| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Department of Radiology, ACC Steyr, Stadtplatz 30, A-4400 Steyr, Austria (O.G., G.H., C.G.); Department of Radiology, Medizinische Universität Wien, Vienna, Austria (T.H.H.); and Department of Radiology, University of California Medical Center, San Francisco, Calif (E.A.S.). Received February 10, 2006; revision requested April 11; revision received July 12; accepted August 22; final version accepted November 1. Address correspondence to O.G. (e-mail: Oswald.Graf{at}utanet.at).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
Materials and Methods: The local ethics committee approved this study; informed consent was not required. In 409 women, 448 nonpalpable solid masses were identified and classified as probably benign at US; at mammography these masses were either partially or completely obscured by dense fibroglandular tissue. Of the 448 masses, 445 were followed up, while biopsy was performed after initial imaging in the remaining three. False-negative rates, negative predictive values (NPVs), and exact 95% confidence intervals (CIs) were calculated.
Results: Of the 445 masses, 442 remained stable at follow-up (range, 2–5 years; mean, 3.3 years). Two masses increased (fibroadenomas at biopsy). One mass became palpable, and cancer was diagnosed at biopsy. The three masses in which initial biopsy was performed were fibroadenomas. The false-negative rate was 0.2% (one of 448; NPV, 99.8%; 95%CI: 0.0%, 1.23%).
Conclusion: Follow-up US appears to be an acceptable alternative to biopsy for solid masses with benign morphologic features seen at US owing to the extremely high NPV (99.8%).
© RSNA, 2007
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
According to the recently introduced Breast Imaging Reporting and Data System (BI-RADS) lexicon for US, a solid mass with circumscribed margins, oval shape, and parallel orientation can be classified as probably benign (category 3) (10). These lesions are estimated to have a very low probability (<2%) of being malignant, so short-term follow-up is proposed as a management strategy. However, the proper management for these findings remains controversial because long-term follow-up has not yet been established. It is feared that the incidental detection of solid masses at US will subsequently increase the negative biopsy rate unless reliable US criteria for probably benign masses are validated (11–13).
For circumscribed solid breast masses that are radiographically visible, follow-up with mammography is now accepted as standard practice in many institutions. This strategy is supported by scientific data and is used to reduce the number of low-yield biopsies (14–23). However, to our knowledge, there are no studies that report outcome of masses that are classified as probably benign only at US and are followed up only with US. Therefore, the purpose of our study was to retrospectively determine whether nonpalpable, solid, circumscribed masses that were partially or completely obscured at mammography and diagnosed as probably benign only at US can be safely managed with US follow-up.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
In our breast imaging unit, mammography screening is performed online when a radiologist is present. After conventional mammograms in the mediolateral oblique and craniocaudal projections have been viewed, immediate work-up of the patient with additional mammographic views, clinical breast examination, and/or US examination is performed if necessary. Women are not recalled for these additional examinations; all US examinations are performed by the interpreting radiologist.
When a mass is new or growing at mammography, when compared with prior studies, a probably benign assessment is not made and biopsy is recommended at our institution. For a circumscribed mass visualized at mammography, the final assessment of a probably benign finding is made only if the mass is completely visible (or <25% of its circumference is obscured by dense fibroglandular tissue) and if the mass shows no malignant features at US. At our institution, if at least 25% of the circumference of a mass is obscured, the final assessment is based only on the US study. Furthermore, at our institution, women with dense breasts on screening mammograms are routinely examined with US for mammographically occult breast cancer, allowing circumscribed solid masses visible only at US to be encountered.
In the study period, US helped identify one or more nonpalpable solid breast masses that were classified as probably benign at US in 466 women. Twenty-nine women younger than 30 years of age (our lower limit for mammography) were excluded because only US was performed. Twenty-eight women were excluded because they were lost to follow-up. Thus, 409 women were included in this study (mean age, 47.9 years; range, 32–76 years).
The women were referred either for screening or for evaluation of diffuse breast pain. None of the women had palpable breast lesions. Three hundred fifty-three of the 409 women were aged 40 years or older, which is the recommended age range for screening at our institution. In 406 of the women, both breasts were examined. In the other three, only one breast was examined because a mastectomy had been performed 6–23 years earlier.
Imaging and Image Interpretation
Standard mammograms were obtained in mediolateral oblique and craniocaudal projections by using a dedicated mammography unit (Senographe 800T; GE Healthcare, Milwaukee, Wis). Film processing optimized for mammography was used (Kodak, Rochester, NY). Mammograms were evaluated by one of four radiologists (O.G., G.H., two nonauthors) with 5–12 years experience in mammography and US. Routine interpretation of a mammogram includes determination of breast density according to the American College of Radiology categories 1–4. In women with dense breasts (categories 3 and 4) US examination is performed by the same radiologist who interpreted the mammograms.
In the case of a circumscribed mass that was partly obscured by breast tissue on standard projections, a mediolateral whole-breast view and spot-compression magnification views in two other projections were ordered by the radiologist to help characterize the findings with greater precision. US was then performed to determine the cystic or solid nature, border characteristics, and internal features of the mass.
US was performed by using 11–14-MHz transducers (Powervison 6000; Toshiba, Tokyo, Japan), and findings were documented on laser film hard-copy prints (Kodak). The following US criteria were used to define a probably benign solid breast mass: shape (oval or macrolobulated); circumscribed margins of the entire circumference of the lesion; width greater than height (long axis parallel to the skin surface); echogenicity (isoechoic or mildly hypoechoic); and no acoustic shadowing (10). All criteria were required for a probably benign assessment. A circumscribed area of low or identical echogenicity relative to that of subcutaneous fat was called a mass only if the lesion was seen in two planes. Women with typically benign cysts or intramammary lymph nodes (BI-RADS category 2) or with masses suspicious for malignancy (BI-RADS category 4 and 5) at US were not included in this study.
Patient Management
Patients were told that a mass was found whose morphologic characteristics at US suggested that the lesion was probably benign. They were informed that the estimated rate of malignancy in these lesions is less than 2%, and follow-up was offered. We advised the patients that preliminary data suggest a low likelihood of malignancy but that this is currently an estimate. Patients were also told that both benign and malignant lesions may increase in size (24–27). Enlargement of the mass of more than 10% of its initial diameter and/or change of the benign border characteristics at follow-up was considered an indication for biopsy (28).
The number, side, location (clock position), distance from nipple, and size (greatest diameter) of the lesions were recorded. A lesion-to-lesion comparison was performed at follow-up studies. The duration of follow-up was recorded for every woman.
If the mass was seen for the first time, a short-term follow-up study was scheduled for 6 months after. Mammography and a directed US examination were performed in partially obscured masses at this short-term follow-up study; in mammographically occult masses, only a directed US examination was performed. The next follow-up study with mammography and US was scheduled at 12 months. At this time, both breasts were examined with mammography and US because 1 year is the recommended interval for screening at our institution.
The next follow-up examination with mammography and US was performed after another year. After following a mass for 2 years or longer with demonstrated stability, the assessment was changed to benign (BI-RADS category 2, most likely fibroadenoma). After 2 years, the women were followed up in our recommended screening interval of 1 year.
Statistical Analysis
Statistical analysis was performed with the StatXact-6 program (2004; Cytel Software, Cambridge, Mass). The false-negative rate (FNR) and negative predictive value (NPV) were calculated on a per-lesion basis. To address the study assumption that fewer than 2% of the masses were malignant, frequencies were described by using percentages, and corresponding exact 95% confidence intervals (CIs) were calculated assuming independence of the lesions (Blyth-Still-Casella procedure). Furthermore, when clustering dependency of data in patients with multiple lesions is possible, the FNR, NPV, and exact 95% CIs were also calculated on a per-patient basis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
In 90 (22.0%) of the 409 women, a circumscribed mass was visible at mammography in which, after full diagnostic mammographic work-up, more than 25% of its circumference was obscured by adjacent breast tissue. These masses were identified as noncystic circumscribed lesions of low echogenicity at US (Fig 1). In six of these 90 women, additional lesions of similar morphology were detected at US that were not visible at mammography. In five of these six women, one additional lesion was found; in the sixth woman, two additional lesions were found. In total, 97 lesions were detected in these 90 women.
|
|
|
In three (0.7%) of the 409 women, histologic work-up of a solitary probably benign mass was performed when the mass was initially detected. Two of these women had a coincident breast cancer in the ipsilateral breast. In one of these two patients the cancer was diagnosed at both mammography and US (Fig 2). In the other patient, the cancer was mammographically occult and was found only at US. A circumscribed mass that was mammographically occult in both women was also detected at US. In both patients, a fibroadenoma was diagnosed as the second lesion at histologic work-up when the cancers and probably benign masses were surgically excised. The third patient had undergone mastectomy 6 years earlier and chose biopsy of the probably benign mass (12 mm in largest diameter) rather than follow-up. A fibroadenoma was diagnosed after US-guided 14-gauge core biopsy.
|
|
|
|
In a 45-year-old patient, a mammographically occult lesion interpreted as probably benign at US became palpable 4 months after initial imaging. No imaging study was performed prior to surgical excision. A 12-mm invasive ductal cancer (tumor stage pT1cN0) was diagnosed. At initial US, this circumscribed lesion was 9 mm in largest diameter (Fig 3).
|
|
|
If we restrict our patient population to women aged 40 years and older, corresponding to the age range in which screening is routinely performed, one of 385 probably benign masses proved to be malignant, which is a similarly low FNR of 0.26% (exact 95% CI: 0%, 1.37%; NPV, 99.74%) on a per-lesion basis. On a per-patient basis, the FNR was 0.28% (one out of 353; exact 95% CI: 0%, 1.48%; NPV, 99.72%).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
However, US screening for mammographically occult cancer is controversial (11,12,32–37). It is argued that there is no proof that US screening lowers breast cancer mortality, which, if proved, would be the most important justification for applying a screening test in an asymptomatic population. A beneficial effect (ie, mortality reduction) has been demonstrated only for mammographic screening. Thus, US screening should not be used unless the effectiveness of lowering mortality is shown in a randomized controlled trial.
Furthermore, negative consequences resulting from screening healthy people with US are unavoidable. False-positive findings occur for all screening tests, and with screening US these result in anxiety, additional imaging examinations, biopsies, and consequent downstream costs.
For some authors (38–40), a solid mass seen only at US is an indication that a biopsy is needed because there is no proof that US can differentiate between benign and malignant lesions. For others (41–48), US is key in the evaluation of a breast mass. The US subset of BI-RADS classifies breast masses into different categories according to their morphologic features (49). BI-RADS category 3 is used for findings whose morphologic characteristics suggest that a lesion is probably benign. Fewer than 2% of lesions classified in this category should be malignant. It has been suggested that these masses may be managed with periodic imaging surveillance, and that stability at imaging for at least 2 years strongly indicates benignity. This approach should lead to a reduction of biopsies that ultimately yield a benign result. For mammography, this strategy has been supported by many studies (15–23); however, for US, this has not yet been proved, and the risk of malignancy among prospectively identified masses has not been clearly established.
The results of this study support the approach that solid breast masses with the US morphologic features specified in BI-RADS category 3 can safely be managed with periodic imaging surveillance, and that biopsy can be averted. The vast majority of the masses in our study (442 of 445; 99.3%) remained stable for at least 2 years (range, 2–5 years; mean, 3.3 years) and the risk of malignancy was very low (0.2%).
Previous single-institution studies also have reported a high NPV for circumscribed solid breast masses with probably benign morphologic features at US. In articles by Stavros et al (41) and Mainiero et al (44), the NPV was 99.5% and 99.3%, respectively. In our patients, a similarly high NPV was found (99.8%). Only one (0.22%) of 448 lesions that was called probably benign at US proved to be malignant (Fig 3). This cancer became palpable 4 months after initial evaluation and was excised before the next follow-up study was scheduled. In this case, the cancer mimicked a benign lesion at initial imaging.
In the report by Mainiero et al (44), patient age was similar to that in our study (mean, 45 years; range 18–83 years). We have also calculated the NPV for probably benign masses in women aged 40 and older because breast cancer is uncommon in women younger than age 40, so the truly important NPV is limited to those women older than 40 years. By restricting the data to patients aged 40 and older, the NPV was similarly high (99.7%), compared to the NPV when women younger than 40 years of age are included (99.8%). The assertion that US screening may substantially increase the number of negative biopsies is not supported by our data (32). The negative biopsy rate in those women who were initially followed up was only 0.45% (two of 444 lesions).
In 7.6% (31 of 409) of the women with multiple lesions studied, similar probably benign morphologic features were found at US. Multiplicity, especially if bilateral, may contribute to the level of confidence in a probably benign assessment. Moreover, in the two patients who had coincident cancer, US permitted differentiation between malignant and probably benign masses. Nonetheless, we do not recommend follow-up instead of biopsy for a benign-appearing mass in women with ipsilateral cancer.
This report had limitations. This retrospective study does not provide any data on the prevalence of benign solid breast masses at mammography in women with dense breast tissue. Furthermore, our study does not provide evidence of benefit for complementary screening US. However, it does suggest that the negative effects of incidental findings by using US may be limited principally to anxiety and the cost of additional imaging, rather than also including a large number of benign biopsies. More studies are needed to define the role of US in the breast cancer screening process. A multicenter trial in high-risk women with dense breasts, with readers blinded to the mammographic results, is currently underway to determine who may benefit from supplemental US screening (50). This study also will produce large-scale multi-institutional data on the NPV of solid masses that are characterized as probably benign at US. Until these data are reported, the only published experience involves single-institution studies; our study adds to this experience, indicating that circumscribed solid masses that fulfill the morphologic criteria for US BI-RADS category 3 may have an extremely high NPV.
|
ADVANCE IN KNOWLEDGE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
|
FOOTNOTES |
|---|
Abbreviations: CI = confidence interval • FNR = false-negative rate • NPV = negative predictive value
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, O.G.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, all authors; clinical studies, O.G., T.H.H., G.H.; statistical analysis, O.G., C.G.; and manuscript editing, all authors
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCE IN KNOWLEDGEReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. H. Shin, B.-K. Han, E. Y. Ko, Y. H. Choe, and S.-J. Nam Probably Benign Breast Masses Diagnosed by Sonography: Is There a Difference in the Cancer Rate According to Palpability? Am. J. Roentgenol., April 1, 2009; 192(4): W187 - W191. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Mesurolle, V. Sygal, L. Lalonde, A. Lisbona, M.-P. Dufresne, J. H. Gagnon, and E. Kao Sonographic and Mammographic Appearances of Breast Hemangioma Am. J. Roentgenol., July 1, 2008; 191(1): W17 - W22. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
