| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 All authors: Department of Radiology, Breast Imaging Section, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10021.
Received November 4, 2004;
accepted after revision January 28, 2005.
Supported by grant R25 CA20449 from the NIH/NCI Cancer Education
Program.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. A retrospective review was performed of 666 consecutive nonpalpable, mammographically occult lesions that had MRI-guided localization. MRI examinations were performed using a 1.5-T magnet. Lesions were measured by the interpreting radiologist before biopsy. Malignancy rate versus lesion size was determined.
RESULTS. The median MRI lesion size was 1 cm (range, 0.3-7.0 cm). Malignancy was present in 149/666 (22%) lesions, of which 80 (54%) were ductal carcinoma in situ (DCIS), 66 (44%) were invasive cancer, and three (2%) were lymphoma. The frequency of malignancy increased significantly (p = 0.0005) with lesion size, with malignancy found in one (3%) of 37 lesions less than 5 mm, 44 (17%) of 254 lesions 5-9 mm, 37 (25%) of 151 lesions 10-14 mm, 21 (28%) of 74 lesions 15-19 mm, and 46 (31%) of 150 lesions 20 mm or larger. Lesions less than 5 mm accounted for 37 (6%) of 666 lesions that had a biopsy and one (< 1%) of 149 cancers (one DCIS). Among lesions less than 10 mm, the likelihood of malignancy was highest in postmenopausal women (22% malignant) and in the extent of disease setting (22% malignant), and lowest in premenopausal women (10% malignant) and in the high-risk screening setting (10% malignant).
CONCLUSION. The PPV of biopsy for lesions identified at breast MRI using a 1.5-T magnet significantly increased with increasing lesion size. Biopsy is rarely necessary for lesions smaller than 5 mm because of their low (3%) likelihood of cancer. Further work is needed to develop an algorithm that uses size in addition to other patient and lesion factors to guide biopsy recommendations for MRI-detected breast lesions.
Keywords: breast • breast cancer • breast imaging • MRI • oncologic imaging
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
During the study period, 3,206 breast MRI examinations were performed at our facility. Indications for performing breast MRI at our institution included high-risk screening in women with risk factors for developing breast cancer (e.g., genetic predisposition; strong family history of breast cancer; previous breast cancer; previous biopsy-proven diagnosis of atypical ductal hyperplasia, atypical lobular hyperplasia, or lobular carcinoma in situ; or previous mantle irradiation for Hodgkin's disease), extent-of-disease assessment (in a woman with breast cancer diagnosed within the previous 6 months to assess for additional sites of cancer in the ipsilateral or contralateral breast), problem solving (e.g., occult cancer in an axillary lymph node consistent with breast primary with negative mammogram and physical examination, nipple discharge, or questionable but indeterminate findings on conventional imaging studies such as a suspicious lesion identified only on one view), or follow-up (recommended on the basis of a previous breast MRI examination).
Breast MRI Technique
At our institution, diagnostic MRI examinations were performed with the
patient prone in a 1.5-T commercially available system (Signa, GE Healthcare)
using a dedicated surface breast coil. Our imaging sequence included a
localizing sequence followed by a sagittal fat-suppressed T2-weighted sequence
(TR/TE, 4,000/85). A T1-weighted 3D fat-suppressed fast spoiled gradient-echo
(17/2.4; flip angle, 35°; bandwidth, 31.25 MHz) sequence was then
performed before and three times after a rapid bolus injection of 0.1 mmol/L
of gadopentetate dimeglumine (Magnevist; Berlex) per kilogram of body weight,
delivered through an indwelling IV catheter.
Image acquisition started after IV contrast material injection and saline bolus. Sequential images were obtained sagittally, for an acquisition time that was usually under 2 min each per volumetric acquisition. Total imaging time per breast, including three postcontrast acquisitions, was usually less than 20 min. Section thickness was between 2 and 3 mm without gap, using a matrix of 256 x 192 and field of view of 16-22 cm. Frequency was in the anteroposterior direction. After the examination, the precontrast images were subtracted from the first postcontrast images on a pixel-by-pixel basis.
Interpretation of Breast MRI Examinations in Clinical Practice
The institution at which the study was performed is an academic center
where more than 30,000 mammograms, 1,500 breast MRI examinations, and more
than 1,000 new breast cancer cases were evaluated annually during the study
period. During the study period in our practice, each MRI examination was
interpreted by one of nine breast imaging specialists in conjunction with
clinical history and other breast imaging studies when available.
Level of suspicion was reported on a scale of 0 to 6 as follows: 0, needs additional imaging evaluation; 1, no abnormal enhancement; 2, benign enhancement; 3, probably benign, recommend short-term follow-up (specified as either at a different time in the patient's menstrual cycle or in 6 months); 4, suspicious; 5, highly suggestive of malignancy; or 6, known cancer. Images were interpreted on a PACS monitor, which allowed manual adjustment of window and level. Classification was based primarily on lesion morphology; however, kinetic features were visually assessed on the three postcontrast image acquisitions.
The maximal diameter of the lesion was measured by the interpreting radiologist using electronic calipers per clinical practice at our institution, and this measurement was recorded before biopsy. Most lesions referred for biopsy were larger than 5 mm; however, smaller lesions were referred for biopsy at the discretion of the attending radiologist depending on specific lesion features (e.g., irregular or spiculated margins or washout kinetics), solitary nature, appearance different from that of other multiple enhancing areas, and clinical history (e.g., a small but indeterminate lesion in the ipsilateral breast of a woman with known cancer contemplating breast-conserving surgery).
MRI-Guided Needle Localization Technique
For nonpalpable, mammographically occult, MRI-detected lesions referred for
biopsy that had no sonographic correlates, MRI-guided needle localization was
performed using previously described methods
[7], a biopsy compression
device (Biopsy System No. NMR NI 160, or Open Breast Coil [model OBC-63], or
Biopsy Breast Array Coil [model BBC], MRI Devices), and MRI-compatible needles
or hookwires (Tumor Localizer, 18- or 20-gauge, Daum Medical; MRI Breast
Lesion Marking System, 20-gauge, E-Z-EM; or MReye Modified Kopans Spring Hook
Localization Needle, 20-gauge, Cook).
Data Collection and Analysis
Medical records were reviewed to determine the indication for breast MRI,
menopausal status, and histologic findings at biopsy. Cancers were classified
as ductal carcinoma in situ (DCIS) or invasive carcinoma; for purposes of
analysis, breast lymphomas were included with invasive cancers. The frequency
of cancer as a function of lesion size was determined for all lesions and in
subgroups according to indication for MRI and menopausal status. Data were
entered into a computerized spreadsheet (Excel, Microsoft). Statistical
analysis was performed with a software program (Epi-Info, Centers for Disease
Control and Prevention) using the chisquare and Fisher's exact tests.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
For each individual lesion size from 4 to 20 mm, the frequency of malignancy was significantly higher in lesions equal to or larger than that size compared with smaller lesions (Table 3). The likelihood of malignancy significantly increased (p = 0.0005) with increasing lesion size (Table 4). Malignancy was found in one (3%) of 37 lesions measuring less than 5 mm, 44 (17%) of 254 lesions 5-9 mm, 37 (25%) of 151 lesions 10-14 mm, 21 (28%) of 74 lesions 15-19 mm, and 46 (31%) of 150 lesions 20 mm or larger (Table 4). Among lesions smaller than 10 mm, the likelihood of malignancy was highest for lesions detected in the extent of disease setting (22% malignant) and lowest in the high-risk screening setting (10% malignant) (Table 4).
|
|
Extent of Disease
The frequency of malignancy was significantly higher in MRI lesions
detected for extent of disease assessment compared with all other indications
(88/279 = 32% vs 61/387 = 16%; p = 0.000002; odds ratio, 2.5; 95%
confidence interval, 1.7-3.6) (Table
4). Malignancy was present in 67 (42%) of 158 lesions in the
ipsilateral breast, of which 37 (55%) were DCIS and 30 (45%) were invasive
cancer. Malignancy was present in 21 (17%) of 121 MRI-detected lesions in the
contralateral breast, of which 10 (48%) were DCIS and 11 (52%) were invasive
cancer. The frequency of malignancy was significantly (p < 0.0001)
higher in MRI-detected lesions in the ipsilateral compared with the
contralateral breast.
For MRI lesions detected in the assessment of extent of disease, the frequency of malignancy increased significantly (p = 0.005) with lesion size (Table 4). The frequency of malignancy was significantly higher for lesions measuring 5 mm or larger compared with smaller lesions (88/264 = 33% vs 0/15 = 0%; p < 0.004), lesions 10 mm or larger compared with smaller lesions (62/159 = 39% vs 26/120 = 22%; p = 0.003), and lesions 15 mm or larger compared with smaller lesions (37/90 = 41% vs 51/189 = 27%; p < 0.03).
High-Risk Screening
For MRI-detected lesions identified on high-risk screening examinations,
the frequency of malignancy was higher in larger lesions compared with smaller
lesions, but this difference only achieved statistical significance or a trend
toward significance at specific lesion sizes
(Table 4). Among MRI
screening-detected lesions, the frequency of malignancy was higher for lesions
5 mm or larger compared with smaller lesions (38/244 = 16% vs 1/16 = 6%;
p = 0.5), lesions 10 mm or larger compared with smaller lesions
(27/145 = 19% vs 12/115 = 10%; p = 0.097), lesions 12 mm or larger
compared with smaller lesions (24/123 = 20% vs 15/137 = 11%; p =
0.08), lesions 16 mm or larger compared with smaller lesions (18/76 = 24% vs
21/184 = 11%; p = 0.02), and lesions 20 mm or larger compared with
smaller lesions (15/59 = 25% vs 24/201 = 12%; p < 0.02).
Problem Solving
For MRI-detected lesions identified at problem solving, the frequency of
malignancy was higher in larger compared with smaller lesions, but this
difference only achieved statistical significance or a trend toward
significance at specific lesion sizes
(Table 4).
Follow-Up
Among lesions identified at follow-up MRI studies, no significant
difference in the frequency of malignancy was observed as a function of lesion
size, either overall or at specific lesion sizes
(Table 4).
Menopausal Status
Malignancy was identified in 76 (27%) of 281 MRI lesions detected in
postmenopausal women; of these malignancies, 35 (46%) were DCIS, 38 (50%) were
invasive cancer, and three (4%) were lymphoma
(Table 5). Malignancy was found
in 73 (19%) of 385 premenopausal women, of which 45 (62%) were DCIS and 28
(39%) were invasive cancer (Table
5).
|
The frequency of malignancy was significantly higher in lesions identified in postmenopausal women compared with premenopausal women (27% vs 19%; p < 0.02). Among lesions measuring less than 10 mm, the frequency of malignancy was significantly higher in postmenopausal women than premenopausal women (29/134 = 22% vs 16/157 = 10%; p = 0.01; odds ratio, 2.4; 95% confidence interval, 1.2-5.0). The frequency of malignancy increased significantly with lesion size, both in postmenopausal (p = 0.02) and premenopausal (p < 0.007) women (Table 5).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Few data address the use of size criteria to determine the need for biopsy of nonpalpable breast lesions identified on imaging studies. Sickles [8] found no statistically significant difference in the likelihood of cancer as a function of lesion size in "probably benign" solitary, nonpalpable, circumscribed masses on baseline mammograms. Stavros et al. [9] described the use of specific sonographic criteria involving lesion shape, margins, orientation, internal echogenicity, and posterior enhancement patterns in determining the likelihood of carcinoma in a sonographically evident mass, but they did not assess criteria based on sonographic lesion size. In early work involving breast MRI, Harms et al. [10] indicated that biopsy was recommended only for MRI-detected lesions 1 cm or larger, but no data validated this approach.
We studied the association between MRI lesion size and likelihood of cancer in 666 lesions that had MRI-guided needle localization and surgical biopsy. The overall frequency of malignancy in our study was 22%, at the low end of the 22-55% range reported in studies of MRI-guided needle localization at other centers in the past decade [11-19]. The 54% proportion of cancers that are DCIS in our study is higher than the 0-40% range reported by others [11-19]. These differences may reflect differences in patient populations, MRI technique, and interpretation. In our study, the frequency of cancer was highest in the extent of disease setting (32% cancer), particularly in the ipsilateral breast (42% cancer), and lowest in the high-risk screening setting (15% cancer). The frequency of cancer was also higher in lesions detected in postmenopausal women (27% cancer) than in premenopausal women (19% cancer).
The results of our study suggest that MRI lesion size may be useful in predicting the likelihood of malignancy in MRI-detected breast lesions. For all lesions, the frequency of malignancy significantly increased (p = 0.0005) with increasing lesion size. At a variety of cutoff values examined, the frequency of malignancy was significantly higher for larger lesions versus smaller lesions. For example, the frequency of malignancy was higher for lesions measuring 5 mm or larger versus smaller lesions (24% vs 3%; p < 0.006), 10 mm or larger versus smaller lesions (28% vs 16%; p = 0.0002), 15 mm or larger versus smaller lesions (30% vs 19%; p = 0.001), and 20 mm or larger versus smaller lesions (31% vs 20%; p < 0.008).
Our data enable calculation of the potential impact of size criteria on biopsy recommendations. If lesions measuring less than 5 mm had not been referred for biopsy, the PPV would have increased from 22% to 24% (p =0.7); 37 (6%) of 666 lesions would have been spared biopsy, and one DCIS lesion (< 1% of 149 malignancies) would have been missed. If lesions less than 6 mm had not been referred for biopsy, the PPV would have increased from 22% to 24% (p = 0.54); 74 (11%) of 666 lesions would have been spared biopsy at the expense of missing seven (5%) of 149 malignancies (five DCIS and two invasive cancers). If lesions measuring less than 10 mm had not been referred for biopsy, the PPV would have increased from 22% to 28% (p = 0.001); 291 (44%) of 666 lesions would have been spared biopsy at the expense of missing 45 (30%) of 149 malignancies (23 DCIS and 22 invasive cancers).
The likelihood of cancer for any lesion size depends in part on the prior probability of breast cancer; therefore, the optimal size criteria may differ in different clinical scenarios. For example, a smaller lesion has a significantly higher likelihood of cancer in the extent of disease setting (especially in the ipsilateral breast) and in postmenopausal women and a lower likelihood of malignancy in the screening setting and in premenopausal women. If size criteria are used, one could argue for a higher size cutoff in premenopausal women having high-risk screening examinations and a lower size cutoff in postmenopausal women having MRI for an extent-of-disease assessment.
Our study has limitations. Each lesion was measured once by a single interpreting radiologist; variability in measurement of breast MRI lesions may have occurred, as variability occurs in other aspects of interpretation of breast MRI [20], mammography [21, 22], and sonography [23]. Our protocol, however, allowed us to state that if any one of the nine interpreting radiologists judged the lesion to be less than 5 mm at maximal diameter, the likelihood of malignancy was extremely low (3%). Our data provide, to our knowledge, the first scientific validation of the low likelihood of cancer in MRI-detected lesions smaller than 5 mm, lesions described by the American College of Radiology Breast MRI Lexicon [2] as a "focus," defined as "a tiny spot of enhancement that does not clearly represent a space-occupying lesion or mass" [2]. Further study with multiple reviewers and assessment of inter- and intraobserver variability would provide additional information. Whether our results, obtained using a 1.5-T magnet, will be applicable at higher field strengths (such as 3 T) or with further improvements in MRI technique and resolution remains to be determined [24].
In conclusion, we found that the likelihood of malignancy in MRI-detected breast lesions at 1.5 T increased significantly with increasing lesion size. Biopsy is generally not warranted for MRI-detected lesions less than 5 mm due to the low (3%) frequency of malignancy in these lesions. For lesions less than 1 cm, the frequency of malignancy was highest in postmenopausal women and in the extent of disease setting (particularly in the ipsilateral breast) and lowest in premenopausal women and in the high-risk screening setting. The final decision regarding biopsy recommendation should be based not only on lesion size but on other lesion features (e.g., morphology, kinetics, and enhancement pattern) and patient risk factors and clinical history. Further work is necessary to develop a multivariate model including all of these parameters and to reevaluate this model as MRI technique evolves to minimize the number of benign biopsies while maximizing the benefits of MRI in detecting early breast cancer.
Acknowledgments
We acknowledge Cynthia M. Thornton, Indira Gonzalez, Youngduk Paik, and
Anita M. Sanchez for their technical support with MRI-guided needle
localizations, and David C. Perlman for assistance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Preda, S. Rizzo, M. Bellomi, W. Tse Yang, and H. T. Le-Petross MR Imaging of Breast Lymphoma Radiology, July 1, 2008; 248(1): 320 - 321. [Full Text] [PDF]
|
|
|
|
 |
|
 N. Hoogerbrugge, Y. J. L. Kamm, P. Bult, K. M. Landsbergen, E. M. H. F. Bongers, H. G. Brunner, H. J. Bonenkamp, J. A. de Hullu, M. J. L. Ligtenberg, and C. Boetes The impact of a false-positive MRI on the choice for mastectomy in BRCA mutation carriers is limited Ann. Onc., April 1, 2008; 19(4): 655 - 659. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. H. Javid, J. W. Carlson, J. E. Garber, R. L. Birdwell, S. Lester, S. Lipsitz, and M. Golshan Breast MRI Wire-Guided Excisional Biopsy: Specimen Size as Compared to Mammogram Wire-Guided Excisional Biopsy and Implications for Use. Ann. Surg. Oncol., December 1, 2007; 14(12): 3352 - 3358. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. B. Bevilacqua, M. W. Kattan, J. V. Fey, H. S. Cody III, P. I. Borgen, and K. J. Van Zee Doctor, What Are My Chances of Having a Positive Sentinel Node? A Validated Nomogram for Risk Estimation J. Clin. Oncol., August 20, 2007; 25(24): 3670 - 3679. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. A. J. Van den Bosch, D. M. Ikeda, and B. L. Daniel Does Size Matter? Likelihood of Cancer in MRI-Detected Lesions Less Than 5 mm Am. J. Roentgenol., June 1, 2007; 188(6): W571 - W571. [Full Text] [PDF]
|
|
|
|
|
|
F. Pediconi, C. Catalano, A. Roselli, S. Padula, F. Altomari, E. Moriconi, A. M. Pronio, M. A. Kirchin, and R. Passariello Contrast-enhanced MR Mammography for Evaluation of the Contralateral Breast in Patients with Diagnosed Unilateral Breast Cancer or High-Risk Lesions Radiology, June 1, 2007; 243(3): 670 - 680. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Plevritis, D. M. Ikeda, and A. M. Garber Cost-effectiveness of Breast Magnetic Resonance Imaging to Screen BRCA1/2 Mutation Carriers--Reply JAMA, December 13, 2006; 296(22): 2682 - 2682. [Full Text] [PDF]
|
|
|
|
 |
|
 T. G. ODLE Breast MR Radiol. Technol., September 1, 2006; 78(1): 45M - 66M. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
