Radial Scars without Atypia Diagnosed at Imaging-guided Needle Biopsy: How Often is Associated Malignancy Found at Subsequent Surgical Excision, and do Mammography and Sonography Predict which Lesions are Malignant?

Anna Linda; Chiara Zuiani; Alessandro Furlan; Viviana Londero; Rossano Girometti; Piernicola Machin; Massimo Bazzocchi


Am J Roentgenol. 2010;194(4):1146-1151. 

In This Article


The use of percutaneous biopsy for the initial evaluation of clinically occult breast lesions is now widespread and is a practical alternative to open surgical biopsy for many patients. In previous studies, investigators have reported high rates of concordance between the histologic findings of percutaneous biopsy and surgical biopsy.[20] Exceptions are lesions that have high-risk histology results; this heterogeneous group of lesions includes atypical ductal hyperplasia, lobular neoplasia, papillary lesions, fibroepithelial lesions, mucocelelike lesions, columnar cells lesions, and radial scars.[21] Diagnoses of these lesions based on percutaneous biopsy are considered to be unreliable because of the risk that the lesions will be upgraded to malignancy at surgical excision; therefore, when encountered in biopsy specimens, they raise repeated management questions.[22]

In lesions diagnosed as radial scars at percutaneous biopsy, underestimation of malignancy occurs not only because of sampling error (i.e., sampling only the region of radial scar in a lesion containing both radial scar and carcinoma), but also because of diagnostic error resulting from of the difficulty in differentiating radial scar from carcinoma, particularly of tubular histotype.[6] Studies addressing the issue of invasive carcinoma or in situ carcinoma occurring after a percutaneous diagnosis of radial scar have reported variable underestimation rates, ranging from 0 to 40%[6–16] (Table 2). Such variability is attributable primarily to limited study cohort sizes, differences in inclusion criteria, and possible biases with regard to the selection of lesions to undergo surgical excision. Furthermore, in some of these reports, the authors did not provide complete methodologic, radiologic, or clinical details.

Despite being included among the lower published malignancy underestimation rates, the 8% (5/62) incidence of carcinoma at surgical excision in this study shows that percutaneous biopsy of a radial scar is not reliable for diagnosis and should prompt surgical excision with pathologic examination of the entire lesion.

Regarding the histologic features of the five cancers detected at surgical excision, none was a tubular carcinoma. This finding suggests that underestimation was mainly due to sampling error rather than diagnostic error. The results of our study confirm previous findings that DCIS arising in radial scars are likely to be of low or intermediate grade and that invasive carcinomas are likely to be grade 1 or 2.[23] Three of the five malignancies were low- or intermediate-grade DCIS. This finding raises the question of overdiagnosis of slow-growing lesions that do not necessarily progress to invasive malignant disease. Therefore, a diagnostic delay due to a noninvasive management decision (imaging follow-up) of these cases likely would not have had a significant impact on patient prognosis.[24] Nevertheless, studies have shown that most of the malignant foci are contained within the confines of radial scars;[6,25,26] therefore, significant progression would be required before changes in lesion size or contour will become apparent.

In comparing the data from our study with the data from previous studies, it is necessary to consider that all patients with percutaneous diagnosis of radial scar in association with atypia (i.e., lobular neoplasia, atypical ductal hyperplasia, or both) were excluded from our evaluation, whereas in many published studies this exclusion criterion is not explicitly stated. Brenner et al.,[8] Resetkova et al.,[9] and Becker et al.[16] concluded that lesions without associated atypia at percutaneous biopsy performed with a vacuum-assisted device are amenable to follow-up instead of surgical excision in view of the low risk of malignancy. However, our results contradict their conclusions.

Limited or partial sampling of correctly targeted imaging lesions has been identified as the most important cause of underestimation[27] and the amount of tissue obtained per biopsy has become a matter of concern.

Compared with automated or semiautomated biopsy guns, vacuum-assisted biopsy devices provide pathologists with large individual samples obtained with larger needles, thereby inherently improving the architectural definition of difficult-to-diagnose lesions such as radial scars. Lomoschitz et al.[28] reported that the highest diagnostic yield at stereotactically guided vacuum-assisted biopsy can be achieved with 12 specimens per lesion and that this yield is not improved with more than 12 specimens.

Indeed, the malignancy underestimation rate in our series was lower using stereotactically guided 11-gauge vacuum-assisted biopsy devices than using sonography-guided 14-gauge automated or semiautomated guns. However, a false-negative result occurred. Notably, one lesion in which a vacuum-assisted device was used to obtain 13 specimens was underestimated (Fig. 1). This result does not seem to substantiate the findings of Brenner et al.[8] and Becker et al.[16] who reported that a diagnosis of radial scars without atypia at vacuum-assisted biopsy (using 14- or 11-gauge needle) with at least 12 specimens is reliable and that surgical excision is not required (Table 2). Nevertheless, in those series, an average of 14 and 32 cores per targeted lesion, respectively, were obtained. Resetkova et al.[9] had that same finding using an 11- or 9-gauge needle (Table 2). To our knowledge, our case is the first reported case of underestimation with a vacuum-assisted device. However, we must emphasize that the overall vacuum-assisted biopsy underestimation rate obtained by combining our results with those of the three mentioned literature references is only 0.6% (1/180).

In addition, our unique case of underestimation with a vacuum-assisted biopsy device is likely to be explained by inaccurate sampling of the radiologic lesion. In fact, a specimen radiograph obtained after the procedure showed a few calcifications, suggesting that the mammographic lesion was not entirely removed and that residual microcalcifications were left after vacuum-assisted biopsy (Fig. 1).

A complementary explanation for this unusual occurrence might be that the diagnostic quality of sampling in this setting is reflected not only in the amount of sampling relative to the extent of the lesion (the more extensive the sampling, the lower the likelihood of underestimation) but also in the extent of the malignant disease relative to the extent of the entire lesion. In fact, López-Medina et al.[6] showed that the proportion of the radial scar volume involved by carcinoma ranged from 3.7% to 16.2% (mean, 8.3%) and was located invariably at the periphery of the lesion. As a consequence, even very extensive sampling might theoretically miss malignant foci.[23]

Further investigation with a larger patient database and with correlation between complete removal of the radiologically targeted lesion and underestimation is necessary to clarify this issue.

In our series, mammographic and sonographic appearances were not able to predict which radial scars diagnosed at percutaneous biopsy would be malignant at surgical excision. In particular, underestimation of malignancy occurred in two of 11 cases (18%) presenting as microcalcifications, which were diagnosed as radial scars without atypia at percutaneous biopsy but proved to be low-grade DCIS at surgical excision. This result confirms the findings of Becker et al.,[16] who reported that two of 37 radial scars (5%) without atypia diagnosed at percutaneous biopsy and presenting as microcalcifications were upgraded to malignancy at surgery. On the contrary, this finding is in contrast with the results of Brenner et al.[8] and Resetkova et al.;[9] in their studies, those investigators found no case of underestimation among 28 radial scars without atypia and among 55 radial scars (with and without atypia) presenting as microcalcifications, respectively.

With regard to sonographic lesions, of the 17 circumscribed masses, none was found to be malignant at excision. However, this malignancy rate was not statistically different from that found in noncircumscribed lesions (4/28), possibly because of small sample size.

Although further studies are necessary to support these findings, our results show that even calcifications should be excised when well-performed percutaneous biopsy shows radial scar regardless of its association with atypia.

When pathology results from percutaneous biopsy yield radial scar that was not suspected on imaging, it is difficult to determine if the lesion biopsied is an atypical-looking radial scar or if a microscopic radial scar was biopsied in proximity to the lesion being targeted. According to Cawson et al.[7] and Hassell et al.,[19] up to one third of radial scars found on percutaneous biopsy are incidental to the lesion being targeted.

One might speculate that incidental radial scars are at the highest risk of underestimation; that is, the suspicious radiologic lesion and the target for the biopsy are located in the immediate vicinity of the radial scar but are not part of the radial scar. Therefore, the diagnosis of radial scar is likely to be based just on a limited percentage of the sampled tissue. This partial and peripheral sampling of the lesion will not allow malignant foci to be reliably excluded and surgical excision will be required.

Our study has some limitations. First, it is a retrospective study with a limited sample size. Second, heterogeneous biopsy technique was used: 19 lesions were biopsied with an 11-gauge vacuum-assisted device, and 43 with a 14-gauge automated or semiautomated needle biopsy. Third, the pathologic slides of core specimens were not retrospectively reviewed by a pathologist and the original pathologic interpretation has been accepted.

In conclusion, our results showed that, first, a diagnosis of a radial scar based on percutaneous biopsy does not exclude associated malignancy at subsequent surgical excision; and, second, mammographic and sonographic appearances of a lesion diagnosed as a radial scar at imaging-guided percutaneous biopsy are not able to predict which lesions will have associated malignancy at subsequent surgical excision.

Our results suggest that surgical excision is required for lesions yielding radial scars at percutaneous biopsy regardless of the mammographic and sonographic appearance.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.