| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215 e-mail: fhall@caregroup.harvard.edu
Editor:
I enjoyed the thought-provoking article by Dr Kaplan (1) about screening breast ultrasonography (US). If this examination is ready for prime time, I hope that associated automated imaging and computed-aided diagnosis are not far behind because the examination will certainly exacerbate the present shortage of radiologists and imaging technologists.
Dr Kaplan (1) found a 0.3% cancer detection rate at screening breast US; a rate similar to that found at screening mammography. This percentage seems high, because 25%–40% of mammographically diagnosed cancers are identified on the basis of the presence of calcifications without an associated mass. Most of these cancers, albeit not all, would not be diagnosed with breast US.
Women who were offered the opportunity to undergo screening breast US in the study by Dr Kaplan (1) needed to return to the facility for a separate appointment. It is possible that those who chose to do so were disproportionately concerned about, and perhaps at higher risk for, breast cancer. This might elevate the cancer detection rate of the examination.
Dr Kaplan (1) points out that "the similarity in cancer detection rates between screening mammography and bilateral whole-breast US demonstrated in this study would probably not hold true if women with primarily fat-replaced breasts (BI-RADS [Breast Imaging Reporting and Data System] density category 1 or 2) were also included," because masses in dense breasts are seen less well at mammography. Mammography and breast US are also complementary for another reason. Dense parenchyma is more echogenic than fatty parenchyma and, therefore, normally hypoechoic malignancies are more conspicuous on US scans in mammographically dense breasts. This was demonstrated on the images (1).
Dr Kaplan (1) performed the screening in 1,862 selected subjects with BI-RADS density category 3 or 4. We are not informed of the total number of mammographically screened women from which the US-screened group was selected, the number of women who were offered US and refused, or the percentage of women in each of the BI-RADS density categories. The latter information is of concern to me because of the lack of rigorous and uniform criteria for placing breasts in the four BI-RADS density categories and the resulting inability to compare published data (2). Findings from recently published series have shown the prevalence of BI-RADS density category 4 to vary from 2% (3) to 51% (4). In the latter study, 85% of women had BI-RADS density categories 3 or 4 and, therefore, would have qualified for screening breast US with the criteria of Kaplan (1).
Dr Kaplan (1) expresses appropriate concern for the positive predictive value of only 11.8% for lesions that were examined at biopsy and that he identified with US. However, he states that "the number of biopsies performed and the positive predictive value of biopsy recommendations should [be able to] approach that of screening mammography" by using better-trained radiologists, state-of-the-art high-resolution equipment, and adherence to strict criteria for diagnosis. I think not. The reason that most experienced breast sonographers prefer to perform only targeted breast US is because of the relatively large number of atypical findings, such as small complicated cysts, which cannot be definitively distinguished from cancer. In addition to the costs of the many false-positive biopsy findings in this series, I believe there inevitably will be even greater "downstream" expenses from indeterminate US findings that require follow-up rather than biopsy.
I wonder if Dr Kaplan (1) obtained randomly the hard-copy sonograms that documented normal tissue, such as images of each quadrant of each breast, when the US examinations were normal? If so, why? I believe that such commonly obtained images are unnecessary, add to the time and cost of the examination, and might have medicolegal implications if the woman subsequently develops breast cancer, especially in the setting of the present study in which screening was performed by technologists.
REFERENCES
Strax Institute, 4300 N University Drive, Suite E-200, Lauderhill, FL 33351 e-mail: skaplan659@aol.com
I greatly appreciate Dr Hall’s comments concerning my article (1). I have responded separately to each of the points made by Dr Hall.
In regard to the 0.3% cancer detection rate in my study, Dr Hall points out that this percentage seems high given the number of cancers that are detected on mammograms as a result of calcifications. However, the 0.3% cancer detection rate in this study is the same as that in two previous reports (2,3), in which the use of US to detect occult breast cancer in women with dense breast tissue was evaluated. Also, since the study was done only in women with dense breasts, to use Dr Hall’s comparison accurately, we would need to know the percentage of cancers detected on mammograms as a result of calcifications in only women with dense breasts, not the percentage of all screening patients. To my knowledge, no study findings have been published that document this percentage. I suspect the percentage of patients with dense breasts in whom cancers are detected at screening mammography solely on the basis of the presence of calcifications is lower than the 25%–40% percentage quoted by Dr Hall. In addition, although calcifications without an associated mass detected at mammography often do not have an associated US finding, in my experience, many cases in which calcifications are present do have suggestive findings at US as well. This fact is probably underappreciated, since US is not routinely performed as an additional diagnostic test when calcifications without an associated mass are detected at mammography. This would further lessen the difference between screening mammography and screening US in the expected cancer detection rates.
The women who returned for screening US in our study constituted a diverse group of patients representative of our breast imaging practice. There is no evidence to support the theory that these women were more or less concerned about their cancer risk or were at a higher risk for breast cancer than patients who did not return for screening US, although this parameter was not evaluated in our study. The same information was provided for all patients who were recommended to undergo the examination. There is no reason to believe that these patients had any different fears about breast cancer risk or had a difference in risk factors for breast cancer than those who sought screening mammography in the first place.
The point that dense parenchyma is more echogenic on a US scan than fatty parenchyma and, therefore, a hypoechoic solid mass will be more obvious in dense tissue is a valid point that verifies the claim made in the article that imaging of women with fatty breast tissue would be less effective. I thank Dr Hall for this insightful comment.
Unfortunately, I cannot provide specific data regarding the number of patients in the categories requested by Dr Hall. This information was not specifically recorded as part of the study. I can, however, give estimates based on our typical screening population. Our center typically performs between 50 and 60 screening mammographies per day. Some weeks, screening is done 4 days per week; at other times, 5 days per week. Over the 21-month period during which the study was performed, the total number of screening mammographies performed would be approximately 19,000. The percentage of women in each of the BI-RADS density categories in our practice is estimated to be 20% with fat-replaced breasts (category 1), 40% with scattered fibroglandular density breasts (category 2), 30% with heterogeneously dense breasts (category 3), and 10% with extremely dense breasts (category 4). Therefore, approximately 40% of the women in our screening mammography population were eligible for screening US. This percentage is reflective of the average age of our population of patients who are predominantly eligible to receive Medicare. In a younger population, the percentage of women eligible for screening US would likely be higher given the higher frequency of dense tissue in younger women. However, I have also had experience in breast imaging in a younger patient population, and I do not believe that the percentage of women eligible for screening US would approach the 85% cited by Dr Hall.
We have no true estimate of the number of patients who refused screening US, but once the study was underway, and the referring physicians became familiar with its parameters, the majority of patients who were offered screening US underwent the examination. I would estimate that at least 70% of patients, if not more, who were offered the examination underwent it. This estimate is based on the experience of seeing patients who have returned for their annual mammograms after screening US had been recommended the previous year. This type of follow-up experience has now been available for more than 3 years.
The argument that the positive predictive value cannot be improved over the 11.8% demonstrated in our study is not valid. The reasons that the positive predictive value of biopsy may have been lower than optimal are discussed in the article. Dr Hall mentions the "relatively large number of atypical findings" at screening US that would result in a low positive predictive value. However, the majority of these findings are complicated cysts, as stated by Dr Hall. These findings would not affect the positive predictive value of biopsy, as the cysts could be aspirated to verify their benign cystic nature, thus not requiring biopsy. Aspiration of cysts, even cysts smaller than 1 cm, is a procedure that is quick and easy to perform in the hands of an experienced breast-imaging radiologist, and it is also well tolerated by patients. As to the increase in costs related to additional procedures and follow-up examinations that may be necessary, the same argument can be made regarding screening mammography. Yet, we are willing to accept these added costs given the benefits of early detection of breast cancer. The same standard should apply to any test proven effective in the diagnosis of otherwise occult breast cancer.
Last, we did obtain a single image of each quadrant of the breasts bilaterally when the examinaton finding was negative. This was done to document that the area was evaluated. I too agree that this is unnecessary, time consuming, and adds to the cost of the examination. However, without some hard-copy documentation, a challenge from a reimbursement standpoint could be made as to the actual performance of the examination. If the finding of this examination was negative, we obtained only images printed on paper rather than images on film, thus limiting costs. We do obtain hard-copy film images along with paper images for any actual findings, even if they are simple cysts that require no further action. I do not believe that documenting negative areas would have substantial medicolegal implications. As long as there are no images that show a lesion that was overlooked, which certainly could be cause for a malpractice claim in any imaging situation, having negative images of an area in which breast cancer develops subsequently does not seem to imply negligence. This situation is similar to having a negative clinical breast examination finding and developing breast cancer the following year. If there is no documentation that an abnormality existed, there should be no negligence if a cancer develops in the future. This situation is different than an abnormality that is overlooked on a mammogram and is seen in retrospect as something that has grown or changed. Documentation of negative US scans is also a procedure that is routinely done for all areas of the body outside of the breast.
Once again, I thank Dr Hall for his insightful and thought-provoking comments.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
