| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Department of Radiology, Memorial Sloan-Kettering Guttman Diagnostic Center, 55 Fifth Ave, 12th Floor, New York, NY 10003. Received January 28, 2002; revision requested March 21; final revision received October 4; accepted October 14. Address correspondence to L.R.L. (e-mail: latrentl@mskcc.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Retrospective review was performed of 654 consecutive breast MR examinations performed from November 1999 to February 2001. This yielded 64 patients with 93 suspicious, nonpalpable, mammographically occult lesions evident on MR images and recommended for biopsy, for which directed US assessment was performed. Images, pathologic records, and medical records of these patients were reviewed for imaging findings, patient demographics, and histopathologic findings from subsequent biopsy. Statistical analysis was performed with the Fisher exact test.
RESULTS: The median size of MR abnormalities was 0.9 cm (range, 0.3–5.0 cm). The pattern of enhancement was a focal mass in 76 (82%) lesions and nonmass in 17 (18%). A US correlate was identified in 21 (23%) lesions detected with MR—19 (25%) of 76 focal mass and two (11%) of 17 nonmass lesions. Carcinoma was found in nine (43%) of these lesions, of which seven (78%) were invasive carcinoma and two (22%) were ductal carcinoma in situ (DCIS). Among the lesions without a US correlate, 10 (14%) yielded carcinoma, of which five (50%) were invasive carcinoma and five (50%) were DCIS. The frequency of cancer was significantly higher for lesions that were detected with MR imaging and had a US correlate than for those that did not have a US correlate (43% vs 14%, P = .01). Overall, 19 (20%) lesions detected with MR imaging and examined with US yielded carcinoma at biopsy, and nine (47%) malignant lesions were seen on US images. US depicted seven (58%) of 12 invasive cancers and two (29%) of seven instances of DCIS.
CONCLUSION: The likelihood of carcinoma was significantly higher among lesions with a US correlate (43% carcinoma) than lesions without a US correlate (14% carcinoma).
© RSNA, 2003
Index terms: Breast neoplasms, diagnosis, 08.31, 08.32 • Breast neoplasms, MR, 08.1214 • Breast neoplasms, US, 08.1298
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Alternatively, ultrasonography (US) is inexpensive and widely available. Few studies have addressed the usefulness of directed investigation with US of breast lesions that were first detected with MR imaging, specifically the likelihood of finding a US correlate, the histopathologic importance of detecting a correlate, and the ability of US to affect clinical management in these circumstances. The purpose of this study was to determine the utility and histopathologic importance of US depiction of breast lesions that were detected with MR.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
In 236 (36%) of 654 MR examinations, a biopsy was recommended for a suspicious finding. Criteria for classification of a lesion as suspicious included spiculated or indistinct borders, heterogeneous or rim enhancement, and interval change since prior breast MR examinations. Although morphology was the most important factor in determining the need for a biopsy, kinetics of contrast enhancement were used for equivocal lesions. For example, if a lesion had morphologic features that indicated it was probably benign but the time–signal intensity curve demonstrated washout, the lesion was classified as suspicious (12).
In 86 (36%) of 236 women, directed US was recommended to look for a US correlate amenable to biopsy or localization. The decision to recommend directed US correlation was at the discretion of the radiologist who interpreted the MR images. Although there were no standardized criteria, correlative US was recommended for clinically and mammographically occult lesions that were detected with the use of MR imaging and of a sufficient size (>4 mm) and focality to be detected with US. There was no difference among the five radiologists in the frequency with which US was recommended.
Of these 86 women, those with a mammographic or palpable correlate to the lesion detected with MR (n = 8) and those who did not subsequently undergo US (n = 9) were excluded. This yielded 69 women with 101 nonpalpable, mammographically occult lesions that were evident on MR images. Prior to biopsy of these suspicious lesions, directed US was performed. Four women with five lesions were excluded because they refused to undergo a biopsy. One woman with three lesions was excluded because performance of biopsy was deferred due to stage IV disease. The remaining 64 women with 93 lesions constitute the population of this study.
Patients had a median age of 51 years (range, 28–79 years). Of the 64 patients, 23 (36%) underwent a unilateral MR examination, and 41 (64%) underwent a bilateral MR examination. Among these women, 61% had biopsy-proven recent or remote ipsilateral (n = 20) or contralateral (n = 19) breast cancer at the time of the MR imaging. Indications for an MR examination included screening, 30 (47%); preoperative assessment of extent of disease, 12 (19%); search for an occult primary tumor, seven (11%); postoperative examination or rule out residual disease, six (9%); symptomatic, five (8%); abnormal mammogram, three (5%); and cyst aspiration yielding atypia, one (1%). The indications for performance of screening MR were prior history of breast carcinoma, BRCA1 or BRCA2 positivity, strong family history, or prior history of atypical ductal hyperplasia or lobular carcinoma in situ. The five symptomatic patients in this study reported unilateral nipple discharge (n = 3), a palpable mass (n = 1), or pain (n = 1).
Diagnostic Breast MR Imaging Technique
At our institution, breast MR examinations were performed with the patient prone in a 1.5-T commercially available imager (Signa; GE Medical Systems, Milwaukee, Wis) with use of a dedicated surface breast coil (MRI Devices, Waukesha, Wis). Our imaging sequence included a transverse T1-weighted localizing sequence followed by a sagittal T2-weighted sequence (4,000/85 [repetition time msec/echo time msec]). A T1-weighted three-dimensional, fat-suppressed fast spoiled gradient-echo sequence (17/2.4; flip angle, 35°; bandwidth, 31.25) was then performed before and three times after a rapid bolus injection of 0.1 mmol/kg gadopentetate dimeglumine (Magnevist; Berlex, Wayne, NJ).
Contrast material was administered as a rapid bolus injection through an indwelling intravenous catheter that was placed in the patient’s arm. Image acquisition began immediately after administration of the contrast material and saline bolus. Images were obtained sagittally, which resulted in an acquisition time per volumetric acquisition of less than 2 minutes each. Total examination time per breast, including three examinations that were performed after the administration of contrast material, was approximately 15 minutes. Section thickness was 2 mm without an intersection gap with a 256 x 192 matrix and a 16–18-cm field of view. Frequency was in the anteroposterior direction. Subtraction images were obtained by subtracting the precontrast images from the first postcontrast image on a pixel-by-pixel basis.
Breast US Technique
Breast US examinations were performed by one of five breast imaging radiologists (L.R.L., E.A.M., A.F.A., D.D.D., or L.L.) with a 7–10-MHz linear transducer (128-XP, Acuson, Mountain View, Calif; or HDI 5000, Advanced Technology Laboratories, Bothell, Wash). MR images were available for direct correlation when the US images were obtained.
Data Collection and Analysis
For these 93 lesions in 64 women, medical records, histologic findings, and the results of imaging studies were reviewed by one author (L.R.L.). Imaging information, including density of breast tissue and the number, description, size, and location of breast lesions, was obtained from the imaging report. Criteria for breast density were the same as those used for mammography (0%–25% parenchyma = fatty, 25%–50% = mild, 50%–75% = moderate, and 75%–100% = dense). A breast MR lexicon was not finalized at the time of this study; therefore, lesions detected with MR imaging were described by the interpreting radiologist as either mass or nonmass. Nonmass lesions were further classified as linear, nodular, both linear and nodular, or clumped.
Medical records were examined by one author (L.R.L.) for patient age, indication for breast MR imaging, and history of prior breast carcinoma. Histopathologic reports were reviewed by one author (L.R.L.) for pathologic size and specific histologic type. Statistical analysis was performed with the Fisher exact test with use of statistical software (Epi-Info; Centers for Disease Control and Prevention, Atlanta, Ga).
Biopsy Methods
Eighteen patients with 21 lesions that were evident on US images underwent a biopsy. Nine patients with 11 (52%) lesions subsequently underwent a US-guided biopsy. Nine patients with 10 (48%) lesions underwent MR-guided localization (six lesions), US-guided localization (two lesions), quadrantectomy (one lesion), or mastectomy (one lesion). The remaining 46 patients had 72 US occult lesions that were evident on MR images. Forty (87%) patients underwent MR-guided needle localization and surgical biopsy, five (11%) underwent mastectomy, and one (2%) underwent MR-guided biopsy at another institution. Each MR- or US-guided localization was reviewed by the radiologist who performed the procedure to ensure that the pathologic findings were concordant with the imaging findings. If the findings were discordant, a postoperative MR examination was performed to confirm retrieval of the suspicious lesion.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The median size of the MR abnormality in these 93 lesions was 0.9 cm (range, 0.3–5.0 cm). Forty-seven lesions were located in the right breast and 46 in the left. Seventy-six (82%) lesions were described as mass, and 17 (18%) were described as nonmass. The nonmass lesions were described as linear or both linear and nodular in nine, nodular in seven, and clumped in one.
Carcinoma was present in 19 (20%) lesions (median size, 1.0 cm), of which 12 (63%) were invasive carcinoma and seven (37%) were DCIS. The invasive component was ductal in 11 (92%) of 12 lesions and multifocal lobular in one (8%). Carcinoma was found in 16 (21%) of 76 mass lesions and three (18%) of 17 nonmass lesions. Most (11 [69%] of 16) malignant mass lesions were invasive carcinoma, and most (two [67%] of three) malignant nonmass lesions were DCIS.
Histologic Characteristics and US Correlation
Directed US identified a correlate for 21 (23%) of 93 lesions that were evident on MR images (Table 1). A US correlate was seen in nine (47%) of 19 malignant lesions versus 12 (16%) of 74 benign lesions (P = .01). A US correlate was observed more frequently for mass lesions that were evident on MR images than nonmass lesions and for invasive carcinoma rather than DCIS, but these differences were not statistically significant. The frequency of detecting a US correlate was independent of breast density.
|
|
|
|
|
|
|
|
|
Overall, 10 (53%) of 19 carcinomas detected with MR in this study were detected with only MR imaging. The mean size of these lesions was 1.1 cm (range, 0.6–2.1 cm). Pathologic study yielded pure DCIS in five patients, DCIS with microinvasion in one patient, and DCIS with invasive carcinoma in four patients.
Biopsy Results of Lesions Detected with US
Of the 21 lesions that were evident on the US images of 18 patients, 11 (52%) lesions subsequently underwent US-guided biopsy, which yielded benign results in six and malignancy in five. The benign findings of fibroadenoma (n = 2), fibrosis (n = 2), and benign lymph node (n = 2) were believed to be concordant with the imaging findings, and these patients did not undergo subsequent surgery. Biopsy of five lesions in four patients yielded invasive ductal carcinoma. All four of these patients had a known synchronous ipsilateral or contralateral breast cancer at the time of the MR examination and underwent a mastectomy as their first surgical procedure.
Preoperative US-guided biopsy was not performed in 10 (48%) of the 21 lesions that were evident on US images for the following reasons: uncertain correlation with the MR images (n = 3), subtle nature of the US finding (n = 2), suspicion of a papillary lesion (n = 2), recent result of an ipsilateral axillary node biopsy was positive for breast carcinoma (n = 2), ordiscretion of the surgeon (n = 1). Four of the five patients with uncertain correlation with MR images or subtle US findings underwent MR-guided localization that yielded fibrosis/lymph node in one patient, fibrosis/papilloma in two patients, and DCIS in one patient. One patient with a subtle US finding had a diagnosis of multifocal carcinoma that was based on US-guided biopsy of an additional lesion evident on MR images; this patient underwent mastectomy, which yielded invasive ductal carcinoma in the expected location from the MR image. One patient in whom an ipsilateral axillary node biopsy was positive for breast carcinoma did not undergo a preoperative percutaneous biopsy because the US and MR images of the lesion appeared suspicious. She underwent US-guided needle localization and surgical excision, which yielded carcinoma. The other patient with axillary metastasis underwent quadrantectomy without localization at the discretion of the surgeon. Additionally, two patients had previously undergone surgical biopsy, which yielded papilloma. A retrospective review of papillary lesions diagnosed at core biopsy by Liberman et al (13) suggested that primary surgical excision may be prudent; therefore, we recommended surgical excision in these two patients because we had a strong suspicion that a papillary lesion was present. Surgical excision yielded benign papilloma in both cases.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
When an MR abnormality cannot be localized with a physical examination or mammography, the biopsy methods are limited. MR-guided percutaneous biopsy is being developed at a few centers but is not yet widely available (8–10). MR-guided needle localization can be performed with commercially available equipment, but this procedure requires experience and necessitates surgical biopsy (11). If a lesion that is detected at MR can be depicted with US, it may be amenable to US-guided percutaneous biopsy. This procedure could spare the patient a surgical biopsy if the percutaneous biopsy yields benign results and could facilitate preoperative planning if it yields malignancy (18).
Among our patients who had suspicious lesions on MR images and who underwent directed US, a US correlate was identified in 23%. Carcinoma was found in 43% of lesions that had a US correlate versus 14% of lesions that lacked a US correlate. A US correlate was more often observed for mass than for nonmass lesions (25% vs 12%) and for invasive carcinoma than for DCIS (58% vs 29%), but these differences were not statistically significant. We hypothesize that with more cases of breast cancer, we might observe a substantially higher frequency of US correlates for invasive carcinoma than for DCIS because of the higher sensitivity of US for invasive than in situ carcinoma (19). Lesion size did not affect the likelihood of a US correlate in our study.
The 14% frequency of carcinoma among lesions detected with MR and that lacked a US correlate is within the 9%–47% range of reported positive predictive values for mammographically guided needle localization (20). Brown et al (21) described 30 "incidental" lesions that appeared only on MR images; in seven, there was tissue diagnosis (yielding carcinoma in one), 22 underwent stable mammographic follow-up for mean of 22 months (of which two also underwent stable MR follow-up at 2 and 6 months, respectively), and one underwent no follow-up. The authors concluded that "unless malignancy is diagnosed elsewhere in the breast, incidental foci identified at MR imaging are unlikely to be malignant" (21). Our experience does not support this conclusion. Because areas of carcinoma found on MR images may not be evident on mammograms, mammographic stability for less than 2 years cannot ensure benignity in a lesion detected with MR imaging only. Our data indicate that if a suspicious lesion is found at MR imaging, tissue diagnosis is warranted, regardless of whether the lesion has a correlate on other images or carcinoma has been diagnosed elsewhere in the breast. If we exclude the 39 patients in our study with a remote or recent history of breast cancer, we still found nine malignancies in the remaining 25 patients only by using MR imaging.
Among the 18 patients with 21 lesions with US correlates, nine patients with 11 lesions underwent US-guided biopsy. Biopsy of the US correlate obviated a surgical procedure in all of these patients, five of whom had benign results and four of whom had malignant results. The benign lesions did not require surgical excision, and the patients with malignant results underwent mastectomy as their first surgery after the MR examination. Among 64 patients in the study, US-guided biopsy obviated a surgical procedure in nine (14%) patients, and US-guided needle localization was performed in an additional two (3%) patients; however, these data indicate that directed US benefitted a minority of patients who underwent this examination.
Of note, 10 of 21 lesions that were depicted by US were not amenable to US-guided biopsy. In half of these cases, this was due to the subtle nature of the US image or uncertainty as to correlation with the lesion detected with MR. During US, the patient is in the supine or supine-oblique position, but during MR imaging, the patient is in the prone position. In women with large breasts, movement of the breast can cause lesions to project in different axes on the two studies. Careful correlation with MR images at the time of the US examination, including attention to size, shape, and depth of the MR abnormality, is crucial to ensure that the appropriate area of the breast is examined with US and that the US and MR images are concordant. At our institution, we obtain sagittal MR images with the nipple in profile rather than coronal images because this positioning mimics the mammographic mediolateral view and enhances mammographic-MR correlation. With sagittal imaging, we also have the capability for transverse reformatting, if desired.
Our study has several limitations. The majority, 39 (61%) of 64 patients, had a history of synchronous or metachronous breast cancer that may have inflated the number of cancers detected. Only a subset of patients with suspicious lesions at MR was referred for directed US. This was either because the lesion detected on the MR image was palpable or mammographically evident, or because the interpreting radiologist believed the lesion had a low likelihood of being evident on a US image. Also, some lesion subgroups, such as nonmass enhancing lesions, occurred relatively infrequently. However, even with preselection of the lesions considered most likely to have US correlates, only 23% were visible on US images. Additionally, we performed US after, not before, MR imaging. Although it is possible that some of the US-evident lesions could have been detected with screening breast US prior to MR imaging, Panizza et al (22) showed a higher likelihood of identifying lesions at "second-look" directed US following MR imaging rather than during a US screening examination. Although all US in this study was performed by five breast imaging radiologists and carefully correlated with additional imaging studies, US is operator and equipment dependent and results may vary at other centers.
In conclusion, we found a US correlate in 23% of suspicious lesions detected at MR that also underwent directed US examination. The likelihood of carcinoma was significantly higher in lesions detected with MR that had a US correlate (43% carcinoma) than in lesions that lacked a US correlate (14% carcinoma). Absence of a US correlate did not, however, spare the need for biopsy of suspicious lesions detected with MR. Our findings indicate the importance of continued research in performance of MR-guided biopsies so that histologic diagnosis may be obtained for suspicious lesions detected with MR imaging, regardless of whether they can be identified with other imaging modalities.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, L.R.L.; study concepts, L.R.L., E.A.M., L.L.; study design, L.R.L., L.L.; literature research, L.R.L., J.H.M., L.L.; clinical studies, L.R.L., J.H.M., E.A.M., D.D.D., A.F.A., L.L; data acquisition, all authors; data analysis/interpretation, L.R.L., J.H.M., E.A.M., L.L.; statistical analysis, L.R.L., L.L.; manuscript preparation, definition of intellectual content, editing, revision/review, and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. A. Causer, C. A. Piron, R. A. Jong, and D. B. Plewes Preliminary In Vivo Validation of a Dedicated Breast MRI and Sonographic Coregistration Imaging System Am. J. Roentgenol., October 1, 2008; 191(4): 1203 - 1207. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Bartella, S. B. Thakur, E. A. Morris, D. D. Dershaw, W. Huang, E. Chough, M. C. Cruz, and L. Liberman Enhancing Nonmass Lesions in the Breast: Evaluation with Proton (1H) MR Spectroscopy Radiology, October 1, 2007; 245(1): 80 - 87. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Causer, R. A. Jong, E. Warner, K. Hill, J. W. Wong, B. N. Curpen, and D. B. Plewes Breast Cancers Detected with Imaging Screening in the BRCA Population: Emphasis on MR Imaging with Histopathologic Correlation RadioGraphics, October 1, 2007; 27(suppl_1): S165 - S182. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Kuhl The Current Status of Breast MR Imaging * Part I. Choice of Technique, Image Interpretation, Diagnostic Accuracy, and Transfer to Clinical Practice Radiology, August 1, 2007; 244(2): 356 - 378. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Liberman, A. E. Holland, D. Marjan, M. P. Murray, L. Bartella, E. A. Morris, D. D. Dershaw, and R. T. Wynn Underestimation of Atypical Ductal Hyperplasia at MRI-Guided 9-Gauge Vacuum-Assisted Breast Biopsy Am. J. Roentgenol., March 1, 2007; 188(3): 684 - 690. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. A. Causer, C. A. Piron, R. A. Jong, B. N. Curpen, C. A. Luginbuhl, J. E. Glazier, E. Warner, K. Hill, J. Muldoon, G. Taylor, et al. MR Imaging-guided Breast Localization System with Medial or Lateral Access Radiology, August 1, 2006; 240(2): 369 - 379. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Bartella, L. Liberman, E. A. Morris, and D. D. Dershaw Nonpalpable mammographically occult invasive breast cancers detected by MRI. Am. J. Roentgenol., March 1, 2006; 186(3): 865 - 870. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Orel, M. Rosen, C. Mies, and M. D. Schnall MR Imaging-guided 9-gauge Vacuum-assisted Core-Needle Breast Biopsy: Initial Experience Radiology, December 1, 2005; 238(1): 54 - 61. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Kawase, D. J. DiMaio, S. L. Tucker, T. A. Buchholz, M. I. Ross, B. W. Feig, H. M. Kuerer, F. Meric-Bernstam, G. Babiera, F. C. Ames, et al. Paget's Disease of the Breast: There Is a Role for Breast-Conserving Therapy Ann. Surg. Oncol., May 1, 2005; 12(5): 391 - 397. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. E. Deurloo, S. H. Muller, J. L. Peterse, A. P. E. Besnard, and K. G. A. Gilhuijs Clinically and Mammographically Occult Breast Lesions on MR Images: Potential Effect of Computerized Assessment on Clinical Reading Radiology, March 1, 2005; 234(3): 693 - 701. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Lilienstein, B. L. Daniel, and D. M. Ikeda In Vivo Sonography Through an Open MRI Breast Coil to Correlate Sonographic and MRI Findings Am. J. Roentgenol., March 1, 2005; 184(3_supp): S49 - S52. [Full Text] [PDF]
|
|
|
|
|
|
L. Liberman, E. A. Morris, D. D. Dershaw, C. M. Thornton, K. J. Van Zee, and L. K. Tan Fast MRI-Guided Vacuum-Assisted Breast Biopsy: Initial Experience Am. J. Roentgenol., November 1, 2003; 181(5): 1283 - 1293. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
