| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Joyce Eisenberg Keefer Breast Center, John Wayne Cancer Institute, St Johns Health Center, 1328 22nd St, Santa Monica, CA 90404 (R.J.B.). Affiliations for all other authors are listed at the end of this article. Received January 26, 2000; revision requested March 3; final revision received July 19; accepted August 28. Address correspondence to R.J.B. (e-mail: James.Brenner@stjohns.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX REFERENCES |
|---|
MATERIALS AND METHODS: CNB was performed in 1,003 patients, with results validated at surgery or clinical and mammographic follow-up. Mammographic findings were scored according to the American College of Radiology Breast Imaging Reporting and Data System with a similar correlative scale for histopathologic samples obtained at either CNB or surgery. Agreement of CNB findings with surgical findings or evidence of no change during clinical and mammographic follow-up (median, 24 months) for benign lesions was used to determine results. Three forms of diagnostic discrimination measures (strict, working [strict conditioned by tissue sampling error], applied [working conditioned by concordance of imaging and CNB findings) were used to evaluate the correlation of CNB, surgical, and follow-up results.
RESULTS: Strict, working, and applied sensitivities were 91% ± 1.9; 92% ± 1.8, and 98% ± 0.9, respectively; strict, working, and applied specificities were 100%, 98% ± 0.8, and 73% ± 0.9; strict, working, and applied accuracies were 97%, 96%, and 79%.
CONCLUSION: Percutaneous stereotactic CNB is an accurate method to establish a histopathologic diagnosis of nonpalpable breast lesions. Accuracy increases when additional surgery is performed for lesions with anticipated sampling error or when CNB findings are discordant with mammographic findings. An understanding of the interrelationship among these parameters is necessary to properly assess results.
Index terms: Breast, biopsy, 00.1261, 00.1267 • Breast neoplasms, diagnosis, 00.31, 00.32 • Breast radiography, utilization, 00.11
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX REFERENCES |
|---|
Two interrelated factors have limited the assessment of reports on stereotactically guided CNB. First, in only three studies is the degree of suspicion of lesions assessed with CNB reported. Of these studies, in only one were lesions categorized according to the standardized system recommended by the American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) (10). Second, results of follow-up in patients who did not undergo immediate surgery have been inadequate for determination of the sensitivity and specificity of stereotactically guided CNB. Such information was either not reported, reported for too few patients, or reported for an inadequate period of time (1–15,20). Follow-up data of longer than 6 months are unavailable for more than 95% of all reported lesions that were not subjected to subsequent surgical excision. Recently, two reports (21,22) have appeared from separate institutions in which sufficient follow-up data were provided to allow a better approximation of the true accuracy of CNB.
The purpose of our long-term, multiinstitutional, prospective study was to clarify the accuracy of stereotactically guided CNB of nonpalpable lesions of the breast in relation to a standardized image classification system and to determine the effect of this procedure on overall patient management.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX REFERENCES |
|---|
Biopsies were performed of nonpalpable lesions by using 14-gauge needles with a spring-loaded Biopty gun (Bard, Covington, Ga) or Promag gun (Manan, Northbrook, Ill), with a 22-mm excursion. All investigators underwent at least a 2-day course of on-site training with an application specialist. A minimum of five core biopsies were recorded. Although the equipment used in this study was commercially available and CNB was already a part of clinical practice, institutional review board approval was sought and obtained as appropriate, based on policies in existence at the seven institutions at the time the study began. Patients were not formally recruited, but rather the data were obtained for each sequential patient undergoing CNB at one of the seven institutions during 2 years.
Mammographically visible lesions were categorized into five groups to account for recommendations for biopsy: namely, mass, clustered microcalcifications, focal asymmetries, architectural distortions without definite mass, and mass with calcifications. The size of the lesion was determined as the average of three diameters measured on the mammogram. The lesions were graded according to the diagnostic categories prescribed by the BI-RADS lexicon as negative (BI-RADS category 1): benign finding (BI-RADS category 2); probably benign finding, short-term follow-up recommended (BI-RADS category 3); suspicious abnormality (BI-RADS 4); and highly suggestive of malignancy (BI-RADS category 5). Lesions were assigned to a BI-RADS category, and data were entered by the investigators (R.J.B., L.W.B., L.L.F., D.D.D., W.P.E., R.H., C.L., I.T., P.F., M.M., V.P.J., S.A.F., E.B.M., F.R.M.) at each respective institution, including three at one institution, two at another, and one at each of the others.
CNBs and surgical excisional biopsies were categorized according to their histopathologic diagnosis as follows: category 1, nonspecific benign findings (any benign epithelial or mesenchymal abnormality not specific for a given lesion: eg, fibrocystic change; ductal hyperplasia, usual type; stromal fibrosis); category 2, specific benign finding (eg, fibroadenoma); category 3, benign finding with anticipated tissue-sampling error necessitating further excision (eg, lobular hyperplasia with atypia, ductal hyperplasia with atypia, radial scar, phyllodes tumor, lobular neoplasia); category 4, ductal carcinoma in situ (DCIS); and category 5, invasive malignancy of any origin, including invasive ductal or lobular cancer. Categories were assigned by using the most clinically relevant histopathologic sample (eg, if one sample showed invasive cancer alone and the others showed DCIS alone, category 5 was assigned). Where data entry was omitted for imaging or CNB results, the sample was assigned to category 0.
Cases that involved calcifications underwent specimen radiography to confirm recovery of calcific particles. Specimen radiography for mass lesions was performed at the discretion of the operator. To ensure that the results reflected current clinical practice, no central pathologist was used. Rather, a pathologist at each institution rendered histologic diagnoses according to generally accepted criteria. The investigator thereafter assigned the correlative CNB category as designed for this study to the histologic diagnosis.
Patients with lesions assigned either a benign or probably benign mammographic diagnosis underwent core biopsy when the patient or patient’s physician was unsatisfied with a recommendation for surveillance, thereby resulting in referral for excisional biopsy. After CNB, all women with a CNB category 3, 4, or 5 diagnosis were referred for surgical excision. Those with concordant benign CNB and mammographic diagnoses (CNB category 1 or 2; BI-RADS category 1, 2, or 3) entered a surveillance mammography program. Patients with suggestive mammographic findings (BI-RADS category 4 or 5) were referred for surgical consultation, regardless of the CNB results. In all patients, clinical and mammographic surveillance were planned to last 18–24 months; this period was chosen on the basis of results from prior longitudinal studies (23–27) in which the development of interval cancers was evaluated. Surveillance was designed for a unilateral study (of the breast in which biopsy was performed) at 6 and 18 months and for a bilateral study at 12 and 24 months. Attempts were made by each institution to obtain follow-up data in all patients, including those referred for CNB by other institutions or health care providers.
Data were collected at each institution by using a commercially available software program (Q&A, version 4.0; Symantec, Cupertino, Calif) and were then forwarded to a central corporate collection unit. Data were monitored, analyzed, and subsequently audited under the direction of the principal investigator (R.J.B.) with the assistance of a biostatistician (J.S.). The sensitivity, specificity, and accuracy of CNB were determined by comparing the results of CNB with those of surgical excision or with results of clinical and mammographic follow-up. These parameters were determined in a manner to critically assess the reliability and role of CNB.
A definition of strict sensitivity was established such that any CNB diagnosis of a nonmalignant lesion that subsequently proved to be malignant was considered a false-negative. Because previous reports in the literature (28,29) provided a basis for the anticipation of tissue-sampling error with CNB for those benign histologic lesions in CNB category 3 (eg, atypical ductal hyperplasia), surgical excision was generally prompted by this CNB finding. Hence, a definition of working sensitivity was constructed, whereby category 3 lesions would be considered true-positive (concordant) when excisional biopsy samples showed malignancy and false-positive (discordant) when excisional biopsy samples showed no malignancy. Finally, a definition of applied sensitivity was developed to assess the effect of imaging findings together with CNB results in terms of optimal patient management.
When the imaging diagnosis was sufficiently suggestive to prompt excision, even though CNB demonstrated no malignancy or atypical hyperplasia (or any other CNB category 3 findings), the designation of "discordant mammographic findings" was applied, and a "positive" imaging diagnosis was included in the determination of true- and false-positive results, in addition to CNB category 3 lesions. Corresponding definitions of specificity and accuracy were also determined (see Appendix), and receiver operating characteristic curve analysis was performed to evaluate the effects of strict, working, and applied diagnostic discrimination criteria.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX REFERENCES |
|---|
Of the original 1,003 patients, 230 (23%) underwent surgical excision of their lesion within 4 weeks of CNB because of a suggestive BI-RADS or CNB category. Of the remaining 773 patients, another 177 (18%) underwent surgery that was delayed by the clinician or patient beyond 4 weeks and was prompted by the recommendation of another interpreting radiologist with a higher degree of suspicion or that was initiated during surveillance because of a change in management decision secondary to either a perceived interval mammographic change or a change in the clinical decision within the first 12 months. Five hundred ninety-six (59%) patients with concordant benign imaging and CNB results who did not undergo surgery were followed up with clinical and imaging surveillance. Of these, follow-up data were available for 475 patients (80%) at 12 months, for 388 (65%) patients at 18 months, for 338 (57%) patients at 24 months, and for 25 (4%) patients at 36 months. The overall mean follow-up was 19.3 months (median, 24 months; range, 0–36 months). During surveillance, an additional six patients underwent surgery because of a perceived change in the mammographic appearance of the original lesion (Fig 1).
|
Overall, 165 patients in whom CNB yielded benign histologic results underwent surgery because of discordant mammographic findings (BI-RADS category 4 or 5) or patient or physician preference for biopsy (or with no BI-RADS classification), including four cases with a BI-RADS category 3, resulting in 21 (8%) of 254 cases of malignancy diagnosed in this study. The relationships among the BI-RADS categories, CNB results, and surgical results are summarized in Table 1.
|
Surgery was not performed in five patients with a BI-RADS category 5 lesion and in 197 patients with a BI-RADS category 4 lesion because the CNB was considered definitively benign and was concordant with the mammographic findings (eg, clustered microcalcifications showing a hyalinizing fibroadenoma). None of these lesions accounted for malignancy identified during follow-up. Four patients in whom mammographic findings were suggestive and whose CNB results were category 3 did not undergo surgery, off protocol, because the extent and cytologic features showing atypical ductal hyperplasia were to be considered minimal in two patients; the other two patients declined surgery.
The sensitivity, specificity, and accuracy of CNB were determined with results from surgical biopsy and clinical and/or mammographic follow-up to determine true-positive, false-negative, and false-positive findings. True-negative findings were determined with concordance among BI-RADS 1, 2, and 3 categories with either surgical validation or no change during clinical and mammographic follow-up. By using these values, the positive predictive value (PPV) and negative predictive value (NPV) were determined (Table 2). Results were as follows: strict sensitivity, 91% ± 1.9 (SD); working sensitivity, 92% ± 1.8; applied sensitivity, 98% ± 0.9; strict specificity, 100%; working specificity, 98% ± 0.8, applied specificity, 73% ± 0.9; strict accuracy, 97%; working accuracy, 96%; applied accuracy, 79%; strict PPV and NPV, 100% and 97%, respectively; working PPV and NPV, 93% and 97%, respectively; applied PPV and NPV, 56% and 99%, respectively.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX REFERENCES |
|---|
The validity of the use of imaging findings and associated CNB results to identify cases suitable for surveillance or further surgery was further assessed by using a median follow-up of 24 months, which we believe is reasonable for several reasons and provides a basis for a sensitivity analysis. First, with deliberate and systematic attempts to obtain follow-up data, success rates have ranged from 54%–99% for single institutions using this procedure and reporting their results (21,22). Second, the geographic relocation of patients, many of whom were referred only for CNB to one of the participating institutions soon after the introduction of this technique into clinical practice, makes it increasingly difficult to obtain longer follow-up, especially for lesions identified as benign. Third, in previously published studies (23–27,36–41) in which cancers detected during dedicated clinical and mammographic surveillance were identified, almost all interval cancers were diagnosed within 12–24 months. Fourth, in the two published studies with patients undergoing CNB with longer duration follow-up of most patients, the interval "false-negative rate" was reported to be between 1.2% and 2%, which is commensurate with our results of 1.6%–2.4%, as well as with those who have reported on lesions considered to be probably benign by using mammographic criteria alone (23–27,36–40). In other words, an expected false-negative rate for probably benign lesions according to mammographic criteria alone or to CNB when mammographic lesions initially considered to be suspicious were reconciled with benign CNB histopathologic results may reasonably be expected to range from 1%–2%. In this study, four (1%) of the 388 patients with mammographic and CNB results consistent with a benign lesion were found to have malignancy at the site of core biopsy within an 18-month follow-up. If a similar or even higher percentage of patients who were lost to follow-up (n = 322) demonstrated malignancy, this would mean an additional four patients (3.3 [1%] patients) with interval malignancy. This would lower the strict sensitivity from 91% to 90%, but the applied sensitivity would remain unchanged at 98%. This potential increase in interval malignancy rates remains one of the limitations of this study.
Our results may actually underestimate the eventual accuracy of CNB in clinical practice, because these results reflect CNB limited to five samples obtained with 14-gauge needles and spring-loaded biopsy devices, which reflects a second limitation of this study. As experience was gained during the study, additional core biopsies were frequently obtained for cases of calcifications, but our data are presented in the context of five CNB samples for each type of lesion, to provide a comparison of accuracies for different types of lesions. Prior studies (42,43) of palpable masses that have shown better results for fine-needle aspiration biopsy than for CNB have been compromised by the paucity of samples obtained, with unsatisfactory results found after one or three samples were obtained for palpable lesions. In the present study, five cases of infiltrating cancer and nine cases of DCIS (90% of which manifested as clustered calcifications) would have been overlooked had only five CNB samples been obtained initially, which corroborates the experience of other investigators (8,44–47). The finding of DCIS at excisional biopsy in four of 18 cases where the CNB diagnosis was atypical ductal hyperplasia, which increased the sensitivity by 1%, emphasizes the potential tissue-sampling error and the need for more extensive histopathologic material in these cases (26,27). Of particular importance was the reconciliation of mammographic findings with CNB histopathologic results. The decision to obtain more tissue by means of CNB (off protocol) or subsequent excisional biopsy because of high mammographic suspicion (discordant mammographic findings) affected the correct diagnosis in 15 (6%) of 254 cases. The increasing sensitivity of this method (from 91% to 92% to 98%) was achieved with a concomitant decrease in specificity (from 100% to 98% to 73%), which was considered both acceptable and advisable.
The rate of correct upstaging of the diagnosis to invasive cancer as determined at surgical excision because of a finding of DCIS at CNB in this study was 9%, similar to an 8% result reported previously (46). These findings, as well as potential underestimation of malignancy because of a finding of atypical ductal hyperplasia at CNB, are important to consider when needle biopsy technologies are proposed for complete excision of small lesions (8,45). Although CNB understaging of both DCIS and invasive cancer has been reported (45) to be mitigated by using larger biopsy needles and obtaining more biopsy samples, reliability is still not sufficient to obviate final surgical diagnoses in most cases.
A third limitation of this study may be related to the study period, which coincided with the early use of CNB. It is difficult to reconcile why 52 (5%) patients with benign mammographic features underwent CNB. Some patients or referring physicians may seek a tissue diagnosis even when there is high confidence regarding the likely benign appearance of a nonpalpable lesion. Whether or not this procedure was seen as a compromise between an imaging diagnosis and a surgical tissue diagnosis is difficult to determine, but the number of patients with benign mammographic lesions may be associated with an artificially improved accuracy. The same argument applies to patients undergoing core biopsy for "probably benign" mammographic lesions, where analysis has shown a potential negative cost impact with similar outcome as compared with patients undergoing surveillance (48). More important, a relatively high percentage of cases diagnosed as benign on the basis of CNB findings went on to excision because of discordant mammographic findings. Although this approach increased the sensitivity by 6%, the large number of cases may reflect early lack of confidence in the CNB procedure. As mentioned earlier, increased experience and more efficient methods of tissue retrieval at stereotactic biopsy are likely to decrease the anticipated discordance reported in this study.
We reported the sensitivity, specificity, and accuracy with assigned variations, to permit the radiologist to more accurately determine the deliberate use of both CNB and mammographic results in final management determinations. Although the use of CNB results is an accurate method for establishing a histopathologic diagnosis of a nonpalpable lesion identified at screening mammography (strict sensitivity), certain CNB diagnoses are associated with such a high rate of underestimating malignancy that excisional biopsy is indicated (working sensitivity), which increased the sensitivity by 1% in our study. This definition most closely approximates the manner in which results have been previously reported (1–15) and, in our study, would correspond to a PPV of 93% and an NPV of 97%. Where CNB results were otherwise benign but the mammographic findings were sufficiently suggestive (discordant mammographic findings), additional tissue sampling increased the sensitivity by 6% (applied sensitivity). For example, when a cluster of microcalcifications is adequately sampled and shown by CNB to represent evidence of a hyalinizing fibroadenoma, the otherwise "suspicious" nature of the mammographic finding can be reconciled. Conversely, a spiculated mass that is highly suggestive of malignancy cannot be reconciled with the finding of benign nonspecific fibrotic changes and would warrant additional surgery. The applied specificity may be artificially lower than it should be because 10 patients with benign mammographic findings and CNB results underwent surgery secondary to the preference of the patient, physician, or both.
Ultrasonographically (US) guided biopsies were not included in this protocol. While many operators prefer US guidance for CNB of masses, others favor stereotactic guidance. Stereotactic guidance for CNB of masses is included as a requirement for voluntary accreditation by the American College of Radiology. The U.S. Food and Drug Administration is also considering stereotactic biopsy certification. Emerging technologies, including larger needles and multidirectional vacuum-assisted biopsy devices, will likely help improve the already encouraging results reported here. However, this protocol was developed prior to the introduction of improved biopsy devices, to achieve sufficient follow-up from multiple institutions participating in a prospective trial. An appreciation of both the limitations and potentials of CNB diagnosis is essential for proper management. By recognizing the effect of tissue-sampling error and sufficient concordance with mammographic findings, high accuracy by using CNB is achievable.
|
APPENDIX |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX REFERENCES |
|---|
Strict Sensitivity, Specificity, and Accuracy
The strict measures reflect any discordance between CNB histopathologic and surgical results. False-positive results were determined where CNB showed malignancy (category 4 or 5) but surgery showed no malignancy. Strict sensitivity was determined with the following: TP/(TP + FN), where TP and FN are the number of true-positive and false-negative results, respectively. Specifically, CNB4,5/(Surgery4,5 + interval malignancies) = 230/(248 + 6) = 91% ± 1.9, where the subscripted numbers indicated the CNB and surgical categories. Strict specificity was determined with the following: TN/(TN + FP), where TN and FP are the number of true-negative and false-positive results, respectively. Specifically, CNB1,2,3/(Surgery1,2,3 + CNB4,5) = 734/734 = 100%. Strict accuracy was determined with the following: TP + TN/(TP + FN + TN + FP). Specifically, 964/992 = 97%.
Working Sensitivity, Specificity, and Accuracy
CNB category 3 is included as true-positive because such a diagnosis is intended to prompt wider surgical resection; thereafter, it would be considered to be false-positive if surgery showed no malignancy. As a true-positive finding, these results approximate those previously reported (1–15) but differ in that prior reports do not indicate the FP result when surgery demonstrates benign results.
Working sensitivity was determined with the following:TP/(TP + FN). Specifically, CNB3,4,5/(Surgery4,5 + interval malignancies) = 230 + 4/(248 + 6) = 234/254 = 92% ± 1.8. Working specificity was determined with the following: TN/(TN + FP). Specifically, CNB1,2/(Surgery1,2,3 + CNB4,5) = 717/734 = 98% ± 0.8. Working accuracy was determined with the following: TP + TN/(TP + FN + TN + FP). Specifically, 947/992 = 96%.
Applied Sensitivity, Specificity, and Accuracy
CNB category 3 is included as true-positive, as for working sensitivity. In addition, BI-RADS category 4 or 5 lesions are included as true-positive when any negative CNB histopathologic result cannot be reconciled with suggestive mammographic findings. For example, if clustered calcifications (BI-RADS category 4 or 5) show histopathologic features that indicate hyalinizing fibroadenoma, then the CNB diagnosis is considered to be reliable. If CNB results for the same mammographic feature show only nonspecific fibrocystic changes, the CNB diagnosis is not considered to be reliable, and further surgery may be prompted on the basis of the mammographic features. Thus the true-positive rate is increased by the additional cases where surgery is performed because of suggestive mammographic findings that could not be reconciled with a benign CNB result (discordant mammographic findings). Where surgery shows no malignancy, such cases are then considered to be false-positive.
Applied sensitivity was determined with the following: TP/(TP + FN). Specifically, (CNB3,4,5 + DMF)/(Surgery4,5 + interval malignancies) = (230 + 15 + 4)/(248 + 6) = 249/254 = 98% ± 0.9, where DMF is the number of discordant mammographic findings. Applied specificity was determined with the following: TN/(TN + FP). Specifically, CNB1,2/(Surgery1,2,3 + CNB3 + DMF) = 537/734 = 73% ± 0.9. Applied accuracy was determined with the following: TP + TN/(TP + FN + TN + FP). Specifically, 986/992 = 79%.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, R.J.B.; study concepts, R.J.B.; study design, R.J.B., L.W.B., L.L.F., D.D.D., W.P.E., P.F., V.P.J., S.A.F., F.R.M.; definition of intellectual content, R.J.B., L.W.B., L.L.F., D.D.D., W.P.E., P.F., V.P.J., S.A.F., E.B.M., F.R.M.; literature research, R.J.B.; clinical studies, L.W.B., L.L.F., D.D.D., W.P.E., P.F., V.J.J., S.A.F., E.B.M., F.R.M., C.L.; data acquisition, R.J.B., L.W.B., L.L.F., D.D.D., W.P.E., R.H., C.L., I.T., P.F., M.M., V.P.J., S.A.F., E.B.M., F.R.M.; data analysis, R.J.B., L.W.B., D.D.D., J.S.; statistical analysis, R.J.B., J.S.; manuscript preparation, R.J.B.; manuscript editing, R.J.B., L.W.B., D.D.D., V.J.P., R.B.; manuscript review, all authors.
Author affiliations: Departments of Radiology (L.W.B., M.M.C.) and Biostatistics (J.S.), University of California at Los Angeles School of Medicine; Department of Radiology, Johns Hopkins Outpatient Center, Baltimore, Md (L.L.F.); Department of Radiology, Memorial Sloan-Kettering Cancer Center, New York, NY (D.D.D.); Susan G. Komen Alliance Breast Center, Baylor University Medical Center, Dallas, Tex (W.P.E.); Health Sciences Center, University of Arizona, Tucson, Ariz (R.H.); Department of Radiology, Yale University School of Medicine, New Haven, Conn (C.L., I.T.); Department of Radiology, Allegheny University of the Health Sciences, Philadelphia, Pa (P.F.); Department of Radiology, Indiana University Medical Center, Indianapolis (V.P.J.); Department of Radiology, Mount Sinai Medical Center, New York, NY (S.A.F.); Department of Radiology, Western Pennsylvania Hospital, Pittsburgh, Pa (E.B.M.); Department of Radiology, California Pacific Medical Center, San Francisco, Calif (F.R.M.); and Lorad, Danbury, Conn (R.B.).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX REFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. Calhoun, A. Giuliano, and R. J. Brenner Intraoperative Loss of Core Biopsy Clips: Clinical Implications Am. J. Roentgenol., March 1, 2008; 190(3): W196 - W200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. H. Youk, E.-K. Kim, M. J. Kim, J. Y. Lee, and K. K. Oh Missed Breast Cancers at US-guided Core Needle Biopsy: How to Reduce Them RadioGraphics, January 1, 2007; 27(1): 79 - 94. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. K. Koskela, M. Sudah, M. H. Berg, V. J. Karja, P. K. Mustonen, V. Kataja, and R. S. Vanninen Add-on Device for Stereotactic Core-Needle Breast Biopsy: How Many Biopsy Specimens Are Needed for a Reliable Diagnosis? Radiology, September 1, 2005; 236(3): 801 - 809. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Kirshenbaum, T. Voruganti, C. Overbeeke, M. D. Kirshenbaum, P. Patel, G. Kaplan, V. Maker, C. August, and R. P. Cavallino Stereotactic Core Needle Biopsy of Nonpalpable Breast Lesions Using a Conventional Mammography Unit with an Add-On Device Am. J. Roentgenol., August 1, 2003; 181(2): 527 - 531. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. C. Gombos, R. J. Poppiti Jr., and R. J. Brenner Percutaneous Core Needle Biopsy of Radial Scars of the Breast Am. J. Roentgenol., July 1, 2003; 181(1): 275 - 275. [Full Text] [PDF]
|
|
|
|
|
|
S. Huber, M. Wagner, M. Medl, and H. Czembirek Benign Breast Lesions: Minimally Invasive Vacuum-assisted Biopsy with 11-Gauge Needles—Patient Acceptance and Effect on Follow-up Imaging Findings Radiology, March 1, 2003; 226(3): 783 - 790. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Woo, T. Yu, and T. C. Hurd Breast Cancer in Pregnancy: A Literature Review Arch Surg, January 1, 2003; 138(1): 91 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. D. Friedman, L. M. Sanders, C. Menendez, L. Kalisher, and G. Petrillo Retrieval of Lost Microcalcifications During Stereotactic Vacuum-Assisted Core Biopsy Am. J. Roentgenol., January 1, 2003; 180(1): 275 - 280. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Brenner, R. J. Jackman, S. H. Parker, W. P. Evans III, L. Philpotts, B. M. Deutch, M. C. Lechner, D. Lehrer, P. Sylvan, R. Hunt, et al. Percutaneous Core Needle Biopsy of Radial Scars of the Breast: When Is Excision Necessary? Am. J. Roentgenol., November 1, 2002; 179(5): 1179 - 1184. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. R. Margolin, J. W. T. Leung, R. P. Jacobs, and S. R. Denny Percutaneous Imaging-Guided Core Breast Biopsy: 5 Years' Experience in a Community Hospital Am. J. Roentgenol., September 1, 2001; 177(3): 559 - 564. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Rasuli Breast cancer diagnosis: What are we waiting for? Can. Med. Assoc. J., August 1, 2001; 165(3): 303 - 304. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
