Percutaneous Core Needle Biopsy of Radial Scars of the Breast: When Is Excision Necessary?

R. James Brenner 1, 2 Roger J. Jackman 3 Steve H. Parker 4 W. Phil Evans III 5 Liane Philpotts 6 Beth M. Deutch 7 Mary C. Lechner 8 Daniel Lehrer 9 Paul Sylvan 10 Rebecca Hunt 11 Steven J. Adler 12 Nancy Forcier 13 Received June 4, 2001; accepted after revision April 9, 2002. 1 Tower St. John s Imaging, Joyce Eisenberg Keefer Breast Center, John Wayne Cancer Institute, St. John s Hospital and Health Center, 1328 22nd St., Santa Monica, CA 90404. Address correspondence to R. J. Brenner. 2 Department of Radiology, UCLA School of Medicine, 200 UCLA Medical Plaza, Rm. 165-47, Los Angeles, CA 90095-1721. 3 Department of Diagnostic Radiology, Palo Alto Medical Clinic, 795 El Camino Real, Palo Alto, CA 94301. 4 Sally Jobe Breast Center, Radiology Imaging Associates, 1001 E. Layton Ave., Englewood, CO 90110-7017. 5 Susan G. Komen Breast Center, 3535 Worth St., Ste. 340, Dallas, TX 75246. 6 Department of Radiology, Yale University, 333 Cedar St., New Haven, CT 06520-8042. 7 Jacqueline M. Wilentz Comprehensive Breast Center, Monmouth Medical Center, 300 Second Ave., Long Branch, NJ 07740. 8 Department of Radiology, Jane Brattain Breast Center, Park Nicollet Medical Center, 3900 Park Nicollet Blvd., St. Louis Park, MN 55416. 9 CERIM, Instutucion De Avanzada, Pionera en el Diagnstico, De Las Enfermedades Mamarias, Azcuenaga 970, (1115) Buenos Aires, Argentina. 10 La Jolla Radiology, Ste. B, 7301 Girard Ave., La Jolla, CA 92037. 11 Health Science Center-Breast Division, University of Arizona Medical Center, 1501 N. Campbell Ave., Tucson, AZ 85724. 12 Mammography Section, Virginia Mason Medical Center, 1100 Ninth Ave. C5-XR, Seattle, WA 98111. 13 Mission Breast Center, c/o 456 28241 Crown Valley Pkwy., Ste. F, Laguna Niguel, CA 92677. AJR 2002;179:1179 1184 0361 803X/02/1795 1179 American Roentgen Ray Society Percutaneous Core Needle Biopsy of Radial Scars of the Breast: When Is Excision Necessary? OBJECTIVE. This study was conducted to evaluate the outcome of cases of radial scar diagnosed by percutaneous core needle biopsy. MATERIALS AND METHODS. Of 198 nonpalpable lesions diagnosed with radial scars found at core needle biopsy, 157 lesions constituting the study group had undergone surgical excision (n = 102) or mammographic surveillance after biopsy for at least 24 months (median, 38 months; n = 55). Mammographic lesion type, lesion size, biopsy guidance method, biopsy device, number of specimens per lesion, and presence of atypical hyperplasia at percutaneous biopsy were retrospectively analyzed. Results were compared with histologic findings at surgery or mammographic findings during surveillance. RESULTS. Carcinoma was found at excision in 28% (8/29) of lesions with associated atypical hyperplasia at percutaneous biopsy and 4% (5/128) of lesions without associated atypia (p < 0.0001). In the latter group, carcinoma was found at excision in 3% (2/60) of masses, 8% (3/40) of architectural distortions, and 0% (0/28) of microcalcification lesions. Malignancy was missed in 9% (5/58) of lesions biopsied with a spring-loaded device and in 0% (0/70) of lesions biopsied with a directional vacuum-assisted device (p = 0.01); and in 8% (5/60) of lesions sampled with less than 12 specimens per lesion and 0% (0/68) sampled with 12 or more specimens (p = 0.015). Lesion type, maximal lesion diameter, and type of imaging guidance (stereotactic or sonographic) were not significant factors in determining the presence of malignancy. CONCLUSION. Diagnosis of radial scar based on core needle biopsy is likely to be reliable when there is no associated atypical hyperplasia at percutaneous biopsy, when the biopsy includes at least 12 specimens, and when mammographic findings are reconciled with histologic findings. When the lesion diagnosed by core needle biopsy as radial scar does not meet these criteria, excisional biopsy is indicated. R adial scars of the breast have been described by various names including radial sclerosing lesion, scleroelastotic lesion, indurative mastopathy, nonencapsulated sclerosing lesion, sclerosing papillary proliferation, and, if larger than 1 cm, complex sclerosing lesion [1 6]. The term radial scar was introduced by Hampert s translation of Strahlige Narben in 1975 [7]. Characterized microscopically by a central fibroelastotic core containing entrapped glandular elements, radial scars contain ducts that radiate outward, giving the lesion its characteristic stellate appearance. Because of the entrapped glands at the center of the lesion, radial scar has sometimes been mistaken for tubular carcinoma [6, 7]; the former can be differentiated by a surrounding layer of myoepithelial cells that sometimes are difficult to detect with routine H and E staining. In addition to the morphologic similarity of radial scar to tubular carcinoma, the coincident presence of hyperplasia, adenosis, and papillomatosis has led some investigators to speculate that radial scars represent a stage in the development of invasive carcinoma [8]. Although most radial scars have been reported as incidental microscopic lesions seen at biopsy for another lesion or at autopsy [9], the increased use of mammography has identified features sometimes considered typical of this lesion [10]. However, rigorous analysis of this criterion does not support a basis for sufficiently distinguishing radial scar from carcinoma by mammographic features, so biopsy of suspected radial scars is still considered mandatory [1, 11 14]. Because of the reported association of radial scar with surrounding proliferative disease and malignancy, some authors have advocated excisional biopsy in all cases [12, 15, 16]. The consequence of this recommendation is to require excisional biopsy to follow any core needle biopsy that re- AJR:179, November 2002 1179

Brenner et al. sults in the diagnosis of radial scar. This rationale has been validated in the case of atypical ductal hyperplasia diagnosed by core needle biopsy in which the subsequent excisional biopsy changes the final diagnosis to intraductal or invasive carcinoma in 18 54% of cases [17 19] so that surgery is indicated. However, to our knowledge, the recommendation for performing excisional biopsy for all cases of radial scar diagnosed at core needle biopsy has not been similarly justified, perhaps because so few cases are studied in any individual practice. We therefore conducted a retrospective review of records from 11 institutions to determine the conditions under which the diagnosis of radial scar by core needle biopsy is reliable. Materials and Methods From January 1990 through December 2000, records from 11 institutions in both academic and private settings were reviewed to determine which patients had undergone imaging-guided breast core needle biopsy for nonpalpable mammographically detected lesions that resulted in a histologic diagnosis of radial scar. Lesions with associated atypical hyperplasia at percutaneous biopsy were included in the study. Lesions with associated malignancy (invasive carcinoma or ductal carcinoma in situ [intraductal carcinoma]) at percutaneous biopsy were excluded from the study. Age of patient, lesion characteristics, and procedural variables were recorded. Imaging guidance was by a stereotactic or sonographic technique, each using a variety of springloaded devices or a directional vacuum-assisted device (Mammotome; Ethicon Endo-Surgery, Cincinnati, OH). Needle size and number of specimens obtained during each biopsy were also recorded. Each operator at his or her respective institution decided which approach stereotactic or sonographic would be used, as well as the size of needle, biopsy device, and number of samples obtained when biopsying the lesions. Lesions were characterized as masses, architectural distortions, or clustered calcifications with combinations of characteristics noted on data entry sheets. As a multiinstitutional retrospective study, the methodology introduced a degree of variability, although all operators were experienced with both procedures and Breast Imaging Reporting and Data System (BI-RADS) terminology [20]. Informed consent was obtained from the 192 patients having a total of 198 lesions (with three lesions in one patient and two lesions in each of four other patients). After core needle biopsy was performed, patients underwent either excisional biopsy or clinical and mammographic follow-up, with surgical results compared with core needle biopsy results. To date, 157 lesions in 152 patients (age range, 26 79 years; median age, 51 years) have undergone surgical excision of the percutaneous biopsy site (n = 102) or mammographic follow-up for at least 24 months after biopsy was performed (n = 55), and these patients constituted the study group. The remaining 41 lesions that did not qualify for the study had mammographic follow-up for less than 24 months (range, 3 21 months; median, 12 months) (n = 22) or no follow-up to date (n = 19). The decision to excise the percutaneous biopsy site was predicated on the findings at core needle biopsy (particularly the presence of atypical hyperplasia), on individual operator preference based on prior reported studies or lack of confidence in reconciling mammographic findings with those of core needle biopsy (or both), or on progression of the lesion on mammography after percutaneous biopsy [17 19]. Statistical analysis was performed using chisquare methods; p values of less than 0.05 were considered significant. Results In data analyses of the 157 lesions, certain mammographic characteristics of lesions and certain procedural variables were combined. Lesions were considered to be masses (n = 73) whether they were associated with calcifications; focal architectural distortion (n = 47), whether they were associated with calcifications; or clustered microcalcifications (n = 37). Maximal mammographic lesion diameter, measured from one of two standard mammographic views, ranged from 1 to 41 mm (median, 10 mm; mean, 13 mm). The size of the lesion was reported as mammographically measured because after core biopsy, histologic size, as determined at subsequent surgery, is usually not reported and not reported, by definition, for lesions in patients subject to follow-up. Imaging guidance was stereotactic (n = 129) (whether the patient was in a prone position on a dedicated table [n = 128] or upright [n = 1]) or by sonography (n = 28). Biopsy devices were spring-loaded (n = 65) (whether 14-gauge [n = 63] or 12-gauge [n = 2]) or directional vacuum-assisted (n = 92) (whether 14-gauge [n = 6] or 11-gauge [n = 86]). By guidance system and biopsy device, the number of specimens obtained per lesion was between three and 20 (median, n = 7) for 51 stereotactic biopsies with a springloaded device; between four and 38 (median, n = 14) for 82 stereotactic biopsies with directional vacuum-assisted device; between two and 14 (median, n = 5) for 16 sonographic biopsies with a spring-loaded device; and between five and 24 (median, n = 15) for eight sonographic biopsies with a directional vacuum-assisted device. Of 157 study lesions, 102 cases diagnosed as radial scar were excised. Ten of these cases underwent surgery on the basis of a single institution s modification in management recommendations after an interval change in one case that prompted excision and the diagnosis of malignancy. In another eight cases, surgical biopsy was prompted by apparent interval mammographic change or change in patient or physician preference. Mammographic follow-up in the 55 cases that did not undergo excision was performed for 24 92 months (median, 38 months; mean, 45 months) after percutaneous biopsy, with none showing increased density or progression at the site of the lesion. At surgical excision, carcinoma was histopathologically detected in 13 (8%) of 157 cases. Malignancies were ductal carcinoma in situ (n = 8), invasive ductal carcinoma (n = 3), and a combination of invasive and intraductal carcinoma (n = 2). Carcinomas were found in 28% (8/29) of lesions with associated atypia at percutaneous biopsy, compared with 4% (5/ 128) of lesions without associated atypia, a statistically significant difference (p < 0.0001). The 29 radial scar lesions with associated atypia presented as masses in 13 lesions (45%), areas of architectural distortion in seven lesions (24%), or calcifications in nine lesions (31%). Malignancies with associated atypical hyperplasia were found in four (31%) of 13 masses, two (22%) of nine clusters of microcalcifications, and two (28%) of seven foci of architectural distortion and biopsied with either the directional vacuum-assisted device (n = 7) or the spring-loaded device (n = 1). The types of atypical tissue (n = 29) associated with radial scar at percutaneous biopsy were atypical ductal hyperplasia (n = 19); atypical lobular hyperplasia (n = 2); lobular carcinoma in situ (n = 1); atypical ductal hyperplasia and lobular carcinoma in situ (n = 2); and atypical tissue, not specified (n = 5). All carcinomas were associated with atypical ductal hyperplasia at core needle biopsy except for one case of calcifications associated with only atypical lobular hyperplasia. Of these cases, stereotactic biopsy was performed using an 11-gauge needle in seven cases (number of specimens obtained per lesion, respectively, was 10, 12, 15, 18, 18, and 21), and sonography was performed using a 12-gauge spring-loaded device in one case (with six specimens obtained). Two carcinomas (13%) were found in lesions with a maximal diameter of 2 10 mm and two (13%) with a maximal diameter of 11 40 mm. The 128 radial scar lesions in our study without associated atypia presented as masses in 60 lesions (47%), foci of architectural distortion in 40 (31%), or calcifications in 28 (22%). Of the five malignancies found histopathologically at surgical excision, carcinomas were found in 3% (2/60) of masses, 8% (3/40) of architectural distortions, 0% (0/28) of clustered microcalcifications, 3% (2/61) of lesions with a maximal diameter of 2 10 mm, and 5% 1180 AJR:179, November 2002

Core Needle Biopsy of Radial Breast Scars (3/64) of lesions with a maximum diameter of 11 40 mm (diameter was not recorded for three lesions). Of five carcinomas among the 128 cases without atypia, four (4%) of 104 carcinomas were missed with stereotactic guided biopsy and one (1%) of 124, with sonographic guidance. None of these parameters showed statistically significant differences. In cases without atypia, for spring-loaded devices and directional vacuum-assisted devices, fewer than 12 biopsy samples were obtained in 48 and 12 cases, respectively, whereas 12 or more samples were obtained in 10 and 58 cases, respectively. Carcinoma was found at subsequent surgery in 9% (5/58) of lesions biopsied with a spring-loaded device, compared with 0% (0/70) of lesions biopsied with a directional vacuum-assisted device. Moreover, carcinoma not identified by core needle biopsy was found in 8% (5/60) of lesions sampled with fewer than 12 specimens per lesion versus 0% (0/68) sampled with 12 or more specimens. Both of these parameters directional vacuum-assisted devices versus spring-loaded devices and springloaded devices with fewer than 12 samples versus those with more than 12 samples showed statistically significant differences (p = 0.01 and 0.015, respectively). The five lesions showing no atypia but with carcinoma found at subsequent excision involved spring-loaded devices in all, stereotactic guidance in four lesions TABLE 1 Lesion Core Needle Biopsies Resulting in Missed Cancers Size (mm) Note. DCIS = ductal carcinoma in situ. a With calcifications. Biopsy Guidance Method (number of specimens per lesion were five, five, six, and 11), and sonographic guidance in one lesion (with five specimens obtained). Core needle biopsy showing radial scar apparently accurately excluded malignancy in 144 (92%) of 157 lesions on the basis of surgical correlation or mammographic follow-up with a median of 38 months (minimum, 24 months; mean, 45 months). If the 29 cases with associated atypical hyperplasia at percutaneous biopsy are excluded, then core needle biopsy apparently excluded malignancy in 122 (95%) of 128 lesions. All these malignancies involved masses or architectural distortions biopsied with spring-loaded devices and fewer than 12 samples. A summary of all cases of malignancy diagnosed as radial scar with or without atypical ductal hyperplasia at core needle biopsy is shown in Tables 1 and 2. Discussion Diagnostic surgical excision of suspicious mammographic lesions has been avoided in a high percentage of cases with the use of imaging-guided core needle biopsy in a cost-effective manner [21, 22]. Certain histologic diagnoses found at core needle biopsy, however, have been shown to be associated with a sampling error rate that is high enough to warrant surgical excision. One such example Needle Size Directional Vacuum-Assisted No. of Samples is atypical ductal hyperplasia. Although certain mammographic findings of atypical ductal hyperplasia have been described [23], they are neither sufficiently sensitive nor specific to avert biopsy so that the histologic finding of atypical ductal hyperplasia at core needle biopsy still requires surgical excision in most cases [17 19, 22]. In a like manner, imaging findings of radial scar on mammography or sonography have been described but are also inadequate to avoid histologic confirmation and exclusion of carcinoma [1, 15, 16]. Like atypical ductal hyperplasia, radial scar may be a microscopic incidental finding or may account for the imaging findings, especially because the size of such lesions vary, including complex sclerosing lesions that exceed 1 cm. Several reports based on review of surgical specimens have indicated an increased incidence of associated proliferative disease, malignancy, or both [8, 12, 16, 21, 22, 24 28] when radial scars are diagnosed so that an issue similar to the discovery of atypical ductal hyperplasia arises namely, if, and under what conditions, can the core needle biopsy diagnosis of radial scar be reasonably trusted? To our knowledge, the issue has not been previously studied. Nielsen et al. [9] reported an increased risk for breast cancer in a follow-up study of pa- Core Needle Biopsy Results Surgical Results Architectural distortion 10 Sonographic 12-gauge No 6 Radial scar, atypical ductal hyperplasia DCIS Architectural distortion 30 Stereotactic 14-gauge No 5 Radial scar DCIS Mass 10 Sonographic 14-gauge No 5 Radial scar DCIS Mass a 20 Stereotactic 14-gauge No 6 Radial scar Invasive ductal carcinoma Mass 20 Stereotactic 14-gauge No 5 Radial scar DCIS Mass a 7 Stereotactic 14-gauge No 11 Radial scar Invasive ductal carcinoma, DCIS Mass 15 Stereotactic 11-gauge Yes 18 Radial scar, atypical ductal hyperplasia Invasive ductal carcinoma Mass 12 Stereotactic 11-gauge Yes 18 Radial scar, atypical ductal hyperplasia DCIS Mass 20 Stereotactic 11-gauge Yes 10 Radial scar, atypical ductal hyperplasia DCIS Mass 15 Stereotactic 11-gauge Yes 20 Radial scar, atypical ductal hyperplasia DCIS Mass 8 Stereotactic 11-gauge Yes 15 Radial scar, atypical ductal hyperplasia Invasive ductal carcinoma, DCIS Microcalcifications 4 Stereotactic 11-gauge Yes 10 Radial scar, atypical ductal hyperplasia, lobular carcinoma in situ Microcalcifications 5 Stereotactic 11-gauge Yes 12 Radial scar, atypical lobular hyperplasia DCIS Invasive ductal carcinoma AJR:179, November 2002 1181

Brenner et al. TABLE 2 Specimens per Lesion Number of Radial Scar Lesions Without Atypia and Subsequent Number of Malignancies Found at Surgery a Gauge was 14 (n = 57) or 12 (n = 1). b Gauge was 14 (n = 6) or 11 (n = 64). c Five, five, five, six, and 11 specimens per lesion. tients diagnosed with radial scar, although no attempt was made to evaluate a causal relationship between the anatomic focus of radial scar and the development of carcinoma [29]. Linell et al. [8], in discussing the difficulty of distinguishing tubular carcinoma from radial scar because of atypical epithelium, suggested that tubular carcinoma may develop in an area of prior radial scar. Their study was conducted before the clinical implementation of immunohistochemical studies that may have modified such conclusions. Frouge et al. [14] reported on 40 patients with a coincident malignancy associated with 20 radial scars detected mammographically and removed at surgery, as well as one radial scar associated with atypical ductal hyperplasia and one associated with lobular carcinoma in situ. Hassell et al. [16] also found a large incidence of atypia and malignant disease associated with radial scars in a series in which pathology files identifying radial scars in the specimen initiated a retrospective determination of mammographic lesions; the authors recommended surgical excision when radial scar was identified as a possible cause. This suggestion is difficult to implement because as shown in our study radial scars may be associated with a variety of mammographic features and thus be considered a possible cause for virtually any mammographic lesion, rather than the more classically described focus of architectural distortion [11]. The diagnosis of radial scar can be made by core needle biopsy, as has been previously shown [30] as well as in this study. The spiculated configuration of the lesion and shape of tubules may resemble tubular carcinoma, but the presence of myoepithelial cells shown on either Spring-Loaded Device a Directional Vacuum-Assisted Device b Radial Scar Lesions Malignancies Found at Surgery Radial Scar Lesions Malignancies Found at Surgery Masses and architectural distortion <12 40 5 c 8 0 48 12 9 0 43 0 52 Microcalcifications <12 8 0 4 0 12 12 1 0 15 0 16 Total 58 0 70 0 128 Total H and E or immunohistochemical staining (e.g., actin) excludes the latter diagnosis. However, based on the association of radial scar and other proliferative or malignant processes, the decision regarding tissue sampling error requiring further surgery, rather than misdiagnosis of tubular carcinoma, should still be considered. Two considerations arise in evaluating the results of our study and the appropriate management of radial scars after core needle biopsy namely, the relative lack of cases of radial scar reported in the imaging literature and the absence of prior studies on the accuracy of core needle biopsy for reliably establishing this diagnosis. Among the larger series published during the past 10 years in the American literature, Ciatto et al. [12] reported 38 cases; Frouge et al. [14], 40 cases; Mitnick et al. [11], 14 cases; Adler et al. [1], seven cases; and Orel et al. [13], four cases [13]; a total of 103 cases. The retrospective series of Hassell et al. [16], appearing in the Canadian literature, reported 96 cases of radial scar, one third of which were considered incidental to the biopsy performed. In a recent series of presumed high-risk lesions found at core needle biopsy, Philpotts et al. [30] reported nine radial scars, eight of which were subject to excisional biopsy, four of which were associated with atypia, and none of which was associated with malignancy. The finding of atypical ductal hyperplasia associated with radial scar in our study is sufficient to prompt excision, not because of the radial scar, but because of the atypical hyperplasia. In our series of patients, 28% (8/ 29) of radial scars with associated atypical ductal hyperplasia (n = 7) or atypical lobular hyperplasia (n = 1) showed malignancy at excision. Another five radial scars showed unspecified atypia, but none of the five was associated with malignancy. Without the presence of atypical ductal hyperplasia or other atypia, five (4%) of 128 cases initially diagnosed by core needle biopsy as radial scars were associated with malignancy. None of these cases involved calcifications, a finding consistent with a report by Orel et al. [13] describing cases of radial scars in which calcifications were associated with both ductal and stromal elements. One case involved a focus of architectural distortion and four involved masses. In both these circumstances, no lesion was missed with directional vacuum-assisted devices, and all were missed with spring-loaded devices with fewer than 12 samples, although outcomes based on lesion type did not achieve statistical significance (p = 0.28). Four (1%) of 104 lesions were missed with stereotactic guidance and one (4%) of 24, with sonographic guidance. The use of directional vacuum-assisted biopsy devices is associated with potentially larger volumes of tissue per core biopsy sample [31, 32], although the volume of each sample was not measured in our study. The ease and speed of multiple samples retrieval with directional vacuum-assisted devices likely prompt retrieval of more specimens. For cases without atypia, fewer than 12 samples were obtained by use of spring-loaded devices and directional vacuum-assisted devices in 48 and 12 cases, respectively, whereas 12 or more samples were obtained with spring-loaded devices and directional vacuum-assisted devices in 10 and 58 cases, respectively (Table 2). The median number of samples obtained for spring-loaded devices for stereotactic and sonographic biopsies was seven and five, respectively, and for directional vacuum-assisted devices, 14 and 15, respectively. This finding may be one of the reasons that vacuum-assisted biopsies showed no missed cancers when the diagnosis of radial scar without atypia was made. This trend showed statistical significance. Although no cancers were missed using directional vacuum-assisted biopsy, the number of cases with fewer than 12 samples was relatively low, and 12 or more samples using any device spring-loaded or directional vacuum-assisted should be considered as an appropriate guideline. Four of the five lesions were missed with six or fewer samples, and only one with 11 samples. Two limitations should be considered in assessing the results of our study. The first relates to the multiinstitutional retrospective nature of 1182 AJR:179, November 2002

Core Needle Biopsy of Radial Breast Scars the data. Although all operators were experienced with both stereotactic and sonographically guided biopsies, specific skill in performing these procedures and decisions regarding how many samples to obtain for a given lesion were not standardized and may have affected results. These circumstances apply to current clinical practice among a wide variety of venues represented in our study group. The second limitation relates to the combined reporting of all results validated by both surgery, which accounts for two thirds of the patients in this study, and follow-up with a median of 38 months for the other third. This follow-up period is likely to be a valid basis for excluding malignancy for two reasons. First, without the aid of any intervention or tissue sampling, most lesions considered probably benign but found to be malignant were identified by interval change within 18 months of surveillance [33 35]. Second, Reynolds [36], in a review of 18,542 cases with interventional core biopsies, identified no interval cancers during a follow-up period of 24 31 months. Nevertheless, it may be worth considering the two subsets of patients those with surgical validation (n = 102) and those with clinical and mammographic surveillance (n = 55) and comparing certain results because of inherent selection biases that can affect management decisions. Optimal management decisions are made by recognizing potential sampling errors as well as by reconciling imaging and histologic results. In a prospective multiinstitutional study of core needle biopsies using only springloaded devices, Brenner et al. [18] reported an increased sensitivity of 1% when so-called high-risk lesions associated with sampling errors were included and an additional increased sensitivity of 6% when histologic findings were thought discordant with imaging findings [18]. Those patients who underwent surgery in our study may have had imaging findings that were considered less concordant with benign histologic findings, and results in these patients may tend to overestimate the expected incidence of malignancy missed by core needle biopsy. Conversely, those patients who underwent surveillance after core needle biopsy might have had less worrisome imaging findings, and the results in these patients may tend to underestimate the expected incidence of missed malignancy. In our series, 28 of 29 patients with associated atypia underwent excision, with malignancy found in eight cases (28%). If analysis is restricted to surgical cases, this incidence remains unchanged. The one patient who did not undergo surgery had only minimal atypical cytology and has been followed up without change for 52 months. However, for the 128 patients without atypical hyperplasia, the 4% (5/128) missed cancer rate cancer rate would be 7% (5/74) if analysis is limited only to patients with subsequent surgery, a difference without statistical significance (p = 0.37). This sensitivity analysis reinforces the importance of reconciling histologic and imaging findings and assessing procedural variables in making management decisions. Both Liberman et al. [37] and Brenner et al. [38] reported high overall accuracy of core needle biopsy for the diagnosis of mass lesions using a total of five 14-gauge spring-loaded biopsy samples, although Brenner et al. indicated that the same accuracy was less for architectural distortion as the presenting mammographic feature. Based on our study, the diagnosis of radial scar as the cause of a mass may be an exception to the high accuracy to be expected from core needle biopsy of masses with five samples only. Eight cases of ductal carcinoma in situ (two associated with lobular carcinoma in situ: one at core needle biopsy, the other at surgery), three cases of invasive ductal carcinoma, and two cases of combined invasive and intraductal carcinoma were found in our series. The 8% (13/ 157) incidence of malignancy in our series with surgery or follow-up for a median of 38 months (all > 24 months) is considerably lower than the 29% reported by Frouge et al. [14], the 31% reported by Hassell et al. [16], or the 29% adjusted incidence reported by Vazquez et al. [24] (excluding lobular carcinoma in situ included in that study). Our results reinforce the more recent findings by Philpotts et al. [30] of nine radial scars with core needle biopsy with no malignancies found at surgery. The prior reports without the use of core needle biopsy are difficult to reconcile with our experience because the number of cases with surgical excision in our series exceeds the total reported by both Frouge et al. and Vazquez et al., as well as that of Hassell et al., and the total reported number of cases in our series nearly doubles those previously published. In a retrospective analysis, radial scars have been reported to confer an increased risk of the patient s developing malignancy, increased further by associated proliferative disease, although the site of malignancy does not necessarily correspond to the site of radial scar [39]. The same issue pertains to the review of Nielsen et al. [29]. Wider local excision of the site of radial scar that has been biopsied with a sufficient number of samples and that shows no associated atypical ductal hyperplasia would not necessarily identify additional malignancies according to this longitudinal clinical study, similar to our results. However, specimen size of surgical excisions was not reported in our study, and we cannot determine the exact proximity of the 13 malignancies with respect to the radial scar; by definition, all were proximate. Although radial scar is often considered in the differential diagnosis of architectural distortion, our study indicates that other manifestations such as masses and calcifications may occur. Some cases of architectural distortion may have been characterized by different investigators as masses, and results should be viewed with this caveat. However, the presence of both mammographic and sonographic masses indicates that the presentation of radial scar is variable and that it may not always represent an incidental finding under these circumstances. Whether incidental to or directly related to the imaging findings, radial scars diagnosed at core needle biopsy require deliberate management recommendations, and the data presented in our study support conditions under which the reliability of this diagnosis may be reasonably determined. In summary, our series of patients represents the largest cohort of patients so far reported with the diagnosis of radial scar and the largest series subject to core needle biopsy. Although many cases of radial scar may represent incidental findings, the inclusion of different types of lesions in our study with directed imaging-guided biopsy and with the variable sizes of radial scars that exist pathologically permit a strategy for further management when the diagnosis of radial scar is reported. Prior studies suggesting excisional biopsy for all these lesions have been based primarily on retrospective reviews of surgical specimens without interventional core needle biopsy data. Core needle biopsy is likely to be a reliable method of diagnosing radial scar when there is no associated atypical hyperplasia at percutaneous biopsy, when the biopsy is performed with more than 12 samples especially when performed with a directional vacuum-assisted device and when mammographic findings are reconciled with histologic findings. If any of these conditions is not met, surgical excision is indicated. References 1. Adler DD, Helvie MA, Oberman HA, Ikeda DM, Bhan AO. Radial sclerosing lesion of the breast: mammographic features. Radiology 1990;176: 737 740 2. Eusebi V, Grassigli A, Grosso F. Lesioni focali AJR:179, November 2002 1183

Brenner et al. scleroelastotiche mammarie simulanti il carcinoma infiltrante. Pathologica 1976;68:507 518 3. Fisher ER, Palekar AS, Kotwal N, Lipana N. A nonencapsulated sclerosing lesion of the breast. Am J Clin Pathol 1979;71:240 246 4. Rickert RR, Kalisher L, Hutter RVP. Indurative mastopathy: a benign sclerosing lesion of breast with elastosis which may simulate carcinoma. Cancer 1981;47:561 571 5. Page DL, Anderson TJ. Diagnostic histopathology of the breast. Edinburgh: Churchill Livingstone, 1987:89 102 6. Schnitt SJ, Connolly JL. Pathology of benign breast disorders. In: Harris J, Morrow M, Hellman S, eds., Diseases of the breast. Philadelphia: Lippincott-Raven, 1996:34 35 7. Rosen PP. Rosen s breast pathology. Philadelphia: Lippincott-Raven, 1997:76 82 8. Linell F, Ljungberg O, Andersson I. Breast carcinoma: aspects of early stages, progression and related problems. Acta Pathol Micobiol Scand 1980;272[suppl]:1 223 9. Nielsen M, Jensen J, Andersen, JA. An autopsy study of radial scar in the female breast. Histopathology 1985;9:287 295 10. Tabar L, Dean PB. Teaching atlas of mammography. New York: Thieme-Stratton, 1985:85 136 11. Mitnick JS, Vazquez MF, Harris MN, Roses DF. Differentiation of radial scar from scirrhous carcinoma of the breast: mammographic-pathologic correlation. Radiology 1989;173:697 700 12. Ciatto S, Morrone D, Catarzi S, et al. Radial scars of the breast: review of 38 consecutive mammographic diagnoses. Radiology 1993;187:757 760 13. Orel SG, Evers K, Yeh I-T, Troupin RH. Radial scar with microcalcifications: radiologic-pathologic correlation. Radiology 1992;183:479 482 14. Frouge C, Tristant H, Guinebretiere JM, et al. Mammographic lesions suggestive of radial scars: microscopic findings in 40 cases. Radiology 1995;195:623 625 15. Finlay ME, Liston JE, Lunt LG, et al. Assessment of the role of ultrasound in differentiation of radial scars and stellate carcinomas of the breast. Clin Radiol 1994;49:52 55 16. Hassell P, Klein-Parker H, Worth A, Poon P. Radial sclerosing lesions of the breast: mammographic and pathologic correlation. Can Assoc Radiol J 1999;50:370 375 17. Jackman RJ, Nowels KW, Shepard MJ, Finkelstein SI, Marzoni FA. Stereotaxic large-core needle biopsy of 450 nonpalpable breast lesions with surgical correlation in lesions with cancer or atypical hyperplasia. Radiology 1994;193:91 95 18. Brenner RJ, Bassett LW, Fajardo LL, et al. Percutaneous core breast biopsy: a multiinstitutional prospective trial. Radiology 2001;218:866 872 19. Liberman L, Cohen MA, Dershaw DD, Abramson AF, Hann LE, Rosen PP. Atypical ductal hyperplasia diagnosed at stereotaxic core biopsy of breast lesions: an indication for surgical biopsy. AJR 1995;164:1111 1113 20. American College of Radiology. Breast imaging reporting and data system (BI-RADS), 3rd ed. Reston, VA: American College of Radiology, 1998 21. Liberman L, LaTrenta LR, Dershaw DD, et al. Impact of core biopsy on the surgical management of impalpable breast cancer. AJR 1997; 168:495 499 22. Lee CH, Egglin TK, Philpotts L, Mainiero MB, Tocino I. Cost-effectiveness of stereotactic core needle biopsy: analysis by means of mammographic findings. Radiology 1997;202:849 854 23. Stomper PC, Cholewinski BS, Penetrante RB, Harlos JP, Tsangaris TN. Atypical hyperplasia: frequency and mammographic and pathologic relationships in excisional biopsies guided with mammography and clinical examination. Radiology 1993;189:667 671 24. Vazquez MF, Mitnick JS, Pressman P, et al. Radial scar: cytological evaluation by stereotactic aspiration. Breast Dis 1994;7:299 306 25. Dahlstrom JE, Jain S, Sutton T, et al. Diagnostic accuracy of stereotactic core biopsy in a mammographic breast cancer screening programme. Histopathology 1996;28:421 427 26. Meyer JE, Smith DN, Lester SC, et al. Large-needle core biopsy: nonmalignant breast abnormalities evaluated with surgical excision or repeat core biopsy. Radiology 1998;206:717 720 27. Philpotts LE, Shaheen NA, Carter D, Lange RC, Lee CH. Comparison of rebiopsy rates after stereotactic core needle biopsy of the breast with 11-gauge vacuum suction probe versus 14-gauge needle and automatic gun. AJR 1999;172:683 687 28. Jackman RJ, Nowels KW, Rodriguez-Soto J, Marzon FA Jr, Finkelstein SI, Shepard MJ. Stereotactic automated, large-core needle biopsy of nonpalpable lesions: false-negative and histologic underestimation rates after long-term follow-up. Radiology 1999;210:799 805 29. Nielsen M, Christensen I, Andersen J. Radial scars in women with breast cancer. Cancer 1987;59:1019 1025 30. Philpotts LE, Shaheen NA, Jain KS, Carter D, Lee CH. Uncommon high-risk lesions of the breast diagnosed at stereotactic core-needle biopsy: clinical importance. Radiology 2000;216: 831 837 31. Jackman RJ, Marzoni FA, Nowels KW. Percutaneous removal of benign mammographic lesions: comparison of automated large-core and directional vacuum-assisted stereotactic biopsy techniques. AJR 1998;171:1325 1330 32. Berg WA, Krebs TL, Campassi C, Magder LS, Sun CCJ. Evaluation of 14- and 11-gauge directional vacuum-assisted biopsy probes and 14-gauge biopsy guns in a breast parenchymal model. Radiology 1997;205:203 208 33. Sickles EA. Periodic follow-up of probably benign mammographic lesions: results for 3184 consecutive cases. Radiology 1991;179:463 468 34. Varus X, Leborgne F, Leborgne JH. Nonpalpable, probably benign lesions: role of follow-up mammography. Radiology 1992;184:409 414 35. Vizcaino I, Gadea L, Andreo L, et al. Short-term follow-up results in 795 nonpalpable probably benign lesions detected at screening mammography. Radiology 2001;219:475 483 36. Reynolds HE. Core needle biopsy of challenging benign breast conditions: a comprehensive literature review. AJR 2000;174:1245 1250 37. Liberman L, Dershaw DD, Rosen PR, Abramson AF, Deutsch BM. Stereotaxic 14-gauge breast biopsy: how many core biopsy specimens are needed? Radiology 1994;192:793 795 38. Brenner RJ, Fisher P, Fajardo L, et al. Percutaneous core biopsy of the breast: effect of operator experience and number of samples on diagnostic accuracy. AJR 1996;166:341 346 39. Jacobs TW, Byrne C, Colditz G, Connolly JL, Schnitt SJ. Radial scars in benign breast-biopsy specimens and the risk of breast cancer. N Engl J Med 1999;340:430 436 1184 AJR:179, November 2002