| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||
ABSTRACT |
Monday, May 7, 1:30 PM–3:30 PM
Abstracts 068-078
Moderator(s): Lonie Salkowski and Alexis Nees
1:30 PM
Keynote Address: Current Utilization Trends and Implications: MRI, CAD, Digital Mammography
Larry Bassett, UCLA, Los Angeles, CA
1:40 PM
068. BI-RADS 3 Lesions on MRI: Classification, Outcomes, and Frequency of Malignancy
Moy L.*; Levy A.; Mercado C.; Toth H. New York University, New York, NY
Address correspondence to L. Moy (moyl02{at}med.nyu.edu)
Objective: To identify the types of breast lesions that are classified as "probably benign" and the frequency of malignancies among these lesions detected on MR imaging.
Materials and Methods: Retrospective review of 1246 breast MR exams performed over a nine-year period identified 112 (9%) lesions which were placed into the BI-RADS 3 category. These MR exams were reviewed to determine the type of lesion that was placed into this category and the frequency of subsequent malignancy in these women. Radiology and pathology databases were reviewed.
Results: Follow-up MRI was performed in 89 (79%) of 112 women, with a mean follow-up time of 8 months (range 1.5–30 months). Eighteen women underwent a biopsy, seven for a finding that was reinterpreted as suspicious on follow-up exam, eight due to progression on follow-up imaging, three for new palpable abnormality in the region of the MRI abnormality. Six (5.3%) malignancies were found, three DCIS and three invasive ductal carcinomas. Of 112 lesions, 36 (31.3%) were asymmetric non-mass enhancement, likely related to changes in the patients' menstrual cycle. Twenty-two (61%) of these women were imaged during the correct time of the menstrual cycle. An MRI was performed in 18 of 22 (81.2%) women six months later demonstrated the lesion had resolved. No cancers were detected. Twenty (17.8%) of 112 lesions were subcentimeter, oval or lobulated, heterogeneous enhancing masses with type 2 curve, (slow or medium initial rise). No cancers were detected. Twenty-five (22.3%) of 112 lesions were linear or segmental enhancement with stippled or heterogeneous distribution. Two were reinterpreted as clumped enhancement on follow-up MRI, both cases were DCIS. Thirty-one (27.7%) of 112 lesions were interpreted as a focal region of non-mass enhancement. On follow-up, two were reinterpreted as an irregular mass, two had progressed. At surgery three invasive ductal carcinomas and one DCIS was found. Three of 4 malignancies were downgraded from BI-RADS 4 to 3 after a negative second-look ultrasound.
Conclusion: The frequency of malignancies among "probably benign" breast lesions detected on MR imaging was 5.3%. We believe that short interval follow-up is worthwhile, especially for changes likely secondary to the menstrual cycle. Despite adopting the BI-RADS MR lexicon, there are discrepancies in describing lesion morphology. Also, strict adherence to guidelines was not always followed.
069. Incidental MR-Detected Lesions—How Do We Deal with Them?
Abe H.*; Newstead G.; Sennett C.; Shimauchi A.; Schmidt R. University of Chicago, Chicago, IL
Address correspondence to H. Abe (habe{at}uchicago.edu)
Objective: To assess the significance of the incidental lesion detected initially on MR.
Materials and Methods: Retrospective review was made on1304 consecutive breast MR studies performed at our institution between July 1, 2003 and July 31, 2005, to look for cases with unexpected enhancing mass lesion(s). Lesions visualized on prior studies (including mammography, ultrasound, and MR) were excluded. Palpable lesions were also excluded. The location and measurement of the lesion was documented, and lesions were characterized according to ACR Lexicon. Results of subsequent imaging work-up, biopsy and/or surgery, and imaging follow-up were reviewed. The MR (1.5 T) protocol was: axial T2-weighted FSE, dynamic contrast study including 1 pre- and 4 postcontrast coronal T1-weighted SPGR.
Results: 252 unexpected enhancing mass lesion(s) were found in 198 cases (15%) on the MR images. Size of these lesions was; 5 mm or less in 91 lesions, 6–10 mm in 111 lesions,11–20 mm in 43 lesions, and more than 20 mm in 7 lesions. Additional study, including imaging work-up, biopsy and/or surgery and imaging follow-up, was performed to 230 lesions in 177 cases. Secondlook ultrasound was performed on 192 lesions, and 124 of them (65%) were successfully correlated. Pathologic results were available in 126 (55%) lesions; 61 by US-guided biopsy, 40 by MR-guided biopsy, 17 by surgery, and 8 by FNA. 38 (31 cases) lesions (17%) were malignant. Sizes of these malignant lesions was; 5 mm or less in 8 lesions, 6–10 mm in 13 lesions,11–20 mm in 13 lesions, and more than 20 mm in 4 lesions. Malignant lesions were found in the 19 cases with known cancer, 6 high-risk screening cases, and 6 cases with pathologically suspicious for breast primary cancer. Follow-up MR was performed to 56 lesions in 42 cases in 3–24 months. Results of the follow-up were; 21 lesions disappeared or significantly decreased, 8 lesions found out to be normal parenchymal enhancement, 25 lesions showed no change, and 2 lesions increased in size. One cancer (lesion that increased in size) was found from these follow-up lesions.
Conclusion: Unexpected MR-enhancing mass lesion should be worked up with ultrasound, as many of them can be correlated and better characterization is possible. If lesions are not sonographically correlated, MR-guided biopsy should be considered depending on MR findings and patient's risk status, because malignant lesions are likely to be detected in patients with known cancer and high-risk patients.
070. Pathologic Diagnoses of MRI-Detected Breast Lesions in High-risk Patients
Raza S.1*; Teo S.2 1. Brigham & Women's Hospital, Boston, MA; 2. KK Women's & Children's Hospital, Singapore, Singapore
Address correspondence to S. Raza (sraza1{at}partners.org)
Objective: To report the pathologic diagnoses of breast lesions visible only on MRI in different high-risk patient groups.
Materials and Methods: We retrospectively reviewed all MRI-directed breast biopsies (MRI-guided wire localized surgical excision or MRI-guided 9G vacuum-assisted core needle biopsy) performed at our institution from May 2004 to August 2006. During this period, we performed biopsies on a total of 156 lesions in 142 patients. Of these, we identified 71 patients as being at high risk for breast cancer. This high-risk group was divided into the following categories: 29 patients had personal history of previous invasive breast cancer; 7 patients had previously been diagnosed with ductal carcinoma in situ (DCIS); 8 patients had previous biopsy diagnoses of high-risk lesions (ADH, LCIS); 19 patients had family history of breast cancer (first-degree relative); and 8 patients were genetic mutation carriers (7 BRCA 1 or 2, 1 p53). All patients had recent mammograms and second-look ultrasound after MRI, unless the area of MRI abnormality was too small or indistinct to be detected by ultrasound. In all patients the lesion was detectable on MRI only.
Results: Of the total study population of 71 patients, 16 (22.5%) were found to have malignant disease on biopsy. In the sub-group of 29 patients with history of previous invasive breast cancer, 6 (20.7%) had malignancy, including 3 (10.3%) invasive cancers and 3 (10.3%) DCIS. Of the 7 patients with history of previous DCIS, 4 (57.1%) malignancies were found, including 2 (28.6%) invasive cancers and 2 (28.6%) DCIS. Of the 19 patients with family history of breast cancer, there were 4 (21.1%) malignancies, including 1 (5.3%) invasive cancer and 3 (15.8%) DCIS. In the two groups of 8 patients each with prior high-risk lesions and genetic mutations respectively, there was 1 (12.5%) patient with DCIS in each group and no patient with invasive cancer.
Conclusion: In our study population the positive predictive value of MRI for the detection of breast malignancy in all high-risk patients is 22.5%. This is similar to the positive predictive value of MRI in non high-risk patients. The positive predictive value for the individual high-risk groups is the highest for patients with previous DCIS (57.1%) and lowest for patients with prior highrisk lesions and genetic mutations (12.5%). Currently these values are limited by the small number of patients in each group, and will require continued outcomes analysis.
071. MRI as Part of the Pre-operative Assessment of Lobular Carcinoma—An Update
Sharma N.*; Manuel D. D.; Dall B. J. MRI Department, Cookridge Hospital, Leeds, United Kingdom
Address correspondence to N. Sharma (nishas{at}doctors.org.uk)
Objective: This is an audit of our practice with retrospective review of 79 cases. We have previously demonstrated that MRI provides important additional information in the diagnosis of lobular carcinoma (BJS 2002,89,1296–1301). It is our practice to offer MRI preoperatively to women with a core biopsy diagnosis of lobular carcinoma who are eligible for wide local excision (WLE).
Materials and Methods: Three breast radiologists (2 consultant radiologists and a clinical fellow in breast radiology) undertook retrospective consensus review of the MRI examinations and developed an MRI proforma categorizing invasive lobular carcinoma (ILC) into either a mass or an area of enhancement. If the lesion was seen on T2-weighted images and postcontrast subtraction T1-weighted images it was described as a mass lesion. If only seen on postcontrast subtraction T1-weighted images it was described as an area of enhancement. The mass lesions were further subdivided into focal and multifocal lesions. We also recorded the enhancement profiles. The MRI examinations were reviewed with the mammography and ultrasound findings. Extra disease on MRI, which was either additional foci of disease or disease more than double the expected diameter was correlated with the surgical pathology report.
Results: Final histology confirmed lobular carcinoma in 62 cases; 17 cases had alternative diagnoses. Core biopsy reported lobular features in 12/17 cases; 5/17 cases were inappropriately referred. Based on the MRI appearances, lobular carcinoma can be subdivided into areas of enhancement (24%, n = 15) and mass lesions (74%, n = 46); 1 case was MRI occult. The mass lesions can be further subdivided into focal (29%,n = 18) or multifocal lesions (45%, n = 28). The enhancement curves are variable; 28 (47%) had a malignant curve, 20 (33%) had a suspicious curve and 12 (20%) had an indeterminate or benign curve. Review of subsequent treatment suggests MRI led to a change in treatment in 50% of patients (n = 31) by demonstrating more extensive disease than predicted by mammography and ultrasound. MRI overestimated disease in 1 case, potentially resulting in a mastectomy. Mammography, ultrasound and MRI underestimated disease in 1 case with no clinical consequences. Only 1 patient had 2 surgical procedures but there was no residual disease at the second operation. Without MRI 31 patients would probably have required two procedures.
Conclusion: We consider this to be a good use of a scarce and expensive resource. As always, case selection is key.
072. MRI Enhancement Patterns and Volume Kinetics of Unbiopsied Breast Cancers
Comstock C. E.*; Eradat J.; Middleton M. University of California, San Diego, San Diego, CA
Address correspondence to C. Comstock (ccomstock{at}ucsd.edu)
Objective: The purpose of the study is to correlate the MRI enhancement patterns and volume kinetics with the type, grade, and size of unbiopsied breast cancers.
Materials and Methods: A retrospective review was performed of dynamic contrast enhanced (DCE) breast MRI studies from 2003 through 2006. Thirty-seven unbiopsied cancers were identified in 32 patients for whom there was subsequent imaging, core needle biopsy and surgical pathology correlation. DCE MRI protocol: Siemens Symphony 1.5 T, both breasts, fat saturation, axial, 348 x 512 matrix, 1.6 mm slice thickness, 88 slices, 5 postcontrast one-minute sequences were obtained. Uncorrected subtracted images were generated by the Syngo Siemens software package at time of study acquisition. Images registration correction, subtraction, lesion segmentation, enhancement rate analysis, and volume kinetic analysis were performed using the Confirma CADstream software package. A rapid enhancement pattern is classified as a 100% increase at 2 minutes and a medium enhancement pattern as a 50% increase at 2 minutes. Tumor grades are based on the modified Bloom Richardson (mBR) criteria.
Results: The 37 unbiopsied breast cancers were composed of the following tumor types: 7 DCIS, 16 invasive ductal, 4 invasive mixed ductal/lobular, 7 invasive lobular, 2 tubular and 1 colloid. The average percent rapid/medium enhancement ratios were: DCIS 36/64, Invasive ductal 57/43, Invasive mixed ductal/lobular 55/45, invasive lobular 23/77, tubular 20/80 and colloid 76/24. The average percent washout was: DCIS 0, invasive ductal 14, invasive mixed ductal/lobular 2, invasive lobular 0.1, tubular 0 and colloid 0. None of the DCIS, tubular carcinomas or the colloid carcinoma showed washout. Only one of the invasive lobular carcinomas showed washout (with only 1% volume). Three of the 16 invasive ductal carcinomas did not show washout. One was mostly DCIS with foci of microinvasion. The other 2 were 0.7 cm and 0.9 cm grade I and II invasive ductal carcinomas, respectively. One of the invasive mixed ductal/lobular carcinomas did not show washout (1.0 cm grade I)
Conclusion: Our initial data suggest that enhancement rate and percent washout correlate with the type of breast cancer. There was no apparent correlation of tumor size or grade with the percent washout for invasive ductal carcinomas.
073. MRI Detection of Breast Cancer: Computer-Aided Detection using a Novel Combined Morphology and Kinetic-Based System
Brem R.2*; Penn A.1; Thompson S.1 1. OnCAD, Bethesda, MD; 2. The George Washington University, Washington, DC
Address correspondence to R. Brem (rbrem{at}mfa.gwu.edu)
Objective: Commercially available breast MRI CAD systems rely on kinetic angiogenesis maps to identify suspicious areas based on enhancement characteristics. Recent studies have demonstrated that morphology is a critical feature of differentiating benign from malignant breast lesions. Prior studies have incorporated subjective, radiologist-determined lesion morphology. However, objective assessment of lesion morphology might improve exam interpretation. To investigate sensitivity and false-positive (FP) rates of a novel objective morphology and kinetic-based computer-aided detection (ONCAD) on clinical breast MRI examinations
Materials and Methods: Breast MRI cases with biopsy-proven malignancies or who were cancer-free with 1 year follow-up were included. 23 cancers and 23 cancer-free cases were identified. There were 15 IDC, 2 ILC, and 6 DCIS. Imaging included a 70-sec dynamic sequence of 3D T1 images. The CAD system (ONCAD) generates objective measures of morphological blooming (MB) and subjective pharmacokinetic maps to evaluate kinetic and morphological features. Enhancing regions that have MB = 9 are positive as are MB = 7 or 8 if: (a) = 16% of voxels show washout kinetics or, (b) fractal dimension over the enhancement is low (an indication of a highly irregular surface). A true positive (TP) is a positive ONCAD finding which overlapped the cancer; a FP was an ONCAD positive finding within a cancer-free breast.
Results: The MB feature generated sensitivity of 87% at 2.83 FP per cancerfree breast. ONCAD detected 5/6 (83.3%) DCIS, 13/15 (86.7%) IDC, 2/2 (100%) ILC.
Conclusion: ONCAD using objective morphology as well as kinetic data in breast MRI exams detected cancers at sensitivities comparable to reader interpretation levels reported in recent multi-center studies as interpreted by expert readers.
Clinical Relevance/Application: This is the first MRI CAD that utilizes both objective morphologic as well as subjective kinetic data and has the potential to improve sensitivity, accuracy, and rapidity of breast MRI interpretation.
074. MR Spectroscopy (MRS) of the Breast at 3.0 T: Is a Positive Choline Peak Seen at Higher Field Strengths in Normal Fibroglandular Breast Tissue
Moy L.*; Salibi N.; McGorty K.; Mercado C.; Singer C.; Ruff J. New York University, New York, NY
Address correspondence to L. Moy (moyl02{at}med.nyu.edu)
Objective: Single-voxel MRS studies at 1.5 T assume that choline is detected only in malignant lesions. With the increased SNR at high field magnets, literature suggests that choline may be detected in benign lesions and in normal tissue. The purpose of this study was to evaluate the presence or absence of a choline peak from MRS of normal breast parenchyma at 3.0 T.
Materials and Methods: An IRB-approved prospective analysis was performed on 50 consecutive women who underwent dynamic contrast-enhanced breast MRI followed by MRS at the 3.0-T magnet (TIM Trio, Siemens, Invivo 4-channel breast coil). Bilateral single voxel MRS exam was performed over a region of normal parenchyma if no enhancing lesion was visualized. If a lesion was identified, MRS was performed in a region of normal enhancing tissue in the contralateral breast. MRS was defined as positive if a choline peak was seen, negative otherwise. The choline resonance peaks were not quantified. MRS findings were correlated with radiologic, pathologic, and/or clinical follow-up.
Results: Eighty MRS studies were performed on 50 women (median age 44, range, 28–74) who underwent a bilateral breast MRI. The indications were high risk screening (28%), personal history of breast cancer (22%), extent of disease evaluation (25%), and problem solving (25%). Forty-two of 80 MRS studies were performed in 34 of 50 (68%) women where shimming was successful and comprise our sample population. Fourteen (41%) were postmenopausal and 20 (59%) were premenopausal. Of 20 premenopausal women, 3 were scanned during week 1 of their cycle, 14 during week 2, 2 during week 3, 1 during week 4. Three women were on oral contraceptives, and two on hormone replacement. No choline was detected in normal tissue. A false-positive choline peak was seen in a single normal lactating patient. Careful postprocessing resolves a composite choline resonance at 3.22 ppm as normal lactose and choline peaks. BI-RADS assessment categories were: BI-RADS 1 or 2 in 26 (76%), BI-RADS 3 in 2 (6%), and BI-RADS 4 in 6 (18%), in areas separate from the sites that had MRS. Two of the BI-RADS 4 lesions had benign biopsies, two DCIS, and two invasive ductal carcinoma. No choline was detected in the two benign biopsies.
Conclusion: No choline peak is identified in normal non-lactating breast parenchyma at 3.0 T, regardless of the time of the menstrual cycle. The absence of a choline in normal fibroglandular tissue at 3.0 T may help improve the specificity of breast MRI.
075. Dynamic MR Imaging of Invasive Ductal Carcinoma: Studying Kinetics by Estrogen Receptor, Progesterone Receptor and Her2/Neu Amplification Status
Arkani S.*; Newstead G.; Abe H.; Shimauchi A.; Karczmar G. University of Chicago, Chicago, IL
Address correspondence to S. Arkani (sarkani{at}uchicago.edu)
Objective: To perform a systematic evaluation of the kinetic characteristics of 145 invasive ductal carcinoma (IDC) lesions classified by estrogen receptor (ER), progesterone receptor (PR) and Her2/Neu gene amplification status.
Materials and Methods: 138 patients with 145 histologically proven IDC lesions with known ER, PR and Her2/Neu status were selected for IRB-approved review. These lesions were classified as: ER positive (n = 101), ER negative (n = 44), PR positive (n = 76), PR negative (n = 69), Her2/Neu positive (n = 25), and Her2/Neu negative (n = 120). Dynamic MR protocol: 1 pre- and 5 postcontrast images acquired in the coronal plane using a 3D SPGR sequence with 68 s timing. Kinetic curves were generated by two experienced radiologists by using homemade software to manually trace a region of interest (ROI) around the most enhancing part of the lesion on subtraction images and plot the signal intensity versus time (or kinetic) curve. Subsequent analysis of kinetic curve shape was made according to the BI-RADS lexicon: initial rise (rapid, medium, slow) and delayed phase (persistent, plateau, washout). In addition, several quantitative parameters were derived from the kinetic curves: initial enhancement percentage (E_1), a measure of washout relative to the first postcontrast point (W_1), and the time to peak enhancement (T_peak).
Results: Overall, 92% of lesions showed rapid initial enhancement, and 74% exhibited washout curves. The average values for the kinetic parameters were: E_1 = 306%, W_1 = 1.12, T_peak = 146 seconds. ER negative lesions had a stronger initial enhancement, stronger washout, and a shorter time to peak enhancement (E_1 = 351%, W_1 = 1.36, T_peak = 114 seconds) compared with ER-positive lesions (E_1 = 286%, W_1 = 1.02, T_peak = 162 seconds), with p values < 0.03 for all parameters. PR-negative lesions exhibited a stronger washout (W_1 = 1.25) compared with PR-positive lesions (W_1 = 1.01), with p value < 0.01, but were otherwise statistically equivalent. Her2/Neu-negative and positive lesions were statistically equivalent.
Conclusion: The kinetic characteristics of ER- and PR-positive and negative lesions showed some statistically significant differences (p < 0.03), with ER-negative lesions showing the strongest initial enhancement, shortest time to peak enhancement and strongest washout. This suggests that PR and in particular ER status may be related to vasculature in a way that Her2/Neu status is not. These findings may provide new ways of understanding kinetic data presented in dynamic MR images.
076. Increased Specificity for Detection of Invasive Breast Cancer Using Combined High-resolution 3.0-T MRI and Vascular Mapping of the Breast
Schmitz A. C.2; Peters N.2; Mali W. P.2; van Diest P. J.2; van den Bosch M. A.1* 1. Stanford University Medical Center, Stanford, CA; 2. University Medical Center Utrecht, Utrecht, The Netherlands
Address correspondence to M. van den Bosch (mbosch{at}stanford.edu)
Objective: To evaluate the diagnostic accuracy of 3.0-T magnetic resonance imaging (MRI) of the breast. Furthermore, to assess the diagnostic value of the difference in total breast vascularity for distinguishing benign and malignant breast lesions.
Materials and Methods: From January 2006 to July 2006, 54 patients (age range 21–85 years) suspected of having breast cancer were prospectively enrolled in the study and underwent 3.0-T MRI of the breast. MR images were evaluated according to the MRI-BI-RADS lexicon criteria. Lesions size, number of lesions and localization in the breast, were systematically assessed. High-resolution contrast-enhanced fat-saturated T1-weighted images were used to obtain maximum intensity projections (MIPS). Breast vascularization was assessed according to Sardanelli et al. by measuring number, diameter, and length of the vessels on these MIPS. The vascular score ranged form S0 (absent or low breast vascularity) to S3 (high breast vascularity). Finally, the diagnostic accuracy of 3.0-T MRI of the breast was assessed by comparing MRI to histopathology.
Results: Histopathologic analysis of 54 lesions revealed 24 (44%) malignant lesions and 30 (56%) benign lesions. Correlation with the MRI-BI-RADS classification revealed all 23 (100%) BI-RADS 2 lesions were benign; 6/7 (86%) of the BI-RADS 3 lesions were benign and 1/7 (14%) was malignant; 1/6 (17%) of the BI-RADS 4 lesions was benign and 5/6 (83%) were malignant; and all 18 (100%) BI-RADS 5 lesions were malignant. Based on morphologic and kinetic data MRI had a sensitivity of 100% (24/24) and a specificity of 76% (23/30). Vascular scoring of the breast revealed S0 in 5 benign cases, S1 in 13 benign and 2 malignant lesions, S2 in 12 benign and 17 malignant lesions and S3 in 5 malignancies. The use of the vascular score as an adjunct to MRI upgraded the specificity to 86% without affecting the sensitivity.
Conclusion: Breast cancer was associated with increased ipsilateral breast vascularization. Combined analysis of lesion morphology and enhancement kinetics together with vascular mapping of the breast ultimately improved the specificity of breast MR imaging to 86%, without affecting sensitivity that was 100%.
077. Multislice High Spectral and Spatial Resolution MRI for Full-breast Coverage
Medved M.2*; Newstead G. M.2; Abe H.2; Fischer S.1; Zamora M. A.2; Karczmar G. S.2 1. Philips Medical Systems, Cleveland, OH; 2. University of Chicago, Chicago, IL
Address correspondence to M. Medved (mmedved{at}uchicago.edu)
Objective: High spectral and spatial resolution (HiSS) MR provides excellent diagnostic images even in absence of contrast media injection. However, HiSS datasets are extremely large and previous work has demonstrated acquisition of these images only from 2–3 slices through suspicious lesions. Here we demonstrate the application of a sagittal HiSS sequence through the entire breast in 9–13 minutes, using conventional echo-planar spectroscopic imaging (EPSI).
Materials and Methods: Twelve women were scanned after informed consent was obtained, using high resolution EPSI on 1.5 T GE SIGNA and Philips Achieva scanners. (1.00–1.25 mm spatial resolution in 3-mm-thick slices, 11-Hz spectral resolution). To generate pure water peak height images, the fat peak in each small (3–4.5 mm3) voxel was fitted using a Lorentzian function, and the fitted peak was then subtracted to leave a spectrum of water only. Maximum intensity projections (MIP) images were generated from individual water peak height images.
Results: 1) Acquisition of HiSS images with modest spectral resolution allowed coverage of the entire breast in 9–13 minutes of scan time. This increase in scan time was well tolerated by the patients. 2) Excellent fat suppression resulted in MIP images showing parenchyma against a dark background. Four lesions were identified and clearly visible. This was achieved prior to contrast agent injection. 3) Background suppression and high dynamic range resulted in improved detail and higher contrast, as compared to the standard clinical images. 4) The slightly reduced spectral resolution possibly obscures the fine structure of the water resonance, but earlier work shows that most of the image contrast is preserved.
Conclusion: HiSS images can be acquired at modest spectral resolution to allow coverage of the entire breast. Further optimization of data acquisition, by optimizing k-space sampling, reducing the spectral bandwidth, and/or using parallel imaging, could lead to faster imaging. In the clinical setting, a full-breast HiSS multislice scan could be obtained prior to contrast agent injection and areas with suspicious findings could be more closely examined at higher spectral resolution later in the scan.
078. Survey: Current National Trends and Practices in Breast MRI
Bassett L. W.2*; Eradat J.2; Dhaliwal S.2; Brenner J. R.3; Farria D. M.4; Liberman L.1 1. Memorial Sloan-Kettering Cancer Center Department of Radiology, New York, NY; 2. UCLA Department of Radiology, Los Angeles, CA; 3. UCSF Department of Radiology, San Francisco, CA; 4. Washington University Department of Radiology, St. Louis, MO
Address correspondence to L. Bassett (lbassett{at}mednet.ucla.edu)
Objective: Breast MRI is rapidly growing relatively new procedure. The purpose of this study is to define the current practices of radiologists involved in the interpretation of breast MRI at imaging centers throughout the U.S.
Materials and Methods: A national survey of the members of the Society of Breast Imaging (SBI) was undertaken, after obtain IRB approval and permission of the SBI. An invitation to participate in a web-based survey was sent via e-mail to 1305 active SBI physician members who had e-mail addresses available. The survey contained 25 questions about their facility's policies regarding breast MRI. Respondents were instructed to assign only one radiologist from their practice to complete the survey.
Results: Selected responses: We received responses from 397 (30.5%) of the 1305 members with e-mail addresses. Breast MRI was offered at 83.8 % of the facilities surveyed. 64.4% of these performed <10 per week. Only 11.8% performed >20. Performance of breast MRI was supervised by an MQSA-qualified radiologist who interprets mammograms more than 50% of their time in 66.2% of the facilities. Breast MRI was interpreted by MQSA-qualified radiologists who interpret mammograms more than 50% of their time in 73.3% of the facilities. Breast MRI was interpreted with hardcopy in 2.1%, a softcopy workstation without CAD in 46.2%, and a soft copy workstation with CAD in 43.7%. 43.9% of facilities never interpreted breast MRI without correlating mammography and/or ultrasonography, and 40.2% only rarely did so. 46% of facilities never interpreted breast MRI performed at outside facility. 68.5% offered screening breast MRI. "Second-look" ultrasound for suspicious MRI findings was always performed at 57.4%, and frequently performed at 30.8%. 69.3% of the facilities never performed MRI-guided procedures without repeating the MRI at their facility first.
Conclusion: It is hoped that the information gained in this study will be an important step in the eventual standardization of the performance and interpretation of breast MRI.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
