| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, George Washington University, 2150 Pennsylvania Ave.,
NW, Washington, DC 20037.
2 Department of Radiology, Park Nicollet Clinic–St. Louis Park, St. Louis
Park, MN.
3 Department of Radiology, Stanford Hospital and Clinics, Stanford, CA.
4 Department of Radiology, The University of Texas Southwestern Medical Center
at Dallas, Dallas, TX.
5 Department of Diagnostic Radiology, Yale University School of Medicine, New
Haven, CT.
Received June 22, 2007;
accepted after revision September 27, 2007.
Address correspondence to R. F. Brem
(rbrem{at}mfa.gwu.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. The records of 32,420 patients who underwent imaging-guided needle biopsy of the breast for mammographic or sonographic abnormalities from 1988 to 2000 were retrospectively reviewed. The 278 cases in which lobular neoplasia was the highest-risk lesion at biopsy were included. Of the 278 cases, 164 proceeded to surgical excision, allowing calculation of rates of underestimation from minimally invasive biopsy.
RESULTS. Of the 32,420 minimally invasive breast biopsies, lobular neoplasia was found in 278 (0.9%). One hundred sixty-four of the 278 (59%) continued to surgical excision, where cancer was pathologically confirmed in 38 (23%). No difference was seen in the underestimation rates for lesions diagnosed as lobular carcinoma in situ (25%, 17 of 67 lesions) versus atypical lobular hyperplasia (22%, 21 of 97 lesions). Statistically significant underestimation of carcinoma was found with biopsy of masses (with or without associated microcalcifications) rather than calcifications only, a higher BI-RADS category (p < 0.0001), use of a core biopsy device rather than a vacuum device (p < 0.01), and obtaining fewer specimens (p < 0.0001).
CONCLUSION. Significant sampling error occurs regardless of the type of core biopsy device, number of specimens obtained, histologic–radiographic concordance, mammographic appearance, and complete excision of the lesion as determined by imaging. For this reason, all patients with lobular neoplasia at core or vacuum-assisted biopsy should undergo surgical excision until further differentiating criteria can be determined.
Keywords: breast imaging • biopsy • lobular neoplasia • mammography
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Lobular neoplasia is a group of pathologic entities composed of monomorphic, regularly spaced polygonal cells with a high nuclear–cytoplasmic ratio. The spectrum of lobular neoplasia ranges from atypical lobular hyper-plasia (ALH), in which few acinar units are involved, to lobular carcinoma in situ (LCIS), which fills and distends the entire lobule [10]. Lobular neoplasia is often diffuse and bilateral. Women with lobular neoplasia are at a higher overall risk for developing breast cancer than the general population [11].
The significance of the incidental finding of lobular neoplasia on core biopsy is disputed. Lobular neoplasia is the highest-risk lesion in less than 2% of all core biopsies [12, 13], making it a difficult entity to study. Single-institution series have found that the diagnosis of lobular neoplasia in these cases is an underestimate anywhere from 0–50% of the time [14]. On the basis of the high potential for underestimation, some centers recommend excisional biopsy for all patients who have lobular neoplasia as the highest-risk lesion. Others advocate excisional biopsy only in certain situations, such as when the core biopsy histopathology does not correlate with the mammographic findings or the histopathology is ambiguous. Others maintain that surgical excision is not necessary because lobular neoplasia is thought to be an incidental finding.
This retrospective multiinstitutional study was designed to better define the rate at which LCIS and ALH are found to be carcinoma at surgical excision, and to determine which patient, lesion, biopsy, and histologic variables are associated with underestimation of disease.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In each case, patients were stratified by age at the time of biopsy (< 51, 51–60, or > 60 years). Lesions were stratified by imaging lesion type (calcifications only, mass only, or calcified mass), size at imaging (< 5 m, 5–10 mm, or > 10 mm), and BI-RADS classification (categories 3, 4, or 5). Biopsy methods were stratified by guidance method (stereo tactic or sonographic), biopsy device (automated large-core [core] or directional vacuum-assisted [vacuum], including the gauge of the biopsy device), number of specimens obtained (< 11, 11–20, or > 20 specimens), postbiopsy mammo graphy findings (completely or partially excised), and specimen radiography (calci fications seen or not seen). Histologic variables were stratified by percutaneous histology (LCIS or ALH) and imaging–histologic concordance (con cordant or discordant).
Pathologists at each institution independently performed tissue diagnoses from core and surgical biopsies. The histologic findings at percutaneous biopsy were reported as LCIS or ALH. In lesions that contained both, LCIS was used in the analysis because it is the higher-grade lesion. The pathology at surgical excision was compared with that obtained at percutaneous biopsy to determine if there was an upgrade to invasive carcinoma or DCIS.
The method of defining imaging–histologic concordance varied between institutions. For some institutions, a finding of lobular neoplasia with no other lesion to explain the mammographic abnormality was considered automatically discordant, whereas for others it was not. This study was approved by the institutional review board and was HIPAA compliant.
Statistical Analysis
The data were entered into a computerized spreadsheet (Microsoft Excel,
2003) for analysis. Tests for statistical significance were performed using
chi-square and Fisher's exact tests. A p value of less than 0.05 was
considered statistically significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Disease underestimates were found in 38 (23%) of 164 surgically excised cases, with underestimates from 17 (25%) of 67 LCIS lesions and 21 (22%) of 97 ALH lesions. Of the 17 LCIS underestimates, seven (41%) were DCIS, six (35%) were invasive lobular carcinoma, two (12%) were invasive ductal carcinoma, and two (12%) were invasive carcinomas, unspecified. Four of the underestimates were masses, 12 were calcifications, and one was a calcified mass. Of the 21 ALH underestimates, 15 (71%) were DCIS, three (14%) were invasive lobular carcinomas, and three (14%) were invasive ductal carcinomas. The ALH upgrades included five masses, 13 calcifications, and three calcified masses.
Underestimation rates for the 164 lobular neoplasia lesions having surgical excision are shown in Table 1 for patient, lesion, biopsy, and histologic variables. Statistically significant underestimation of the presence of carcinoma was found with biopsy of masses (with or without associated calcifications) rather than calcifications only, a higher BI-RADS category, use of a core biopsy device rather than a vacuum device, and obtaining fewer specimens. The number of specimens collected ranged from one to 31 (median, 12; mean, 12.4 specimens). The average number of specimens collected in cases found to be underestimates was 10, compared with 13.5 specimens for those that did not represent underestimates, a statistically significant difference (p = 0.0027).
|
Underestimation rates were not statistically significant for patient age, biopsy guidance method, completeness of excision, specimen radiographs for calcifications, LCIS or ALH histology, or concordance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Underestimation with the 14-gauge automated core needle was greater than with 14- or 11-gauge vacuum devices, showing that sampling error was inversely related to the amount of tissue removed [15, 16]. Removal of a greater number of specimens in this study was also associated with less disease underestimation. The highest underestimation was observed when fewer than 10 specimens were retrieved at core biopsy (40%). Removing more than 10 specimens had an improved but still significant amount of underestimation (15%). This is consistent with previous studies that have shown increased diagnostic accuracy with increased numbers of specimens [17].
Mammographically determined incomplete lesion excision was associated with more underestimates than complete excision was (25% vs 15%). However, excision of the entire lesion as determined mammographically did not ensure absence of cancer at surgical excision.
A lower underestimation rate occurred in mass lesions (18%) than in lesions with calcifications (24%). Thus, sampling error occurs whether the targeted lesion is a mass or calcifications.
The design of this study had no consistent criteria to determine radiologic–pathologic concordance. This variance is a limitation of this study and is likely responsible for the higher rate of discordance in the study as compared with those previously reported [18]. However, in this study, as previously reported, discordance was associated with a higher rate of underestimation (27% vs 17% for concordant cases).
The overall 23% rate of underestimation found in our multiinstitutional study, which is the largest report to date, emphasizes the findings of single-institution studies, with underestimation rates ranging from 14% to 27% [13, 14, 19–35].
The lobular neoplasia underestimation rate in this study is comparable to that found with ADH diagnosed on minimally invasive breast biopsy, when subsequent surgical excision is the standard of care [8]. So too, when lobular neoplasia is found on minimally invasive biopsy, all patients should proceed to surgical excisional biopsy to exclude the presence of cancer.
A major limitation of this study is the variability between institutions. There was no retrospective pathology overview, resulting in a potential for variability in the interpretation of lobular neoplasia [36]. However, because the underestimation rates for ALH and LCIS were similar, this likely did not affect our findings.
Note that all patients, regardless of the mammographic appearance of their lesion or the biopsy method, had underestimation of cancer. Therefore, until further research can stratify patients with lobular neoplasia at minimally invasive breast biopsy, all patients with this finding should proceed to surgical excisional biopsy.
In conclusion, we found lobular neoplasia in 278 (0.9%) of 32,420 imaging-guided needle biopsies. Of the 164 (59%) of those lesions having surgical excision, the presence of carcinoma was found in 38 (23%). No difference was seen in the underestimation rates for lesions diagnosed as LCIS (25%, 17 of 67 lesions) versus ALH (22%, 21 of 97 lesions). The 38 carcinomas found at surgical excision were DCIS in 22 (58%) and invasive carcinoma in 16 (42%). Statistically fewer underestimates were found with increased tissue sampling (including complete excisions, > 20 specimens removed, and diagnostic vacuum-assisted biopsy device use) and when a lower BI-RADS score was assigned, but we could identify no subgroup of lobular neoplasia lesions that does not require subsequent surgical excision.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. J. Jackman, F. A. Marzoni Jr., and J. Rosenberg False-Negative Diagnoses at Stereotactic Vacuum-Assisted Needle Breast Biopsy: Long-Term Follow-Up of 1,280 Lesions and Review of the Literature Am. J. Roentgenol., February 1, 2009; 192(2): 341 - 351. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Pediconi, C. Catalano, S. Padula, A. Roselli, V. Dominelli, S. Cagioli, M. A. Kirchin, G. Pirovano, and R. Passariello Contrast-Enhanced MR Mammography: Improved Lesion Detection and Differentiation with Gadobenate Dimeglumine Am. J. Roentgenol., November 1, 2008; 191(5): 1339 - 1346. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. Kopans LCIS Found at Core Needle Biopsy May Not Need Surgical Excision Am. J. Roentgenol., September 1, 2008; 191(3): W152 - W152. [Full Text] [PDF]
|
|
|
|
|
|
R. F. Brem, R. J. Jackman, and M. C. Lechner Reply Am. J. Roentgenol., September 1, 2008; 191(3): W153 - W153. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
