| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorial |
1 From the Department of Radiology, Hospital of the University of Pennsylvania, 3400 Spruce St, Philadelphia, PA 19104. Received November 4, 1998; accepted November 12. Address reprint requests to the author.
Index terms: Breast neoplasms, 00.31, 00.32 • Breast neoplasms, MR, 00.121412, 00.121415, 00.12143, 00.12144 • Editorials • Magnetic resonance (MR), flow studies, 00.12143, 00.12144
Magnetic resonance (MR) imaging is now emerging as a very exciting and potentially powerful tool for the imaging of breast abnormalities. While the indications for MR imaging of the breast have yet to be defined, one of the proposed indications is lesion differential diagnosis: Are there lesion characteristics that can be identified at MR imaging of the breast that would allow for the accurate differentiation of benign from malignant?
In this issue of Radiology, Kuhl et al (1) propose that signal intensity time course data are useful for differentiating benign from malignant enhancing lesions identified on dynamic contrast material–enhanced breast MR imaging studies. The group reports results of prospective breast MR imaging in 230 patients with 266 contrast-enhancing lesions. The only inclusion criterion was the presence of a suspicious enhancing lesion in the breast. Region of interest (ROI)–based time–signal intensity curves of the lesions were obtained, and the curves were classified according to their shapes as type I, steady enhancement with a straight or curved time–signal intensity line; type II, plateau of signal intensity; or type III, washout of signal intensity. Enhancement rates and curve types were compared for benign and malignant lesions.
In the authors' patient population, 83% of the benign lesions exhibited a steady or curved time–signal intensity curve. In contrast, 57% of malignant lesions exhibited a washout time–signal intensity curve. Using the shape of the time–signal intensity curve alone, the authors report a sensitivity of 91% (92 of 101), a specificity of 83% (137 of 165), a positive predictive value of 77% (92 of 120), a negative predictive value of 94% (137 of 146), and a diagnostic accuracy of 86% (229 of 266). The likelihood of breast cancer associated with a type I, II, or III time course was 6% (nine of 146), 64% (34 of 53), and 87% (58 of 67), respectively.
In contrast, the authors also used enhancement rate alone, with relative enhancement less than or equal to 60% used to classify a lesion as benign, more than 60% and less than or equal to 80% used to classify a lesion as indeterminate, and more than 80% used to classify a lesion as malignant. They found that while the enhancement rate of benign and malignant lesions differed significantly (P < .001), due to the large SD, there was a considerable overlap in the range of enhancement rates of benign and malignant lesions, with a reported sensitivity of 91% (92 of 101), a specificity of 37% (61 of 165), a positive predictive value of 47% (92 of 196), a negative predictive value of 87% (61 of 70), and a diagnostic accuracy of 58% (153 of 266).
The authors conclude that the shape of the time–signal intensity curve is an important criterion in differentiating benign from malignant lesions even when the rates of enhancement are comparable, and, specifically, that a type III curve is a strong indicator of malignancy, independent of the enhancement rate.
Background: Why MR Imaging for Lesion Diagnosis?
The majority of mammographically and clinically detected suspicious abnormalities will be benign at biopsy (2). MR imaging, with its rich soft-tissue contrast and thin-section, multiplanar capability, offers the possibility of better lesion characterization than can be obtained with conventional imaging methods. Unfortunately, while the reported sensitivities of MR imaging for breast cancer have been as high as 94%–100%, reported specificities have been much more variable, with a range of 37%–97% (3–9). This wide range of values reflects multiple factors, including differences in magnetic field strength, imaging parameters, patient selection criteria, image interpretation criteria, and, perhaps most important, histologic variability of benign and malignant lesions.
Early investigation of contrast-enhanced MR imaging of the breast demonstrated that breast cancer consistently enhanced after the intravenous administration of contrast material (3,4). However, additional investigation showed that not only did malignant lesions enhance, but also many benign lesions, including fibroadenomas, fibrocystic changes, radial scars, mastitis, atypical ductal hyperplasia, and lobular neoplasia, enhanced (4,5,7,8,10). In addition, presumably normal breast tissue demonstrated contrast enhancement, and this enhancement was variable during different phases of the menstrual cycle (11). Thus, the presence of enhancement alone cannot be used to differentiate benign from malignant lesions; further characterization is necessary.
Image Interpretation
At present, there are no standard interpretation criteria for evaluating breast MR imaging studies. There are basically two questions that need to be answered when reviewing a breast MR study: (a) Is there a clinically important enhancing lesion in the breast? and (b) Is this lesion likely to be benign or malignant? As to the first question, while it would seem a relatively simple task to determine whether there is an enhancing lesion within the breast, there is no universally accepted definition of what constitutes clinically important enhancement.
There are, perhaps, as many criteria for classifying an enhancing lesion as clinically important as there have been published series. Kaiser and Zeitler (3) proposed that clinically important enhancement was characterized by an increase in signal intensity of 100% within the first 2 minutes. Heywang et al (4) used ultimate signal intensity rather than dynamic information, with enhancement above 300 normalized units considered clinically important. Gilles et al (6) classified any enhancement in breast parenchyma concomitant with early normal vascular enhancement as a positive finding. Boetes et al (9) classified a lesion as suspicious if it enhanced within 11.5 seconds after the aorta opacified. Gribbestad et al (12) used the criterion of a 70% signal intensity increase in 1 minute.
In their study, Kuhl et al (1) included patients only after a lesion classified as suspicious was identified. The authors defined suspicious enhancement as early-phase enhancement (an increase in signal intensity of more than 60%) that was apparent on the first enhanced image after the administration of contrast material (40 seconds) or contrast enhancement with suspicious morphology (ill-defined borders or irregular contour).
As the authors note, including patients only after a suspicious lesion was identified is a possible source of bias in their study. Using such an inclusion criterion probably explains the relatively low percentage of fibrocystic changes in the study group. Such inclusion criteria may also explain, in part, the relatively low percentage of ductal carcinoma in situ lesions in the study group. As previously reported (13,14), ductal carcinoma in situ may not demonstrate rapid and intense contrast enhancement. While individual investigators have set criteria for what constitutes a significantly enhancing lesion, it is critical to remember that these criteria are arbitrary; the reported results are not necessarily transferable to all MR imaging techniques, and no one interpretation criterion is uniformly accepted at the present time.
Once an enhancing lesion is identified, it must then be determined whether the lesion is suspicious for malignancy or is benign. Two major approaches have been used in an attempt to differentiate suspicious or malignant lesions from benign lesions: enhancement kinetics and architectural feature analysis. In terms of enhancement kinetics, as noted earlier, multiple approaches have been described, including both quantitative and qualitative approaches. The quantitative approaches have included various mathematic models of the enhancement profile data, measurement of the maximum rate of enhancement (slope of enhancement), the maximum amplitude of enhancement, and the rate of contrast material washout (3,4,6–10,15). The imaging parameters in these studies have varied widely, with a time resolution as fast as 2.3 seconds for single-section studies to 1–2 minutes with three-dimensional volume acquisitions.
Using such quantitative methods, several investigators (3,9) have reported that lesion enhancement rates in the early postcontrast period can serve as a method for differentiating malignant from benign lesions, where malignant lesions enhance more quickly and strongly than benign lesions. However, other investigators (4,6–8,10) have demonstrated that while cancers tend to enhance faster than benign lesions, there is a clear overlap in the enhancement rates of benign and malignant lesions.
The variability of the quantitative methods, the dependence of these methods on the acquisition parameters, and the overlap in the enhancement kinetics in the early postcontrast enhancement period of benign and malignant lesions have led investigators to seek a qualitative approach to lesion enhancement. Kuhl et al present such an approach, in which the shape of the entire time–signal intensity curve is qualitatively assessed, yielding three different curves: steady, plateau, and washout. In their patient population, the use of these time–signal intensity curves resulted in dramatically higher specificity (83%) and accuracy (86%) than were obtained when the enhancement rate (specificity, 37%; accuracy, 58%) was used.
The qualitative approach as presented by Kuhl et al permits simple visual inspection of the enhancement curve as an alternative to complex mathematic models or the need to quantify the amount or rate of enhancement. However, as the authors describe, there are caveats that must be considered. First, critical to obtaining the time–signal intensity curve is the accurate placement of an ROI over the area of most rapid and intense enhancement. Interobserver variability and bias in placing the ROI have been previously reported (16), although this variability can be improved when a semiautomated method using parametric images is used. In the study by Kuhl et al, a resident, not involved in reading the film hard copy, was responsible for placement of the ROI by using parametric images. The authors did not test interobserver variability in the placement of the ROI.
Second, a related caveat concerns the distribution of results of histologic examination of the lesions, in which the percentage of invasive cancers and solid benign lesions is higher than expected, while the percentage of fibrocystic changes and ductal carcinomas in situ is lower than expected. This most likely reflects both the initial inclusion criteria of patients with a suspicious enhancing lesion and the referral pattern to MR imaging in the Department of Radiology, University of Bonn, where patients in whom ductal carcinoma in situ is suspected do not undergo MR imaging. The placement of an ROI and the generation of a time–signal intensity curve for both ductal enhancement (ie, as seen with ductal carcinoma in situ) and regional enhancement (ie, which may be seen with fibrocystic changes and ductal carcinoma in situ) may not be as accurate as they are for solid masses. As the authors note, the time–signal intensity curve analysis appears to be most useful in the differential diagnosis of focal lesions that demonstrate rapid enhancement.
The findings presented by Kuhl et al are promising. However, caution must be exercised in attempting to extend these findings to other clinical practice settings. The results presented in their article were obtained from a select patient population and may not necessarily be reproduced in other patient populations.
Enhancement Kinetics or Lesion Morphology: Which Is Best?
While the focus of the article by Kuhl et al is the evaluation of enhancement kinetics as a method for differentiating benign from malignant lesions, the authors make the very important point that analysis of lesion enhancement kinetics should not be used as a stand-alone criterion for lesion diagnosis, but rather should be evaluated along with lesion morphology. Several investigators (5,7,10,13,14,17) have reported architectural features identified on high-spatial-resolution, contrast-enhanced MR imaging studies that can be used for lesion diagnosis. Architectural features that suggest the possibility of malignancy include a mass with irregular or spiculated borders, a mass with peripheral enhancement, and ductal enhancement. Architectural features suggesting benign disease include a mass with smooth or lobulated borders, a mass demonstrating no contrast enhancement, a mass with nonenhancing internal septations, and patchy parenchymal enhancement (10,17).
It is becoming increasingly clear that while most investigators have used either enhancement kinetics or lesion morphology in an attempt to differentiate benign from malignant lesions identified on contrast-enhanced MR imaging studies, the integration of both kinetic and morphologic information may ultimately be needed to achieve optimal discrimination. Kuhl et al describe such an integration of kinetics and architecture as it is used in their practice. The authors make the very important point that there must be concordance between the kinetic information and the morphologic features. There may be malignant lesions, such as certain invasive ductal and lobular carcinomas and certain ductal carcinoma in situ lesions that will not enhance rapidly but in which lesion morphology (ie, architectural distortion, mass with spiculated borders, or ductal enhancement) suggests the presence of malignancy.
In their breast MR imaging practice, Kuhl et al analyze the time course kinetics only after evaluating lesion morphology on postcontrast images. In those cases in which the morphology suggests malignancy, the kinetic information is not evaluated. When the morphology is indeterminate or suggests a benign lesion, the authors recommend performing a time–signal intensity curve analysis. When the morphology suggests a benign or indeterminate lesion but a washout curve is detected, the authors recommend performing biopsy. In contrast, when a type I curve is identified, this can be used to support the diagnosis of a benign or probably benign lesion, averting performance of biopsy.
The relative importance assigned to enhancement kinetics or lesion morphology will, for now, vary from practice to practice because of differences in experience, available hardware and software, and imaging parameters. Whether high temporal resolution (needed for time–signal intensity curve analysis) should be sacrificed for high spatial resolution (needed for morphologic analysis) or vice versa or whether both high temporal and high spatial resolution will be obtainable with continuing improvements in MR imaging technology remains unclear. However, investigators in ongoing and future clinical trials will be looking for answers to these questions.
The techniques of performing and interpreting breast MR imaging studies continue to evolve. At present, the indications for MR imaging of the breast, the optimal imaging parameters, and interpretation criteria for diagnosis remain undefined. The results of clinical investigation, thus far, suggest that in the near future, MR imaging will play a very important role in the evaluation of breast abnormalities. One of the major limitations of this technology, however, remains its relatively low specificity. Most enhancing lesions identified within the breast will not be malignant, so methods to differentiate these false-positive enhancing lesions from the true-positive malignancies are needed. The time–signal intensity curve analysis presented by Kuhl et al provides one such method.
Footnotes
Abbreviation: ROI = region of interest
See also the article by Kuhl (pp 101–110 ) in this issue.
References
Related Article
This article has been cited by other articles:
|
|
 |
|
  F. Pediconi, C. Catalano, S. Padula, A. Roselli, V. Dominelli, S. Cagioli, M. A. Kirchin, G. Pirovano, and R. Passariello Contrast-Enhanced MR Mammography: Improved Lesion Detection and Differentiation with Gadobenate Dimeglumine Am. J. Roentgenol., November 1, 2008; 191(5): 1339 - 1346. [Abstract] [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]
|
|
|
|
|
|
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]
|
|
|
|
|
|
E. Yeh, P. Slanetz, D. B. Kopans, E. Rafferty, D. Georgian-Smith, L. Moy, E. Halpern, R. Moore, I. Kuter, and A. Taghian Prospective Comparison of Mammography, Sonography, and MRI in Patients Undergoing Neoadjuvant Chemotherapy for Palpable Breast Cancer Am. J. Roentgenol., March 1, 2005; 184(3): 868 - 877. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W. Lee, W. K. Moon, H.-J. Weinmann, S. J. Kim, J. H. Kim, S. H. Park, T. J. Kim, C. J. Yoon, Y. H. Kim, E. Y. Cho, et al. Contrast-enhanced MR Imaging of Postoperative Scars and VX2 Carcinoma in Rabbits: Comparison of Macromolecular Contrast Agent and Gadopentetate Dimeglumine Radiology, October 1, 2003; 229(1): 132 - 139. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Jacobs, P. B. Barker, D. A. Bluemke, C. Maranto, C. Arnold, E. H. Herskovits, and Z. Bhujwalla Benign and Malignant Breast Lesions: Diagnosis with Multiparametric MR Imaging Radiology, October 1, 2003; 229(1): 225 - 232. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Liberman, E. A. Morris, D. D. Dershaw, A. F. Abramson, and L. K. Tan Ductal Enhancement on MR Imaging of the Breast Am. J. Roentgenol., August 1, 2003; 181(2): 519 - 525. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H Neubauer, M Li, R Kuehne-Heid, A Schneider, and W A Kaiser High grade and non-high grade ductal carcinoma in situ on dynamic MR mammography: characteristic findings for signal increase and morphological pattern of enhancement Br. J. Radiol., January 1, 2003; 76(901): 3 - 12. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Liberman, E. A. Morris, M. J.-Y. Lee, J. B. Kaplan, L. R. LaTrenta, J. H. Menell, A. F. Abramson, S. M. Dashnaw, D. J. Ballon, and D. D. Dershaw Breast Lesions Detected on MR Imaging: Features and Positive Predictive Value Am. J. Roentgenol., July 1, 2002; 179(1): 171 - 178. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Orel and M. D. Schnall MR Imaging of the Breast for the Detection, Diagnosis, and Staging of Breast Cancer Radiology, July 1, 2001; 220(1): 13 - 30. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R Warren and P Kessar A method of coil localization for breast lesions seen only on MRI Br. J. Radiol., June 1, 2001; 74(882): 548 - 551. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. T. Agoston, B. L. Daniel, R. J. Herfkens, D. M. Ikeda, R. L. Birdwell, S. G. Heiss, and A. M. Sawyer-Glover Intensity-modulated Parametric Mapping for Simultaneous Display of Rapid Dynamic and High-Spatial-Resolution Breast MR Imaging Data RadioGraphics, January 1, 2001; 21(1): 217 - 226. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Kinkel, T. H. Helbich, L. J. Esserman, J. Barclay, E. H. Schwerin, E. A. Sickles, and N. M. Hylton Dynamic High-Spatial-Resolution MR Imaging of Suspicious Breast Lesions: Diagnostic Criteria and Interobserver Variability Am. J. Roentgenol., July 1, 2000; 175(1): 35 - 43. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
