrick R.E.; Barke L.D.; Radiology, Northwestern University, Chicago, IL.
Address correspondence to E.A. Berns (eberns{at}radiology.northwestern.edu)
Objective: To compare image quality and average glandular dose (AGD) on four FDA-approved full-field digital mammography (FFDM) systems using the manufacturers' recommended techniques across the full range of compressed breast thickness.
Materials and Methods: Contrast-detail (CD) phantom images at 2, 4, 6, and 8 cm thickness of simulated 50% fatty–50% glandular breasts and an ACR phantom image were acquired on each of four different manufacturers' FDA approved FFDM units(GE 2000D and DS, Lorad Selenia, and Siemens Novation) using the manufacturers' recommended techniques using their automatic exposure system. For each image, average glandular dose (AGD) was measured and CD and ACR phantom scores were averaged across three readers. Results were analyzed for statistically significant differences in image quality scores among manufacturers at each breast thickness.
Results: AGD for 2, 4, 6, 8 cm breast thickness and the ACR phantom were 0.54, 1.39, 1.47, 2.76, and 2.11 milliGray (mGy) for the 2000D, 0.61, 0.96, 1.21, 1.71, and 1.11 mGy for the DS, 0.75, 1.63, 2.43, 3.49, and 2.26 mGy for the Selenia, and 0.33, 0.91, 2.34, 4.43, and 1.32 mGy for the Novation, respectively. Comparing doses at each thickness, doses ranged by factors of 2.3 at 2 cm, 1.8 at 4 cm, 2.0 at 6 cm, 2.6 at 8 cm, and 2.0 for the ACR phantom. Mean CD scores for 2, 4, 6, and 8 cm breast thickness were 14.8, 13.8, 12.0, and 11.0 for the 2000D, 14.4, 14.0, 12.2, and 11.3 for the DS, 15.1, 14.8, 13.8, and 12.6 for the Selenia, and 13.8, 13, 12.4, and 12.0 for the Novation, respectively. At 2 cm, CD scores were significantly better for Selenia than DS or Novation DR and better for 2000D than Novation DR (p < 0.05). At 4 cm, CD scores were better for the 2000D and Selenia than Novation (p < 0.05). At 6 cm, CD scores were better for Selenia than DS (p < 0.05). No differences existed at 8 cm. The mean ACR phantom scores for the GE 2000D, GE DS, Lorad Selenia, and Siemens Novation were 4.67, 4.67, 6.0, and 4.17 for fibers, 4.0, 3.83, 4.0, and 3.5 for specks, and 3.83, 3.67, 3.83, and 2.83 for masses, respectively. Fiber scores were significantly better for Selenia than the 2000D and Novation (p < 0.05). No significant differences were seen for speck scores. Mass scores were significantly higher for the 2000D than the Novation (p < 0.05).
Conclusion: These results indicate that AGD varies by factors of up to 2.8 and that there are significant differences in image quality among the full-field digital mammography systems available for clinical use in the U.S.
109. Single X-ray Absorptiometry for Measuring Breast Density Compared with Visual Inspection by Radiologists: Inter- and Intraobserver Variability
Leung J.W.1*; Fan B.1; Kerlikowske K.2; Chen J.1; Shepherd J.A.1; 1. Radiology, University of California San Francisco, San Francisco, CA; 2. Epidemiology and Biostatistics, Internal Medicine, University of California San Francisco, San Francisco, CA.
Address correspondence to J.W. Leung (Jessica.Leung{at}ucsfmedctr.org)
Objective: Breast density is one of the strongest risk factors associated with increased breast cancer risk. Yet, there is no uniformly accepted clinical method of determining breast density. We measured the level of agreement in assessing breast density by visual inspection and correlation of visual inspection with an automated method using single x-ray absorptiometry (SXA).
Materials and Methods: The SXA method measures breast density by comparing extent of opacity present on mammograms to fat and glandular reference standards imaged with the craniocaudal breast view. SXA breast density (BD-SXA) is calculated using mathematical derivations previously described. We applied this automated technique to 286 consecutive film screen mammograms and then digitized the 286 screens to assess BI-RADS breast density. Blinded to BD-SXA results, 2 radiologists assigned BI-RADS density categories (1-fatty; 2-scattered fibroglandular densities; 3-heterogeneously dense; 4-extremely dense) to soft-copy digitized film screen on 2 separate occasions. The second reading occurred at least 1 month after the first reading. Radiologist A was a fellowship-trained mammographer with 7 years of experience. Radiologist B was a fourth-year radiology resident. A weighted kappa statistic was used to assess the variability in assigning BI-RADS categories between radiologists and within radiologists. Statistical significance was shown using 95% confidence interval (CI). Disagreement was considered to exist if there were any categorical differences between readings. Correlation between BI-RADS breast density category and BD-SXA was determined for each radiologist.
Results: Agreement between two radiologists using BI-RADS breast density was substantial (kappa 0.79; 95% CI: 0.72–0.85). Agreement on repeat reading by the same radiologist using BI-RADS was less (kappa 0.67 for Radiologist A [95% CI: 0.59–0.75] and 0.53 [95% CI: 0.45–0.61] for Radiologist B) than agreement between radiologists. Correlation of BI-RADS assessment with BD-SXA (kappa-values of 0.47 [95% CI: 0.73–0.57] and 0.42 [95% CI: 0.33–0.51] for radiologist A and B, respectively) was moderate.
Conclusion: There is moderate to substantial agreement in assigning BI-RADS breast density on screening mammography examinations. Correlation of BI-RADS and BD-SXA was modest. This suggests that BD-SXA is measuring breast density differently than a qualitative assessment. Future studies will examine whether BD-SXA is more strongly associated with breast cancer risk than BI-RADS breast density.
110. Algorithmic Prediction of Lesion Location on Breast Ultrasound from Digitized Mammograms
McNamara M.P.1,3*; Izen S.H.2; Rohler D.P.; 1. Radiology—MetroHealth Campus, Case Western Reserve University, Cleveland, OH; 2. Mathematics, Case Western Reserve University, Cleveland, OH; 3. Radiology, Uniformed Services University of the Health Sciences, Bethesda, MD.
Address correspondence to M.P. McNamara (mmcnamara{at}metrohealth.org)
Objective: The purpose of this study was to devise and test a proof of concept software algorithm to predict breast lesion location on ultrasound from spatial coordinates available on mammography. In this communication we report the performance of the model using only lesion and nipple locations in Cartesian coordinates and the orientation at which the mammograms were obtained.
Materials and Methods: Case material was derived from an anonymized historical data set of mammographic lesions with unambiguous ultrasound correlation. Mammographic images for each case consisted of cranio-caudal and medio-lateral oblique views demonstrating an unambiguous mass. Ultrasound images of the same mass were done with the patient supine and turned to the contralateral side a sufficient amount for the breast flatten evenly against the chest wall. The sonographic location of all lesions is expressed in clock-face position and distance from the nipple. Software was devised to digitally extract and record mammographic lesion coordinates from the anonymized historical data set of mammographic masses. A proof-of-concept theoretical mathematical model was developed to relate the mammographic coordinates of a lesion to ultrasound coordinates. The output of the algorithm is the lesion location in ultrasound coordinates.
Results: When the only mammographic variables considered are Cartesian coordinates and image angle, the ultrasound coordinates can be predicted with an absolute angular error of 27.7 degrees (less than one hour). The mean radial error was 2.4 cm (absolute x coordinate error of 1.4 cm and y coordinate error of 3.3 cm).
Conclusion: In day-to-day practice, it can be difficult at times to be certain that the same lesion is being imaged by mammography and ultrasound. With the advent of digital mammography workstations our data indicates the potential of software to predict the ultrasound location of a mammographic lesion. Our current algorithm provides a reasonable prediction of ultrasound clock face position of a lesion using only mammographic Cartesian coordinates and image angle. To reduce radial error to an acceptable amount, we are assessing the effects of breast size and composition, mammographic exposure factors and compression as well as the contribution of systematic geometric distortion for inclusion in a future model.
111. Evaluation of a Motion Correction Algorithm for Breast MR
Middleton M.1; Moy L.2; Toth H.2; Hermosillo G.4*; Schultze-Haakh H.3; Salganicoff M.4; 1. Radiology, MRI Center, University of California, San Diego, San Diego, CA; 2. Kaplan Breast Imaging Center, NYU Medical Center, NYC, NY; 3. MR R&D Collaborations, Pacific Region, Siemens Medical Solutions, San Diego, CA; 4. Computer Aided Diagnosis & Therapy, Siemens Medical Solutions, Malvern, PA.
Address correspondence to G. Hermosillo (gerardo.hermosillovaladez{at}siemens.com)
Objective: To report preliminary results in radiologists' evaluation of a prototype method that corrects motion artifacts in post-pre image pairs in Breast MRI.
Materials and Methods: 3 modes of the method were tested: high-quality 3D, fast 3D and 2D. 8 readers from 7 sites evaluated 31 sequences acquired with 3D protocols, each evaluating 8 cases on average, comparing the original sequence against 2 modes. 22 of the cases were evaluated by at least 2 readers, 12 of them by 3 or 4. Motion artifacts were evaluated per slice in subtractions selected by the evaluator. MIPs of subtractions were evaluated in a subset. The protocol differed among sites: point scores for amount of skin, ridge/valley and pectoralis artifacts were assigned for each breast at 1 site (13 cases, reviewer A), systematic better/same/worse scores were assigned to each case in 3 sites (22 cases, 4 reviewers) and general verbal feedback was given in 3 sites (33 cases). About half of the cases had strong motion, 9 were coronal, 1 sagittal and 21 axial.
Results: Statistical analysis of the 13 cases read by A using the rating scale showed significant reduction of all types of artifacts in subtractions for both 3D modes (p < 0.0004), and that the high-quality mode is better than the fast mode (p < 0.018). The overall opinion of the physicians was that registration helped to interpret subtraction images, but that these appeared less sharp due to reduction in edge artifacts normallypresent, artifact was not introduced, and detail was not obscured. Opinion trend was that diagnosis would be simplified through better contrast and delineation of lesions when patient motion occurred. Some multi-focal lesions were difficult to identify without correction. There was no evidence that lesions had disappeared and artifactual pseudo-lesions were removed. All MIPs of cases with motion were judged better with correction.
Conclusion: The evaluated motion correction algorithm showed a significant reduction of motion artifacts. The point based systematic rating was the most useful approach for quantitative analysis of the effects of motion correction. Further work is needed to validate the generalization of the above preliminary results to larger case sets and readers.
112. Assessment of Parallel Imaging for Breast MRI at 1.5T
Barke L.D.*; Hendrick R.E.; Department of Radiology, Northwestern University Feinberg School of Medicine, Chicago, IL.
Address correspondence to L.D. Barke (lbarke{at}nmff.org)
Objective: To assess the benefits and restrictions of applying parallel imaging to dynamic, bilateral, high-resolution breast MRI at 1.5 T.
Materials and Methods: Phantom, volunteer, and clinical breast MRI studies without and with parallel imaging were performed on Sonata, Symphony Quantum, and Avanto 1.5 Tesla scanners (Siemens Medical Systems, Iselin, NJ) to obtain high spatial resolution, high temporal resolution contrast-enhanced bilateral studies. 3DFT (volume) gradient-echo techniques with minimum TR (3.8 to 4.6 ms) and TE (1.6 to 1.8 ms) were used to obtain fat-suppressed images with nearly isotropic voxels: submillimeter (mm) in-plane resolution with 1-mm displayed slice thickness. Parallel imaging in both frequency (MSENSE) and spatial domains (GRAPPA) were compared to non-parallel imaging techniques. Temporal resolution was kept less than 120 seconds for both non-parallel and parallel techniques. Resulting images without and with parallel imaging were compared in terms of acquisition time, signal-to-noise ratios (SNR), image artifacts, and overall image quality.
Results: Parallel imaging with either MSENSE or GRAPPA provides an acceleration factor of two without loss of spatial resolution. Keeping all spatial parameters constant, parallel imaging incurred a loss in SNR of 35–45% in phantom and clinical images without significant increase in artifacts when parallel imaging was applied in the axial or coronal planes. Parallel imaging with primary acquisitions in the sagittal plane caused significant artifacts; however, with isotropic or nearly isotropic 3D acquisitions in the axial or coronal plane, post-acquisition reconstructions could be performed in sagittal or oblique planes, along with maximum-intensity projections, without artifacts and without significant loss of spatial resolution.
Conclusion: Applying parallel imaging to breast MRI with currently available commercial software can improve temporal resolution without loss of spatial resolution and without significant loss of image quality or significant increase in image artifacts. Parallel imaging with current software restricts primary acquisitions to axial or coronal planes, but images can be reconstructed in sagittal or oblique planes without significant loss of spatial resolution when isotropic or nearly isotropic 3D images are acquired.
113. PET/CT Detection of Occult Foci of 18F-FDG Uptake in the Breast: Incidence, and Clinical Significance
Litmanovich D.1*; Gallimidi Z.1; Israel O.2; 1. Department of Diagnostic Radiology, Rambam Medical Center, Haifa, Israel; 2. Department of Nuclear Medicine, Rambam Medical Center, Haifa, Israel.
Address correspondence to D. Litmanovich (d_litmanovich{at}rambam.health.gov.il)
Objective: To evaluate the frequency and clinical significance of occult foci of 18F-FDG in the breast in patients undergoing PET for indications other than breast cancer.
Materials and Methods: 6,692 cancer patients referred for 18F-FDG PET/CT over a period of 36 months were retrospectively reviewed. All female patients with known or suspected malignancy other than breast cancer were included. Patients with suspected or known breast cancer were excluded. The incidence of 18F-FDG localized by PET/CT to the breast was determined, as were the correlative CT finding and the intensity of 18F-FDG uptake. The etiology and clinical significance of the findings was confirmed histologically or with long-term follow-up.
Results: Unexpected focal 18F-FDG uptake localized by PET/CT to breast tissue was identified in 16 out of 6,692 patients (0.23%). Follow-up data were available for 15 of these patients (by pathology in 6, by repeat PET/CT in 2, and with clinical follow-up in 7). Malignancy was histologically proven in 4 patients (2 primary breast cancers, 1 relapsed Non-Hodgkin's lymphoma and 1 metastasis of melanoma) and highly suspected (clinically and/or by PET/CT findings) in 5. Benign breast lesions were diagnosed by biopsy in 2 patients (ductal hyperplasia and fibroadenoma). In 4 patients clinical and imaging follow up of mean 19.3 months showed no further evidence of malignancy. There was no statistically significant difference in average maximal standardized uptake of FDG in malignant (4.09) and benign (2.14) lesions (p = 0.068). 9 patients had a focal lesion in the breast on CT (malignant = 5, benign = 4), while CT was negative in 6 patients (malignant = 2, benign = 4). There was no correlation between the presence of a lesion identified with CT and the intensity of 18F-FDG uptake on PET.
Conclusion: In 26.7% of patients, occult foci of 18F-FDG uptake in the breast was due to a proven malignant, unsuspected lesion. In an additional 33% a malignancy was highly suspected. Therefore, an incidental finding of focal 18F-FDG uptake in the breast, although rare, warrants further evaluation with mammography, ultrasound and if necessary MRI to identify the etiology of the finding.
* Will present paper
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
