| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Innovation |
1 All authors: Department of Breast Imaging, ACC-219, Massachusetts General Hospital, 55 Fruit St., Boston, MA 02114.
Received March 14, 2000;
accepted after revision January 3, 2001.
Presented at the annual meeting of the American Roentgen Ray Society,
Washington, DC, May 2000.
Introduction
 Top Introduction Materials and MethodsResultsDiscussionReferences |
|---|
Occasionally, a carcinoma may arise adjacent to microcysts. Therefore, analysis of the characteristics of all the calcifications in a patient is important [2]. It is not unusual to encounter patients with calcifications that do not have a uniform appearance on the true lateral projection. For these patients, the diagnosis of milk of calcium is not certain. In patients with calcifications that appear to be milk of calcium (a benign finding) but do not layer exactly on lateral images or in patients with calcifications that do not layer (suggestive of needing biopsy), we have found that a pendent projection is helpful in confirming the diagnosis of milk of calcium. Calcium precipitated into cysts is influenced by gravity and will have a different morphologic appearance on images obtained with the patient in a prone pendent versus those obtained with the patient in the lateral horizontal-beam projection because of shifting in the cysts as the patient leans forward. This change in morphology allows one to more confidently dismiss the calcifications as benign in patients in whom these criteria are met.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Therefore, a total of 17 patients have been imaged in magnified craniocaudal, true 90° lateral, and pendent lateral positions for further evaluation of indeterminate calcifications detected on their screening mammograms.
We used a standard mammographic unit (Bennett Contour B-115; Trex Medical Systems, Danbury, CT) to image the front of the breast in a pendent position in all our patients. This unit allows the detector system to be rotated, a feature that permits the patient to flex at the hips at a 90° angle (Fig. 1), resulting in the breast falling naturally into the pendent position. The breast was then compressed, and the image was obtained using the standard technical factors for magnification.
|
After obtaining these images, we performed a study to determine if the pendent view was useful. Six interpreters who were unaware of each other's findings gave a Breast Imaging Reporting and Data System (BI-RADS) [3] assessment after viewing the craniocaudal and true lateral images for each patient. Subsequently, each interpreter was shown the image obtained in the pendent position to determine if the image obtained in this additional view altered his BI-RADS assessment and, therefore, a patient's treatment.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In five of the 10 patients, the prone pendent image confirmed that the calcifications did not shift to form a concave appearance, with the concavity toward the chest wall, and so they could not be considered milk of calcium. The possibility of malignant calcifications could not be excluded, and biopsy was suggested. In one of these patients, indeterminate calcifications were adjacent to milk of calcium within a macrocyst (Fig. 2A,2B,2C). The calcifications in two patients were found to be ductal carcinoma in situ and in another patient, invasive ductal carcinoma. The calcifications in the other two patients represented fibrocystic changes.
|
|
|
In the remaining five of the 10 patients, the prone pendent image confirmed the classic appearance of milk of calcium so the diagnosis was made with confidence. In one patient, the calcifications did not layer on the image obtained in the true 90° lateral position; the shift of the calcifications was identified on the pendent image (Fig. 3A,3B,3C). No further workup or biopsy was necessary. None of these patients has developed cancer (mean follow-up time, 27 months; range of follow-up time, 18-116 months).
|
|
|
In the remaining seven patients, a short interval of follow-up was suggested. None of these patients has developed cancer at this time, although the mean follow-up interval has been 20 months (range, 18-23 months) rather than the usual full 24 months.
Our retrospective analysis reveals that our six interpreters changed their BI-RADS assessment and, therefore, altered the treatment of 79.4% (range, 58.8-88.3%) of the patients. On the basis of the additional image, all six interpreters changed their assessments for four patients (23.5%); five of the six changed their assessments for seven patients (41.2%); and four of the six changed their assessments for the remaining six patients (35.3%).
The pendent image was most helpful for the five patients for whom we ultimately recommended a biopsy. Among the five patients and six interpreters, there were 30 analyses. In evaluating the images for these five patients, the interpreters found the pendent image altered the treatment of patients 86.7% of the time, by changing the interpreters' initial assessments from BI-RADS category 2 or 3 to BI-RADS category 4 or 5. For 10% of the patients, the interpreters had already thought the findings were suspicious but did not change their BI-RADS assessments. However, the additional image increased the interpreters' confidence in recommending a biopsy. One (3.3%) of the interpreters did not change his BI-RADS category 3 assessment of one patient. No interpreter down-graded his assessment from a BI-RADS category 4 or 5 to a BI-RADS category 2 or 3.
Similarly, the additional image increased the confidence of all the interpreters in recommending routine screening for 90% of the patients whose mammograms aroused few initial suspicions among the interpreters. They did not favor short-interval follow-up or biopsy (a change from BI-RADS category 3 or 4 to BI-RADS category 2) for these patients. In 10% of the patients, the interpreters were fairly confident that the calcifications were benign, and they did not change their BI-RADS assessments. No interpreter changed his assessment and recommended a biopsy on the basis of this additional image when his initial impression had been one of low suspicion.
Among patients with very small and faint calcifications, this additional image was not as helpful because it was difficult to identify a change in morphology in both the straight lateral and pendent positions. Radiographic images are limited by the physical properties of the calcifications. In 28.6% of the readers' analyses, the pendent image did not change the interpreters' BI-RADS category 3 assessment. The interpreters found the pendent image altered recommendations for the patients' treatment in 71.4% of their analyses, by changing their assessments from BI-RADS category 2, 4, or 5 to BI-RADS category 3.
In these borderline cases, the additional image increased the interpreters' confidence in recommending short-interval follow-up in 54.7% of their analyses (changing from BI-RADS 4 to 3). The additional image led the interpreters to up-grade their assessment from a BI-RADS category 2 to 3 in 16.7% of their analyses. We acknowledge that analysis of calcifications must be based on their morphology on images obtained in all three projections. One should be careful to avoid using an intuitive approach to assign cysts to the "probably benign" category, thereby delaying diagnosis of breast cancer. To date, none of these seven patients has been diagnosed with cancer during a mean follow-up time of 20 months.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Precipitated calcium may have multiple appearances, the most common being milk of calcium. Linden and Sickles [2] reported that the most common mammographic appearance of sedimented calcium is milk of calcium in microcysts, which are found bilaterally in scattered and occasionally clustered arrangements. The second most common appearance is a single unilateral cluster of microcysts containing milk of calcium [2]. Other appearances include bilateral clustered, unilateral scattered, and milk of calcium in a macrocyst [2]. The particles may be a very fine powder or small balls of calcium.
Calcifications are not always free to move in cysts and can form benign concretions in the lobular acini. These deposits likely account for the small (<1 mm) smooth round deposits that are sometimes found tightly packed together. These acinar calcifications may occasionally be heterogeneous and difficult to differentiate from cancer. Cancer can grow from the ducts back into the acini, but when calcifications form in these tumors, they are usually in the necrotic portions of the tumor and form irregular particles [1].
The importance of recognizing milk of calcium breast cysts is establishing this benign process, because these cysts have no known potential for malignancy. The linear and curvilinear calcifications seen on mammograms obtained with a lateral horizontal-beam projection might be misinterpreted as possible carcinoma and lead to unnecessary biopsy [4]. On mammograms obtained with the straight lateral and pendent projections, the upper margin of each calcification represents the interface between water-density cystic fluid above and heavier milk of calcium below; the curving lower margin reflects the rounded inferior margin of the cyst [2]. Recognizing a shift in the morphology of preciptated calcium under the pull of gravity is the principle behind performing this extra projection. Because fixed deposits in cancer cannot shift, the shift of calcium in the pendent view allows the radiologist to make a confident diagnosis. Although we do not have a full 2 years of negative follow-up for all patients, we know that no other kind of calcifications change morphology when the patient is imaged in different positions.
In conclusion, in the most patients, milk of calcium may be diagnosed using the two standard mammographic projections. In patients with calcifications that are equivocally diagnosed as milk of calcium, a mammogram obtained with an additional image, the pendent projection, may be helpful for two important reasons. First, occasionally a carcinoma may arise adjacent to microcysts [5]. Sickles and Abele [6] stressed that even if milk of calcium is present, biopsy should be considered if punctate microcalcifications are found. Second, the visualization of shifting calcifications can increase confidence in a diagnosis of benign cysts and eliminate the need for further unnecessary workup, short-interval follow-up mammography, or biopsy.
|
TopIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  F. M. Hall, L. Moy, and D. B. Kopans Milk of Calcium on Mammography Am. J. Roentgenol., March 1, 2002; 178(3): 762 - 762. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
