| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Education Exhibit |
POH Medical Center, Pontiac, Michigan
 |
Commentary |
|---|
 Top Commentary ReferencesResponse |
|---|
Screening mammography has a relatively high sensitivity of nearly 90% but bears the significant limitation of a specificity of only 35%, even in the best hands and in specialized centers. In everyday clinical practice at community hospitals, specificity may be realistically estimated at 10%–15%. I would consider myself to be a typical diagnostic radiologist, working in a 200-bed community hospital setting and reading about 1,000 mammograms per year. Yet the positive predictive value of my personal diagnoses as verified with excisional biopsy during the last 7 years did not exceed 13% despite generous use of breast US in American College of Radiology (ACR) BI-RADS Categories III ("Probably Benign Findings") and IV ("Suspicious Abnormality"). This is mostly due to lack of confidence in differentiating benign from malignant mammographic lesions and extraordinary malpractice pressure in the "radiologic trenches."
On the other hand, the high sensitivity of mammography is also compromised in patients with dense breast tissue and in cases involving postsurgical and post–radiation therapy follow-up, in which the false-negative rate is approximately 25%–30% (3). Additional challenges for radiologists include "multinodular" breast tissue and diffuse, indistinct microcalcifications.
Nuclear medicine imaging was being used in the differential diagnosis of palpable, superficially located breast lesions as early as 1946 and was performed with intravenous injection of radioactive phosphorus (P)–32 sodium phosphate and a Geiger-Muller handheld shielded probe (4). In Europe, the P-32 technique was used until the mid-1970s. Since that time, imaging with at least a dozen different radiopharmaceuticals (not counting 2-[fluorine-18] fluoro-2-deoxy-D-glucose positron emission tomography, immunoscintigraphy, and receptor scintigraphy) has been performed.
The perfusion radiopharmaceuticals used in nuclear cardiology (thallium-201, Tc-99m sestamibi [Cardiolite; DuPont Merck, Wilmington, Del], and Tc-99m tetrofosmin [Myoview; Amersham, Arlington Heights, Ill]) became the radiopharmaceuticals of choice in the diagnosis of breast cancer. However, only Tc-99m sestamibi (methoxy isobutyl isonitrile [MIBI], Cardiolite, Miraluma [DuPont]) is approved by the U.S. Food and Drug Administration for scintimammography in the United States, and Society of Nuclear Medicine procedural guidelines were recently outlined in detail by Khalkhali et al (5). In 1999, Taillefer (6) presented a detailed, comprehensive review and eloquently tabulated the results of Tc-99m MIBI scintimammography in 2,009 patients, which were published between 1994 and 1998 in 20 journal articles from around the world. The ratio of palpable to nonpalpable lesions was 2.3, and the average total accuracy for scintimammography was 86% (1,992 of 2,304 lesions). Its total average sensitivity was calculated as 84.5% (1,029 of 1,218 lesions) and its specificity as 89% (963 of 1,086 lesions). Total average positive and negative predictive values were calculated as 89% (1,029 of 1,152 lesions) and 84% (963 of 1,147 lesions), respectively (6). In 2000, Khalkhali et al (7) published the results of multicenter trials conducted in 673 women at 42 institutions in the United States and Canada (breast cancer prevalence, 40.1%) and calculated blinded readers’ diagnostic accuracy as 78%–81% and interreader agreement as 95%–100%. The institutional sensitivity and specificity in this trial were 87% and 76%, respectively for palpable lesions and 61% and 87%, respectively for nonpalpable lesions (7). As the authors reported, a 94% negative predictive value in women with a likelihood of breast cancer of 40% or less at mammography is the most important practical outcome of complementary use of scintimammography (7).
In the preceding article, Polan et al (8) presented unique scintimammographic results obtained in women with equivocal mammographic, US, or clinical findings: a sensitivity of 90% and a specificity 93.8%. All 27 cancers that proved to be true-positive at scintimammography were nonpalpable, and 11 were smaller than 1 cm.
The results in Category 1 patients (34 vague new mammographic densities with normal or near-normal US findings) are the most impressive. The sensitivity of scintimammography for 11 focal cancer lesions with a median size of 7 mm was 100%, and the specificity for benign lesions with a median size of 1.5 cm was 92.9%. Equally impressive was the 100% specificity of scintimammography in patients in Category 4 (14 women with a new lump or mammographic changes after surgery).
In my opinion, however, "minimal mammographic ... findings" might be a debatable and misleading description of Category 2 lesions (indeterminate calcifications). In 1993, Knutzen and Gisvold (9) calculated the likelihood of malignancy in 200 cases of indeterminate calcifications as 22%. Nevertheless, a sensitivity and specificity of 80% and 93.3%, respectively in this subgroup is remarkable.
The conclusion that scintimammography is a valuable, noninvasive complementary test in patients with a low or indeterminate likelihood of cancer at mammography is a plausible one.
Scintimammographic outcome is, in my opinion, particularly important in ACR BI-RADS Category III, "Probably Benign Findings," in which the prevalence of malignancy does not exceed 2%. Negative scintimammographic findings allow patients to be treated safely with short-term mammographic follow-up at intervals not exceeding 6 months, but positive findings are an indication for biopsy.
We recently presented the results of our analysis of the findings on 134 scintimammograms obtained at POH Medical Center (Pontiac, Mich) and Lapeer Regional Hospital (Lapeer, Mich) between January 1997 and March 2000 (10). Our results fully support this clinical approach.
Findings on 106 scintimammograms obtained in 78 women in ACR BI-RADS Category III were prospectively analyzed for negative and positive predictive values. Of these 78 women, 31 underwent surgical biopsy and 47 underwent at least 1 year of mammographic or US follow-up. In 72 of 78 studies, results were negative and demonstrated no cancer (positive predictive value = 100%). In four of the six cases that demonstrated positive scintimammographic findings, malignancy was proved at biopsy (positive predictive value = 67%) (10).
Category III patients also include young women with palpable nodules and unlikely malignancy as well as patients with dense breasts or multiple sharply defined, well-circumscribed nodules. Extrapolation of this approach to patients in ACR BI-RADS Category IV ("Suspicious Abnormality"), in whom the prevalence of malignancy is approximately 10%, is more controversial despite its recent introduction by Prats et al (11).
With its high accuracy, cost-effectiveness, and capacity to decrease morbidity and anxiety in patients and bolster confidence in attending physicians, scintimammography deserves the careful attention of a larger group of diagnostic radiologists when Category III findings ("Probably Benign Findings") are encountered.
|
References |
|---|
|
TopCommentaryReferencesResponse |
|---|
Department of Radiology, Los Robles Regional Medical Center, Thousand Oaks, California
|
Response |
|---|
|
TopCommentaryReferencesResponse |
|---|
We agree with Dr Kotlyarov’s observations concerning the limitations of screening mammography. As mammographers, we are daily confronted with mammographic findings (as well as US and physical examination findings) that are not obvious malignancies but still require our input as to management. Scintimammography would seem to be the proper approach to many of these ACR BI-RADS Category III dilemmas.
The authors of the nuclear medicine literature and the Society of Nuclear Medicine should be commended: Over the years, they have provided us with articles and lectures detailing the development and potential role of scintimammography. Many different research groups have separately concluded that scintimammography has a high sensitivity and specificity and have published their findings. As mammographers, we are the physicians most responsible for detecting breast cancers early and for simultaneously decreasing the number of unnecessary breast biopsies. In general, our awareness of recent advances in nuclear medicine is low. It is important that, in the future, the authors of the radiology literature and the leaders in breast imaging help educate general radiologists and mammographers regarding the significance of the complementary role of radionuclide breast imaging in the diagnosis of malignant and nonmalignant lesions.
Related Article
RadioGraphics 2001 21: 641-655.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOGRAPHICS | RADIOLOGY | RSNA JOURNALS ONLINE |
