| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Beth Israel Deaconess Medical Center and Harvard Medical School Boston, MA 02215
Soo et al. [1] report that in 420 patients with 455 palpable breast lesions the negative predictive value of sonography and mammography together was 99.8%. This extraordinarily high value may have been influenced by the authors' not unreasonable apparent definition of a palpable lesion as a lesion for which a patient is referred for breast sonography. However, in only 77% of patients in this study did the sonography report record the findings of the radiologist's physical examination. In these cases, a discrete mass was palpated in 35%, thickening was found in 19%, and palpation was normal in 46%. Readers are not told who originally noted the finding or referred the patient, or whether the patient herself palpated the lesion at the time of sonographic examination.
In our department, most patients referred for sonographic evaluation of a breast mass have never had the lesion confirmed by a physician, nor can any discrete lesion be palpated by the mammographer or technologist at the time of the examination. In fact, many women can no longer palpate the abnormality they may have initially felt, perhaps because the sonographic examination is being performed at a different time in the menstrual cycle. When no abnormality is palpated, the untargeted sonographic examination of the quadrant in question has less satisfactory results than when a lesion can be palpated. This situation arises, in our experience, because most of our referring general physicians and breast surgeons prefer or require that sonography or mammography be performed before they see a woman with self-palpated breast lesions. This process usually expedites the clinical workup. Even if no lesion is palpated by the physician, the clinical assessment alone is often not acceptable to the patient or to the physician because of the fear and medicolegal climate surrounding breast cancer. Our breast surgeons and clinicians point out that even if they palpate no abnormality in the area of patient concern, they remain at considerable medicolegal risk if the patient subsequently develops a cancer in that portion of the breast.
I would like to make several minor comments unrelated to the results of this study. In table I, the authors show that 26/420 (6%) of breasts had fatty parenchyma (Breast Imaging Reporting and Data System [BI-RADS] [2] category 1); 214/420 (51%) of breasts had dense parenchyma (BI-RADS category 4); and 85% were either dense or heterogeneously dense (BI-RADS categories 3 or 4) [1]. These percentages differ substantially in the literature and emphasize the lack of rigorous and uniform criteria for placing breasts into density categories and the resulting inability to compare published data [3]. More conformity is particularly important if clinical judgments are made on the basis of the BI-RADS categories, such as in a study in which the decision to perform whole-breast screening sonography was based on the BI-RADS density categories [4].
When the sonographic findings were normal, I wonder if Soo et al. [1] obtained random images of normal tissue, such as an image of each quadrant of each breast? Although these images are commonly obtained, I believe they are unnecessary, add to the time and cost of the examination, and might have medicolegal implications if the woman subsequently develops breast cancer in this area.
References
Duke University Medical Center, Durham, NC 27710
Dr. Hall's letter points out a common scenario in breast imaging centers: many women referred for sonography of a breast mass have never had a lesion confirmed by a physician. He suggests that this happens either because the imaging is ordered first to expedite the clinical workup or because of the patient's anxiety or the physician's concern about the medicolegal climate surrounding breast cancer. We experience the same situation in our clinic. In our practice, sonography targeted to the area of concern (the area indicated by the patient or, if the patient cannot identify the abnormality, the area indicated by the medical records or referring physician) is performed to address the clinical question—regardless of whether the lesion is considered a significant palpable abnormality by the radiologist. Therefore, we reported that many patients in our study had no palpable abnormality identified during the radiologist's physical examination [1]. We included these patients in our evaluation in an effort to determine the efficacy of breast sonography in confirming a negative result in this common clinical setting.
Hall also raises an interesting point in that the "percentages [of fatty and dense tissue] differ substantially in the literature and emphasize the lack of rigorous and uniform criteria for placing breasts into density categories and the resulting inability to compare published data." Many of us would agree that the assignment of Breast Imaging Reporting and Data System (BI-RADS) [2] density categories 1-4 is subject to substantial interobserver variability and that more rigorous and uniform criteria would help standardize parenchymal density descriptors. This standardization can be particularly important if clinical judgments are made on the basis of BI-RADS categories, as Hall suggests. In a previous letter to the editor [3], Hall describes his proposed categorization of parenchymal density by assigning density distribution to a bell-shaped curve "because categorization of breast density is relatively arbitrary and most biologic data fall into bell-shaped curves." Certainly this premise could be studied to evaluate the accuracy of this method. However, the parenchymal density in one study population may differ from that in another study or in the general population of patients undergoing mammography.
Our population consisted of patients with palpable abnormalities, many of whom were young [1], and certainly age has an effect on the density of fibroglandular tissue. Perhaps women with denser breast tissue identify a greater number of what they perceive to be palpable abnormalities than do women with fatty tissue, prompting a greater percentage of imaging workups. Other factors related to different study groups might also result in different parenchymal distribution patterns. In our study, we could not force the distribution of parenchymal density into a fixed percentage; we could only use the categories defined at the time prospective management was planned for each patient.
Additionally, not all patients with completely fatty breasts who present with a palpable lump require sonographic evaluation. In this scenario, it is our practice for the radiologist to first palpate the area of possible abnormality to determine if a lump is present and ensure that the region was included on the mammogram. The radiologist then decides whether sonography would be useful in the workup of the patient.
Finally, in response to Hall's query, it is our practice to target the sonographic evaluation to the area of concern; we do not screen the entire breast when evaluating a patient with a lump. If findings on targeted sonography are normal, we generally obtain two orthogonal images of the area of concern to document that sonography was performed in that area. We do not scan the entire breast, nor do we obtain random images of normal tissue in any other quadrant of the breast.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  F. M. Hall, T. M. Kolb, J. Lichy, and J. H. Newhouse Screening Breast US [letter] * Dr Kolb and colleagues respond: Radiology, May 1, 2003; 227(2): 607 - 609. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
