| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Diagnostic Radiology, Chinese University of Hong Kong, Prince of
Wales Hospital, Shatin, NT, Hong Kong SAR, China.
2 Present address: Department of Diagnostic Radiology, University of Texas M. D.
Anderson Cancer Center, Unit 57, 1515 Holcombe Blvd., Houston, TX 77030.
3 Department of Anatomical and Cellular Pathology, Prince of Wales Hospital,
Shatin, NT, Hong Kong SAR, China.
Received February 14, 2003;
accepted after revision July 16, 2003.
Address correspondence to W. T. Yang
(wyang{at}di.mdacc.tmc.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. We retrospectively reviewed mammographic and sonographic images of 60 DCIS lesions from 55 symptomatic women. Images were reviewed by a radiologist who knew that the patients had DCIS but had no other information regarding pathology. Lesions were evaluated pathologically and classified using the Van Nuys classification system. Statistical comparisons were made using Fisher's exact test.
RESULTS. Of the 60 lesions, 33 were classified as Van Nuys group 1, 19 as Van Nuys group 2, and eight as Van Nuys group 3. Six (10%) of the 60 lesions were not visible on sonography, and 12 lesions (20%) were not visible on mammography. Sonography revealed a mass in 43 cases (72%), ductal changes in 14 cases (23%), and architectural distortion in four cases (7%). Eight lesions had more than one of these features. A sonographically visualized, irregularly shaped mass with indistinct or angular margins and no posterior acoustic shadowing or enhancement was associated with a high Van Nuys classification (p < 0.05). Microcalcifications were visible on sonography in 13 (22%) of the 60 lesions or on mammography in 25 lesions (42%). Both findings were associated with a high Van Nuys classification (p < 0.05).
CONCLUSION. Although sonography can reveal microcalcifications within masses, it is unreliable in depicting and characterizing the morphology and extent of microcalcifications, particularly when they are in isolation. Therefore, sonography should not be used to replace mammography but instead as an adjunctive tool to increase the sensitivity of mammography in breast diagnosis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The mammographic features of DCIS have been well described in the literature, with microcalcifications being the dominant feature [10–12]. Other findings such as masses, nodular abnormalities, architectural distortions, dilated retroareolar ducts, and developing densities have also been reported [13, 14]. Little systematic description of the gray-scale and color power Doppler sonographic features of DCIS is found in the literature [15–17], although several recent studies have examined the role of sonography in the evaluation of mammographically detected microcalcifications [18–20].
The purpose of this study was to describe the spectrum of sonographic features of symptomatic DCIS and to evaluate the ability of sonography to predict the grade of DCIS on the basis of imaging features. Documenting these sonographic features may help us to increase the sensitivity and specificity of sonography as an adjunctive tool in breast diagnosis, and comparing these features with the histopathologic results will help us to further evaluate the ability of this technique to determine prognosis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Clinical Features
The following clinical features were obtained from the medical histories:
presence of a palpable mass, nipple discharge, Paget's disease, and mastalgia.
Thirty-two women (58%) had a palpable mass, 21 (35%) had spontaneous
uniorificial nipple discharge, two (3%) had Paget's disease, and two (3%) had
palpable thickening. Five women had more than one clinical finding. Three
women who presented with unilateral symptoms had incidental abnormalities
detected in the contralateral asymptomatic breast.
Sonography
Physicians have performed bilateral whole-breast real-time sonography on
all symptomatic patients at this institution. In 1998, color power Doppler
sonography was included in the scanning protocol for all solid and complex
cystic masses. Findings of color power Doppler sonography, including the
presence or absence of intralesional vascularity, were routinely documented on
hard-copy images and in reports of physicians. Lesions that did not fulfill
the criteria of benignity as defined by Stavros et al.
[22] were evaluated using a
sonographically guided biopsy after informed consent was obtained from the
patients. Sonography was performed on 60 DCIS lesions using a high-resolution
unit with a 7.5- or 12.5-MHz linear array transducer with a 6-MHz Doppler
operating frequency on a 700 Loqic Expert/Pro series (General Electric Medical
Systems, Milwaukee, WI) or a Sonoline Elegra scanner (Siemens Medical
Solutions, Erlangen, Germany). Color power Doppler sonography was performed on
32 lesions with the following optimized parameters: pulse repetition frequency
of 750–1,000 Hz, low wall pass filter, medium persistence, high
sensitivity, and dynamic motion differentiation. Color power Doppler gain was
optimized with an increase in gain until the color box was filled with uniform
low-level blue noise with minimal yellow power signal (
75–85% gain)
[23]. Sonograms were reviewed
for masses, architectural distortion, ductal extension and dilatation,
microcalcifications, and color-flow signal. We documented the following
features of masses: size, nature (solid or cystic), shape, margin,
echogenicity, and posterior acoustic phenomena.
Mammography
Standard two-view mammography was performed with a Senographe DMR+ unit
(General Electric Medical Systems) with additional views obtained as
necessary. Mammograms were reviewed for masses, calcifications, architectural
distortion, tubular ductal opacity, and asymmetric density. The shape, margin,
and density of masses were noted. The morphology and distribution of
microcalcifications were also recorded.
Galactography
Galactography was performed successfully in eight (six with bloody and two
with clear fluid) of the 21 patients who presented with nipple discharge and
unsuccessfully in four (all with bloody fluid) because of difficult
cannulation. The procedure was not attempted in nine (four with bloody and
five with clear fluid) patients in whom no discharge was expressible at the
time of examination. This relatively high success rate was likely related to
the preselected study population of women with pathologic diagnoses of DCIS,
resulting in a high pretest probability of cancer.
Histopathology
Histopathologic findings in excisional biopsy or mastectomy specimens were
used as the gold standard. We analyzed the following histologic parameters:
nuclear grade, presence and extent (by percentage area) of comedo necrosis
[8], architectural pattern
[4], and presence and extent of
microinvasion. Lesions were classified using the Van Nuys classification
system [7,
8]. Lesions with pure DCIS and
DCIS with microinvasion (invasive focus of 
1 mm) as defined by previously
published criteria [9] were
included in this study. Available long-term follow-up data on patients with
microinvasive carcinoma suggest that the prognosis after surgery is excellent,
with no difference in local recurrence, disease-free survival, and overall
survival rates when compared with those of patients with pure DCIS
[24]. In general, the
therapeutic algorithm for patients with pure DCIS and DCIS with microinvasion
is similar [25]. Cases of DCIS
associated with minimal invasion and infiltrative ductal cancer were excluded
from the study.
For all cases, sonographic and mammographic features were compared with the histopathologic findings. Statistical comparisons were performed using the Fisher's exact test for nonparametric inference with the statistical software package StatXact (Cytel Software, Cambridge, MA).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
Sonography
The correlation between sonographic and histopathologic findings is
provided in Table 3. Masses
were the most frequent sonographic finding in 43 (72%) of the 60 cases (Fig.
1A,
1B), followed by ductal
dilatation and low-level intraductal echoes (n = 14, 23%) (Fig.
2A,
2B,
2C) and architectural
distortion (n = 4, 7%) (Fig.
3A,
3B,
3C). Eight lesions had more
than one of these features. Microcalcifications were visible using sonography
in 13 (22%) of the 60 cases; they were characteristically noted within a mass
or duct (Fig. 4A,
4B) but were not visible in
isolation. Color power Doppler sonography revealed a positive signal in 22
(69%) of the 32 patients in whom it was performed (Figs.
2A,
2B,
2C and
5A,
5B,
5C). Masses visible on
sonography were typically solid, hypoechoic, irregularly shaped with
indistinct margins and showed the presence of microcalcifications. Typically,
no posterior acoustic phenomenon was present
(Table 3). Positive ductal
findings included a solitary nodule within a fluid-filled duct, a
solid-appearing distended duct (Fig.
2A,
2B,
2C), and an isolated elongated
nodule oriented in the direction of the subareolar duct. No abnormal skin,
nipple retraction, or axillary lymphadenopathy was noted.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Mammography and Galactography
The correlation between mammographic and histopathologic findings is
provided in Table 4.
Microcalcifications were the dominant finding, noted in 25 (42%) of the 60
cases (Figs. 3A,
3B,
3C and
4A,
4B), followed by the presence
of a mass in 24 cases (40%) (Figs.
5A,
5B,
5C and
6A,
6B). Architectural distortion
or asymmetric density was noted in eight lesions (13%) (Figs.
1A,
1B and
3A,
3B,
3C). Microcalcifications were
predominantly pleomorphic and linear with a segmental, clustered, or linear
distribution. Masses were either irregular or round or had high or medium
density and had indistinct margins. The findings of all eight galactograms
available for review were positive and showed intraductal filling defects
(Fig. 2A,
2B,
2C).
|
|
|
Histopathology Correlation
The visibility of microcalcifications on sonography and mammography was
significantly associated with the Van Nuys groups (p < 0.05).
Sonographic analysis of visible masses showed shape, margin, and posterior
acoustic phenomena to be also significantly associated with the Van Nuys
groups (p < 0.05). An irregularly shaped mass with indistinct or
angular margins and no posterior acoustic phenomena was more likely to be
associated with Van Nuys group 3 (p < 0.05) (Fig.
4A,
4B). In contrast, a cystic,
ovoid mass with circumscribed margins and posterior enhancement was more
likely to be associated with Van Nuys group 1 (p < 0.05) (Fig.
6A,
6B).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Histopathology Categorization
The grouping of DCIS into subtypes is based on a variety of microscopic
observations, including the architectural pattern of cell proliferation within
the involved duct, the extent of tumor necrosis, and the degree of cytonuclear
differentiation. The best mammographic–histologic correlation seems to
be between the pattern of microcalcifications and the cytonuclear
differentiation [28]. Most
classification schemes define three histologic grades: high, intermediate, and
low. High-grade DCIS exhibits large, variably sized nuclei with prominent
nucleoli and clumped chromatin. Extensive (comedo-type) tumor necrosis is also
usually present. Low-grade DCIS shows small uniform nuclei with inconspicuous
nucleoli and a diffuse homogeneous chromatin pattern. Intermediate-grade DCIS
is largely a "wastebasket" category with tumors displaying
intermediate nuclear features and variable necrosis. We chose to compare
imaging findings using the Van Nuys classification system because of the more
reliable prognosis it provides.
Sonography
The most frequent sonographic feature of DCIS in this study was a focal
mass, followed by low-level intraductal echoes, ductal extension or
dilatation, and architectural distortion. The finding of microcalcifications
on sonography and mammography was associated with a high Van Nuys grouping.
Masses visible on sonography that were irregular in shape and had an
indistinct or angular margin without associated posterior enhancement were
more likely to be associated with poorer prognosis. Although color power
Doppler sonography revealed the presence of color flow in 22 (69%) of the 32
lesions in which it was performed, this finding was not discriminating in
terms of the Van Nuys classification.
Other authors have reported the sonographic features of DCIS as architectural distortion; an intracystic lesion; or a bulky, hypoechoic vascular mass with ductal extension and prominent microlobules [29, 30]. More recently, Moon et al. [15] analyzed the sonographic findings in 70 patients with DCIS and correlated them with mammographic and histologic findings. The most common sonographic findings in DCIS in this series were a microlobulated mass, mild hypoechogenicity, ductal extension, and no acoustic enhancement or shadowing. Although the presence of spiculated margins, marked hypoechogenicity, a thick echogenic rim, and posterior acoustic shadowing suggested the presence of invasion, it remained difficult to differentiate pure DCIS from DCIS with microinvasion or invasive carcinoma. Although pure DCIS and DCIS with microinvasion based on the analysis of the results of surgical pathology may not differ significantly in terms of management or prognosis, this distinction may be important in core biopsy specimen analysis. Axillary nodal dissection is sometimes performed in patients diagnosed with DCIS with microinvasion found at core biopsy because of possible underestimation of disease extent. This topic however remains controversial.
The findings in our study concur with those of Moon et al. [15] and confirm that when evident on sonography, DCIS appears most frequently as a solid, irregular mass with indistinct margins and without posterior acoustic enhancement or shadowing. Most cases that showed posterior acoustic enhancement in this study were cystic masses. Virtually all solid masses in this study did not show posterior acoustic enhancement or shadowing. A possible explanation may be that DCIS or intraductal cancer lacks the characteristics of invasive cancer that incite desmoplastic reaction (fibrosis, cicatrization, scarring) leading to posterior acoustic shadowing and also lacks the aggressive, rapid cellular turnover that accompanies high-grade invasive cancers leading to acoustic through-transmission. The reason that masses were so commonly found in this study may be in part because our study population consisted of symptomatic women, many of whom presented with a palpable mass. Ductal change was the next most frequent sonographic feature observed in our study. Solitary nodules within fluid-filled ducts, solid-appearing distended ducts, and isolated elongated nodules oriented in the direction of the subareolar ducts were the ductal changes most frequently encountered in this study. Similar features may be found in patients with benign fibrocystic changes, but these fibrocystic changes are usually present in both breasts and are not usually found in DCIS. Ductal extension and branching patterns are further important sonographic features that aid in differentiating benign from malignant processes. Another recent review of the sonographic features of nine cases of DCIS showed that the number of malignant features and the mean size of the lesions increased with the grade of DCIS [17]. We also found such a trend.
Microcalcification, which was noted predominantly within masses and ducts in this study, is another sonographic feature that deserves mention. The ability to visualize microcalcifications using state-of-the-art sonographic equipment has been described elsewhere in the literature [22, 31, 32]. Several recent reports have further described the ability of sonography to depict mammographically detected microcalcifications [15, 16, 18]. These authors have also reported a higher sonographic visibility of malignant versus benign microcalcifications because most malignant calcifications occur in the mass [15, 16, 18]. Moon et al. [20] reported that with sonography, 77% of DCIS cases were visible as a breast mass associated with microcalcifications. They reported that calcifications associated with malignant tumors were more likely to be seen on sonography. However, the sonographic findings in DCIS with calcifications were nonspecific, in that they were also seen in benign lesions such as sclerosing adenosis, atypical ductal hyperplasia, and radial scars.
Other researches have stressed the importance of performing sonography with good-quality equipment, high-frequency transducers in the 10-13–MHz range and an optimal technique, and preferably by a physician [15, 20, 22]. They suggest that sonography, when optimally performed, can complement mammography in detecting and evaluating DCIS, particularly in detecting DCIS without calcifications and in evaluating disease extent in women with dense breasts. The sonographic findings of DCIS without calcifications, which may consist of single or multiple cystic or solid hypoechoic masses with no pseudocapsule, are nonspecific. These features are also seen in benign conditions such as papillomas, duct ectasia, fibrocystic change, and atypical ductal hyperplasia. This study, however, only addresses the sensitivity of sonography in depicting DCIS, without evaluating specificity, because the study cohort consists of DCIS cases only.
Mammography
Mammographic abnormalities were noted in 80% of the cases in this study.
These comprised microcalcifications (42%), dominant masses (40%), and
architectural distortion or asymmetric density (13%). These proportions differ
from previous descriptions of predominantly screening-detected DCIS in which
mammographic abnormalities were present in 95% of cases and included
microcalcifications (76%); asymmetric density (10%); dominant masses (8%);
and, in patients with nipple discharge, abnormal findings on galactograms (6%)
[10,
11]. These findings suggest
that there are differences in the mammographic features of screening-detected
versus symptomatic DCIS. The lower percentage of mammographically visible
microcalcifications and the higher percentage of dominant masses in this study
correspond to the lower percentage of comedo or high-grade (Van Nuys group 3)
lesions in this symptomatic cohort of DCIS. Mammography of low-grade DCIS
without comedo necrosis (Van Nuys group 1) has been reported to be less likely
to show microcalcifications and more likely to either be mammographically
normal or show noncalcified mammographic abnormalities
[33]. Bellamy et al.
[4] also noted an increased
proportion of the comedo carcinoma subtype in mammographically detected
lesions compared with symptomatic lesions. Researchers have hypothesized that
the larger percentage of comedo carcinoma found in the screening group may
reflect the greater invasive potential of these lesions, with relatively few
patients experiencing symptoms before the progression to invasive disease
[3,
11]. Screening-detected DCIS
is reportedly less extensive, has a smaller average calcification cluster
size, is less diffuse, and has less frequent retroareolar involvement
[3]. About 10% of all DCIS
lesions manifest as soft-tissue masses or asymmetric densities on mammography
[13], and up to 16% of DCIS
cases are mammographically occult.
When the ability of sonography and mammography to visualize DCIS was compared in this study, sonography proved superior in visualizing more cases. These findings reflect, in part, the distribution in a symptomatic population and also serve to reemphasize the increasing role of high-resolution sonography as an adjunctive tool in the evaluation of symptomatic breast disease, including symptomatic DCIS.
The primary limitation of this study is that only one radiologist and one pathologist reviewed the findings of radiology and pathology, respectively. Therefore, intra- and interobserver variability was not evaluated, and the potential lack of reproducibility both in the performance and interpretation of the examinations was not addressed. The second limitation is that this population was symptomatic so that only sensitivity was evaluated and specificity was not addressed. We believe that the relatively small sample size of symptomatic women in this study reflects true clinical practice, in which most DCIS cases are diagnosed on mammography in asymptomatic women and symptomatic DCIS is uncommon. It is nonetheless incumbent on sonographers to learn about this disease because although it is usually detected on mammography, it can also be detected on sonography.
The findings in this small study suggest that sonography may have a role in cases of mammographically detected abnormalities in symptomatic women and possibly in every symptomatic patient. This possibility is exemplified in the three women who presented with unilateral symptoms but had bilateral DCIS diagnosed with the aid of sonography.
The optimum role of sonography in this group of patients may be to evaluate palpable abnormalities when findings of mammography are negative or nonspecific, patients with nipple discharge in whom galactography is not possible, and nonspecific mammographic soft-tissue densities without calcification. The role of bilateral whole-breast sonography as a screening tool has also recently been addressed in the literature [34, 35]. Additionally, galactography has been shown in this study to be a useful adjunctive diagnostic tool in the subset group of patients with DCIS who present with spontaneous uniorificial nipple discharge, particularly in those women who present with negative findings on mammography and sonography in whom galactography remains the imaging technique of choice.
Acknowledgments
We thank Peggo K. W. Lam, Center for Clinical Trials and Epidemiological
Research, Chinese University of Hong Kong, Prince of Wales Hospital, Shatin,
NT, Hong Kong, for statistical assistance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. L. Thakur, D. Devadhas, K. Zhang, R. G. Pestell, C. Wang, P. McCue, and E. Wickstrom Imaging Spontaneous MMTVneu Transgenic Murine Mammary Tumors: Targeting Metabolic Activity Versus Genetic Products J. Nucl. Med., January 1, 2010; 51(1): 106 - 111. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. J. Shin, H. H. Kim, S. M. Kim, G. Y. Kwon, G. Gong, and O. K. Cho Screening-Detected and Symptomatic Ductal Carcinoma in Situ: Differences in the Sonographic and Pathologic Features Am. J. Roentgenol., February 1, 2008; 190(2): 516 - 525. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G M Tse, P-H Tan, A L M Pang, A P Y Tang, and H S Cheung Calcification in breast lesions: pathologists' perspective J. Clin. Pathol., February 1, 2008; 61(2): 145 - 151. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. E. Dogan, K. Ceyhan, S. Tukel, S. Saylisoy, and G. J. Whitman Ductal Dilatation as the Manifesting Sign of Invasive Ductal Carcinoma J. Ultrasound Med., October 1, 2005; 24(10): 1413 - 1417. [Full Text] [PDF]
|
|
|
|
 |
|
 W. A. Berg, L. Gutierrez, M. S. NessAiver, W. B. Carter, M. Bhargavan, R. S. Lewis, and O. B. Ioffe Diagnostic Accuracy of Mammography, Clinical Examination, US, and MR Imaging in Preoperative Assessment of Breast Cancer Radiology, December 1, 2004; 233(3): 830 - 849. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
