| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Departments of Radiology (J.A.H.), Pathology (R.E.F.), and Surgery (M.M.M.), University of Virginia, Box 170, Charlottesville, VA 22908. Received September 10, 1998; revision requested October 22; final revision received June 9, 1999; accepted August 18. Address reprint requests to J.A.H.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Mammographic change in size was evaluated by measuring the distance from the nipple to the pectoralis major muscle on the mediolateral oblique view of the diagnostic mammogram and on a preceding mammogram in 30 patients with ILC. Clinical, mammographic, and histopathologic findings were retrospectively reviewed.
RESULTS: Five patients (17%) had an ipsilateral decrease in mammographic size. No patients noticed a physical decrease in breast size. Patients with an ipsilateral decrease in mammographic size most commonly had breast thickening at examination (four of five patients [80%], P < .001) and either a focal asymmetry density (three of five patients [60%]) or architectural distortion (one of five patients [20%]) at mammography; those patients with no change in size most commonly had a palpable mass (six of 25 patients [24%]) or normal findings (19 of 25 patients [76%]) and a mass (13 of 25 patients [52%]) at mammography. The mean tumor size was 66 mm for those with an ipsilateral size decrease and 16 mm for those with no size decrease (P < .001). At histologic analysis, tumors associated with an ipsilateral decrease in mammographic size had more diffuse involvement of the breast, and discrete masses were not seen.
CONCLUSION: An apparent decrease in mammographic size may help identify cases of ILC, especially when associated with thickening at clinical examination and focal asymmetric density at mammography.
Index terms: Breast, abnormalities, 00.327 • Breast neoplasms, diagnosis, 00.11, 00.327, 00.91
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
ILC is also difficult to diagnose at mammography (2). ILC more commonly manifests as a focal asymmetric density or architectural distortion than does infiltrating ductal carcinoma (5). The tumors are often of low density, even when large (1,3–5). Because of the frequent lack of a discrete mammographic mass and the low density of the tumor, ILC is associated with a higher mammographic false-negative rate than is infiltrating ductal carcinoma (2–4). An ipsilateral decrease in breast size has been described with advanced cases of ILC (1). No mammographic findings have been specific to ILC, although the tumor is often seen in only one projection and is commonly best seen in the craniocaudal view (3,5–7).
At microscopic analysis, ILC usually has a different pattern of infiltration than invasive ductal carcinoma, which may account for the increased difficulty in diagnosis with either physical examination or mammography. Although ILC may form a discrete tumor mass similar to that in most cases of invasive ductal carcinoma, it more commonly infiltrates the breast in single rows of cells, thus preserving normal anatomy without forming a dominant mass or associated dense fibrotic reaction. Because of the infiltrative pattern, ILC is often larger at histopathologic examination than at mammography (8,9) and is larger than invasive ductal carcinoma at diagnosis (5,8,9).
Although an ipsilateral decrease in breast size has been described with advanced cases of ILC, we noted an apparent ipsilateral decrease in mammographic size of the breast in cases of ILC that did not manifest as clinically advanced cancer. Since this finding may help identify cases of ILC, it prompted us to retrospectively review our cases of ILC to assess for a mammographic change in breast size. We also retrospectively correlated mammographic decrease in breast size with physical examination, mammographic, and histopathologic findings.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Change in mammographic size of the breast was assessed by one author (J.A.H.) by using the posterior nipple line, such that the distance of a line drawn perpendicular to the nipple to the pectoralis major muscle on the mediolateral oblique (MLO) view was used. For a few older mammograms, where the pectoralis major muscle did not extend to the posterior nipple line, this distance was estimated by extending a line along the pectoralis major muscle. Measurements were also attempted in the craniocaudal projection but were too imprecise because of the lack of a consistent landmark, such as the pectoralis major muscle, and were not used.
Measurements were done for both breasts on both the diagnostic mammogram and a preceding mammogram. Because some cancers may grow slowly, if available a mammogram that preceded the diagnostic study by 2–3 years was chosen for comparison over a mammogram obtained the previous year, which is routine practice at our institution.
To control for variation in nipple-to-pectoralis distance owing to a change in breast size from fluctuation in body weight between studies, the contralateral mammograms were analyzed (J.A.H.). The nipple-to-pectoralis distance on the MLO view of the contralateral breast was measured on the diagnostic and preceding contralateral mammograms. The difference between the two contralateral MLO views was subtracted from or added to the difference noted between the mammograms of the breast with ILC.
Changes on the contralateral mammogram were also used as a control group to establish a meaningful change in the size of the breast affected with ILC. Change in the distance from the nipple to the pectoralis major muscle on the MLO view of the contralateral mammograms between the time of diagnosis of ILC in the opposite breast and the preceding study varied from an increase of 14 mm to a decrease of 6 mm (mean, increase of 0.5 mm). The SD of change was 3.6 mm. Therefore, after correction for fluctuation in body weight, a decrease of 10 mm or more in the nipple–to–pectoralis major muscle distance on the MLO view of the diagnostic and preceding mammograms of the breast affected with ILC was considered important.
For all study patients, mammography charts were reviewed (J.A.H.) for patient symptoms at presentation. Medical records were reviewed (M.M.M.) for clinical findings at the time of surgical consultation. Mammography reports were retrospectively reviewed (J.A.H.) for findings. The gross and microscopic examination results were retrospectively reviewed (R.E.F.) for 28 patients, without knowledge of which patients had a mammographic change in breast size. Histopathologic information was unavailable for the remaining two patients, because the surgical biopsy was performed at an outside institution and only representative microscopic slides were available for review.
The gross appearance of the cancers was classified initially as to whether they were visible. If visible, the cancers were evaluated (R.E.F.) for gross size and configuration (stellate, round, or ill defined). The size of the tumor was determined by its gross measurement and was later coupled with microscopic verification of the gross measurement.
At the microscopic examination, tumors were assessed for grade (Elston-Ellis method), stromal response (paucicellular, densely collagenous stroma vs cellular, loose, fibrous tissue), edema, lymphovascular invasion, and presence of tumor foci ("skip areas") more than 3 mm beyond the main tumor mass. Skip areas were designated as "local" if they were on the same slide as the primary tumor mass and as "distant" if they were on random sections. The Elston-Ellis method of tumor grading was used, in which a score of 1–3 was assigned for (a) tubule formation, (b) pleomorphism, and (c) mitotic count. The total score could range from 3 to 9, with a total of 3–5 representative of grade 1, a total of 6 or 7 representative of grade 2, and a total of 8 or 9 representative of grade 3. Axillary lymph node involvement was also reviewed.
Statistical analysis (J.A.H.) of raw data was performed by using the unpaired, two-tailed, Student t test to compare two groups. The two-tailed Fisher exact test with Yates continuity correction was used to compare the differences between two groups by using a contingency table.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
At the time of mammography, physical findings were reported as abnormal for all five patients with an ipsilateral decrease in mammographic size, whereas the findings were normal for 19 of the 25 patients (76%) with no mammographic change in size (Table 1) (P = .003, Fisher exact test). In the 19 women without a mammographic change in size whose physical examination findings were reported as normal at the time of mammography, at subsequent examination a breast surgeon found a palpable lump in three patients and thickening in one patient. In many of these patients, however, the surgeon had knowledge of the diagnosis of ILC at the time of physical examination because core-needle biopsy is often performed prior to evaluation by a breast surgeon at our institution.
|
Patients with an ipsilateral decrease in mammographic size also had more subtle mammographic findings than did patients with no mammographic change in size (Table 2). Four of the five patients with an ipsilateral mammographic decrease in size had abnormal mammographic findings, which were focal asymmetry (n = 3) and architectural distortion without an associated mass or density (n = 1). None of the patients with a marked ipsilateral decrease in mammographic size had mammographic masses, whereas 13 of the 25 patients (52%) with no mammographic change in size did (Fig 3) (P = .05, suggestive, Fisher exact test). Of note, one of the five patients with an ipsilateral decrease in mammographic size had a mammographically occult tumor (Fig 2). Thus, mammographic findings were subtle in those patients with an ipsilateral decrease in mammographic size and were more discrete in those with no mammographic change in size.
|
|
|
|
In the five patients with an ipsilateral decrease in mammographic size, three tumors measured less than 25 mm at gross examination but covered a much larger area when examined microscopically. Tumor not only extended adjacent to the visible lesion but also was present on random sections of breast that appeared normal at gross inspection (Fig 4). The conservative estimate of the size of the tumor for each of these three lesions was at least 50 mm in diameter. Thus, every tumor in patients with a marked ipsilateral mammographic decrease in size exceeded 50 mm. Moreover, microscopic skip areas were present in all cases (Fig 4).
|
|
Fibrous tissue formation was assessed microscopically. Four cases in the group with no change in mammographic size had dense central fibrous tissue that correlated with stellate masses at gross evaluation (Fig 3). All other cases had a complex architecture in which the fibrous tissue formed strands that irregularly intersected one another and seemed to run in haphazard directions. They varied in width from minute strands to bands 1–3 mm in thickness. Carcinoma cells were frequently located between the immediately adjacent fat cells.
The stromal response was also assessed microscopically and varied markedly within almost all of the tumors. All cases had at least a few foci with numerous fibroblasts and loose collagen formation. The local skip areas that were within 1 cm of the primary lesion and the distant skip areas had small areas of "active" fibroblasts, but most of the tumor cells were surrounded by dense, paucicellular collagen. Distant skip areas often occurred in or near ducts or lobules, and the tumor cells occasionally encircled normal ducts (Fig 4). Other skip areas had tumor within rounded or elongated strands of fibrous tissue situated in fat and devoid of epithelial parenchyma. As with most carcinomas, the desmoplastic response was the majority of the tissue. No association with fascial planes or ligaments was discernable.
Edema was not recognized in any case. Only a single tumor contained a solitary focus of lymphovascular invasion. This case was in the group with a marked decrease in mammographic size and lacked any other unusual features.
When assessing tumor grade by using the Elston-Ellis method, none of the tumors had tubule formation; therefore, the score was always 3. The degree of pleomorphism varied from none, or a score of 1, to moderate, or a score of 2. Mitoses never exceeded 10 per high-power field, or a score of 1. Thus, 12 tumors were grade 1 (scores of 3 + 1 + 1) and 16 tumors were grade 2 (scores of 3 + 2 + 1). No significant differences in histologic grade were present between groups.
Tumor involvement of axillary lymph nodes was higher in patients with a decrease in mammographic size, but not significantly so (P = .12, Fisher exact test) (Table 3).
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
All patients with an ipsilateral decrease in the mammographic size of the breast had abnormal breast examination findings, but for most patients the findings were subtle. Although the tumors were large, only one of the five patients had a discrete palpable mass. The remaining four patients (80%) had thickening at clinical examination, which is a nonspecific finding often associated with fibrocystic change.
Tumors that were associated with an ipsilateral decrease in mammographic size also demonstrated subtle mammographic findings. Focal asymmetry and architectural distortion were the only mammographic findings in four patients (80%), and the remaining patient had a mammographically occult tumor. A discrete mammographic mass was not present in any of these patients. In contrast, discrete mammographic masses were seen in 52% of patients without a change in mammographic size.
At gross examination, tumors that had no mammographic change in size were more sharply delineated than tumors associated with a marked decrease in mammographic size. At microscopic analysis, it was found that most of the tumors (83%) in the group without a mammographic change were well estimated by using their gross appearance. In contrast, all of the tumors associated with an ipsilateral decrease in mammographic size were very underestimated in size at gross pathologic analysis compared with the size at microscopic evaluation.
While there was considerable overlap in the microscopic appearance of the tumors, tumors associated with an ipsilateral decrease in mammographic size had more diffuse involvement of the breast, whereas discrete masses were not seen. Numerous microscopic skip areas were present in all patients with a marked ipsilateral decrease in mammographic size but were not seen in those with no change. Fibrous tissue formation was also more diffuse for patients with an ipsilateral decrease in mammographic size. Stromal response and tumor grade were similar for the study groups and therefore are not likely to be substantial factors contributing to the apparent ipsilateral mammographic decrease in breast size seen in some patients.
The overlap in histologic findings between groups suggests that the tumors that demonstrate an ipsilateral decrease in mammographic size do not represent a specific histologic subtype. However, cases of ILC that manifest as a discrete mass are unlikely to demonstrate this mammographic finding compared with the infiltrating form. The tumors that demonstrated an ipsilateral decrease in mammographic size may represent a more advanced state of the infiltrating form.
The mechanism of an apparent ipsilateral decrease in breast size on the mammogram without a physical decrease in size is unknown. The characteristics of ILC that may explain this finding are the bandlike infiltrating tumor, its association with loose but diffuse collagen formation, and frequent microscopic skip areas. These histologic findings could cause thickening at clinical examination without a discrete mass. During mammography, compression of the breast may be limited by the bandlike infiltrating tumor. Thus, the apparent ipsilateral decrease in mammographic size is probably related to reduced compressibility of the breast. Unfortunately, the degree of compression used for each view was not consistently recorded at our institution during this period, and we are unable to investigate this issue. Further study of this point may be useful to assess the mechanism of this mammographic finding.
Reduced compressibility of the breast with some ILC tumors may also explain why this tumor is commonly best seen in the craniocaudal view (3,5–7). Others have previously hypothesized that better depiction of ILC in the craniocaudal view is due to greater compression in this projection (6,7). This implies that ILC is better seen in this view because of decreased compressibility of the tumor. Although we attempted to measure the breasts in the craniocaudal projection for this study, the measurements were imprecise because of the lack of landmarks. Others have hypothesized that improved tumor visibility in the craniocaudal projection may be due to growth of the tumor around ligamentous structures (3), although we saw no evidence of this microscopically.
An ipsilateral decrease in mammographic size was found in 17% of cases of ILC in our series. This finding will probably aid in the diagnosis of this subtype of breast cancer, which is often difficult to diagnose clinically and mammographically. However, this finding may not be specific to ILC. We do not know how often this finding may be found in cases of infiltrating ductal carcinoma. Given the histologic differences between ILC and infiltrating ductal carcinoma, this finding is probably more common with ILC. Histologically, ILC commonly manifests as a poorly defined infiltrating process, whereas infiltrating ductal carcinoma commonly manifests as a defined stellate mass.
In addition, multicentric disease is twice as common with ILC than with invasive ductal carcinoma (10). We found a large number of ILC cases with gross or microscopic multicentric disease. It is well known that ILC often has microscopic disease in breast tissue that appears normal at gross pathologic inspection well beyond the confines of a palpable tumor. These findings are mirrored mammographically, where infiltrating ductal carcinoma most commonly manifests as a mass but where ILC manifests as a focal asymmetric density or architectural distortion.
To our knowledge, a decrease in breast size has not been described as a finding of invasive ductal carcinoma. Given the known clinical, mammographic, and histologic differences between ILC and invasive ductal carcinoma, we would not expect an apparent ipsilateral decrease in mammographic size without a clinically obvious tumor to be a finding of invasive ductal carcinoma. However, a study comparing mammographic breast changes in both invasive ductal and lobular carcinoma is necessary to prove that this finding is specific to ILC.
In advanced ILC or infiltrating ductal carcinoma, inflammatory changes may develop that are associated with breast edema and enlargement. Although many cases of ILC in our series were large, none manifested clinically as inflammatory carcinoma, and none had breast edema mammographically or histologically. This implies that an ipsilateral decrease in breast size may be a finding of large ILC without clinical or mammographic signs of advanced cancer.
In conclusion, an apparent ipsilateral decrease in mammographic size is a common finding with large ILC. This finding may be accompanied by a physical decrease in breast size in some patients when carefully examined, although this may not be noticed by the patient. Clinical and mammographic findings in these cases are subtle, commonly without discrete palpable or mammographic masses. The apparent ipsilateral mammographic decrease in size is likely due to decreased compressibility of the breast secondary to the infiltrating pattern, loose collagen formation, and skip areas of tumor growth seen in some forms of invasive lobular carcinoma. An apparent decrease in mammographic size may help identify cases of ILC, especially when associated with breast thickening at clinical examination and a focal asymmetric density at mammography.
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, J.A.H.; study concepts and design, J.A.H., R.E.F.; definition of intellectual content, J.A.H., R.E.F.; literature research, J.A.H., R.E.F.; clinical studies, R.E.F., J.A.H., M.M.M.; data acquisition, J.A.H., R.E.F., M.M.M.; data analysis, J.A.H., R.E.F.; statistical analysis, J.A.H.; manuscript preparation, J.A.H.; manuscript editing and review, J.A.H., R.E.F., M.M.M.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. A. Harvey, B. T. Nicholson, and M. A. Cohen Finding Early Invasive Breast Cancers: A Practical Approach Radiology, July 1, 2008; 248(1): 61 - 76. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Harvey Unusual Breast Cancers: Useful Clues to Expanding the Differential Diagnosis Radiology, March 1, 2007; 242(3): 683 - 694. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Qayyum, R. L. Birdwell, B. L. Daniel, K. W. Nowels, S. S. Jeffrey, T. A. Agoston, and R. J. Herfkens MR Imaging Features of Infiltrating Lobular Carcinoma of the Breast: Histopathologic Correlation Am. J. Roentgenol., May 1, 2002; 178(5): 1227 - 1232. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
