Lobular neoplasia on core needle biopsy does not require excision

Authors

  • Chandandeep S. Nagi MD,

    Corresponding author
    1. Lillian and Henry Stratton/Hans Popper Department of Pathology, Mount Sinai School of Medicine, New York, New York
    • Department of Pathology, Mount Sinai School of Medicine, Box 1194, 1 Gustave L. Levy Place, New York, NY 10029
    Search for more papers by this author
    • Fax: (212) 534-7491.

  • James E. O'Donnell MD,

    1. Lillian and Henry Stratton/Hans Popper Department of Pathology, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Mikhail Tismenetsky MD,

    1. Lillian and Henry Stratton/Hans Popper Department of Pathology, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Ira J. Bleiweiss MD,

    1. Lillian and Henry Stratton/Hans Popper Department of Pathology, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author
  • Shabnam M. Jaffer MD

    1. Lillian and Henry Stratton/Hans Popper Department of Pathology, Mount Sinai School of Medicine, New York, New York
    Search for more papers by this author

Abstract

BACKGROUND.

Lobular neoplasia (LN), encompassing atypical lobular hyperplasia (ALH) and lobular carcinoma in situ (LCIS), is often an incidental finding on core needle biopsies (CNBs) performed in instances of radiologic densities and/or calcifications. Because LN is generally considered a risk factor for breast carcinoma, the utility of subsequent excision is controversial.

METHODS.

The authors' database yielded 98 cases of LCIS and/or ALH. Cases containing LN accompanied by a second lesion mandating excision (eg, radial scar, atypical ductal hyperplasia [ADH]) and those failing to meet strict diagnostic criteria for LN (eg, atypical cells, mitoses, single-cell necrosis) were excluded. Radiographic calcifications were correlated with their histologic counterparts in terms of size, number, and pattern.

RESULTS.

Ninety-one biopsies were performed for calcifications and 7 were performed for mass lesions. The ages of the patients ranged from 35 to 82 years. Fifty-three patients were followed radiologically without excision, 42 of whom had available clinicoradiologic information. The 45 patients who underwent excision were without disease at follow-up periods ranging from 1 to 8 years. Of these 45 patients, 42 (93%) had biopsy results demonstrating only LN. The remaining 3 patients had biopsies with the following findings: ADH in 1 biopsy, residual LCIS and a separate minute focus of infiltrating lobular carcinoma (clearly an incidental finding) in the second biopsy, and ductal carcinoma in situ admixed with LCIS in the third biopsy (a retrospective examination performed by 2 blinded breast pathologists revealed foci of atypical cells and mitoses).

CONCLUSIONS.

Excision of LN is unnecessary provided that: 1) careful radiographic-pathologic correlation is performed; and 2) strict histologic criteria are adhered to when making the diagnosis. Close radiologic and clinical follow-up is adequate. Cancer 2008. © 2008 American Cancer Society.

The use of core needle biopsy (CNB) of the breast has dramatically increased in recent years as the technology for breast imaging has improved in terms of both screening and diagnosis. In many cases, CNB has become the initial method of evaluating patients with mammographic calcifications or other nonpalpable masses/lesions. Further management of patients rests on the diagnosis rendered from these specimens.

One of the more controversial issues to have arisen from this shift in procedure is the classification and management of lobular neoplasia (LN) on CNB. Although mentioned in the literature since the early 1900s, to our knowledge the histologic features of lobular carcinoma in situ (LCIS) were first explicitly described by Foote and Stewart in 1941.1 The term was established to articulate the morphologic parallels between the cells of LCIS and invasive lobular carcinoma. The designation of atypical lobular hyperplasia (ALH) consequently evolved to describe those lesions that are similar to LCIS, but which fall quantitatively short of the diagnostic criteria. The term “lobular neoplasia” was devised by Haagensen et al. in 1978 as a means of illustrating the spectrum of proliferative changes encompassing both ALH and LCIS and hopefully decreasing both interobserver and intraobserver variability.2

Classic LN, as determined histologically on CNB, occurs in the absence of any clearly discernable and reliable radiologic abnormality. In cases in which this lesion is diagnosed on CNB, it does not present as a mass and only rarely is observed as an area of suspicious calcifications. LN is found in association with mass lesions (ie, fibroadenomas, cysts, radial scars, intraductal papillomas), fibrocystic changes containing calcifications, atypical ductal hyperplasia [ADH]/ductal carcinoma in situ [DCIS], or, occasionally, as so-called pleomorphic LCIS (pLCIS; which has a significantly higher probability of demonstrating microcalcifications). Therefore, correlating the presence of LN on CNB with invasive disease on subsequent excisional biopsy is problematic. Because another lesion is in most cases ultimately responsible for any abnormal radiologic appearance, the concomitant presence of LCIS in these biopsies can be considered incidental. This observation, along with the finding that LN is largely considered to be only a risk factor rather than a direct precursor to invasive carcinoma, has led to controversies with regard to its management (ie, whether LN requires surgical excision).3–9 In addition, because of the rarity of the lesion compared with DCIS, the accumulation of convincing and meaningful data to help guide therapy has been difficult. We examined cases from our practice in which all CNBs were carefully correlated with the corresponding radiologic findings. Although some patients underwent surgical excision, others were instead followed clinically and radiologically.

MATERIALS AND METHODS

The pathology computer database of our institution, a large tertiary care medical center, was searched using the following phrases: “core biopsies” and “atypical lobular hyperplasia (ALH)” and/or “lobular carcinoma in situ (LCIS)” and “CNB” from January 1994 through June 2005. Only those cases in which the lobular neoplasia (ALH and/or LCIS) was considered incidental to the radiologic findings were included. The term incidental, as used in this study, refers to biopsies in which the diagnosed LN was radiologically occult after careful radiologic-pathologic correlation (ie, the mass lesion and/or the number and pattern of calcifications was explained via histologic findings other than the LN). All the CNBs were either performed under ultrasound (using a 16-gauge needle, an 18-gauge needle, or a 20-gauge needle) or stereotactic guidance using a vacuum suction probe (predominantly an 11-gauge needle but also a 14-gauge needle). Tissue sections were fixed in formalin, embedded in paraffin, and stained with hematoxylin and eosin. The tissue was sectioned at 4-μm intervals, and 4 total levels of each block were examined initially. In each case, radiologic-pathologic correlation was performed. In each mammographically guided biopsy that was performed for calcifications, the specimen radiographs performed by the radiologists were submitted with the biopsies and examined at the time of the initial slide review. Both the number and the pattern of calcifications present on the specimen radiographs were matched to those on the histologic slides. In all discordant cases, additional radiographs of the paraffin blocks and deeper levels were performed as needed until a satisfactory radiologic-pathologic correlation was achieved.

All the breast CNB results were read by 1 of 3 dedicated breast pathologists. All follow-up surgical excision (FSE) specimens acquired at the study institution were also reviewed to confirm the diagnosis. In those cases in which the FSE was performed at an outside institution, the pathology report was obtained and considered to be the definitive result. In all FSE cases, the original CNB site was identified histologically. Clinical information, including patient demographics, reason for biopsy, and subsequent follow-up (whether surgical or radiologic) was gathered. All patients were followed via annual mammography.

When confirming the diagnosis of LN, strict criteria were utilized. The lobular cells were uniform, small, and dyscohesive with a bland nucleus, uniform chromatin, and indistinct nucleoli. These cells also possessed sparse cytoplasm and were devoid of any necrosis. They may or may not have demonstrated pagetoid extension into surrounding ducts. All other cases of LCIS type B, pLCIS, or in situ carcinoma with mixed ductal and lobular features (pleomorphism, mitoses, or single-cell necrosis) were excluded. ALH was defined as histologic changes of LCIS to a lesser quantitative degree.10 Any coexisting benign lesions not requiring excision were documented, as was the exact histologic location of calcifications. In all cases, after careful radiologic-pathologic correlation was performed and the LN was determined to be incidental, the decision to perform a FSE was made at the discretion of the radiologist/surgeon.

RESULTS

Our search produced 98 cases of CNB with a diagnosis of purely incidental ALH and/or LCIS after excluding those cases with a second lesion requiring excision. In all cases, none of the patients was found to demonstrate either clinical or radiologic evidence of malignancy in the breast from which the CNB was obtained. All the cases were from women aged 35 to 82 years (mean age, 57 years). Ninety-one cases (93%) were performed under stereotactic guidance for the evaluation of calcifications (calcifications that were either new, changing over time, or suspicious), and 7 were performed for the evaluation of a nonpalpable mass by ultrasound guidance (7%). The indications for the imaging study performed are listed in Table 1. The exact number of cores varied from patient to patient, but were on average approximately 5 cores for ultrasound-guided CNB and 15 cores for stereotactically guided CNB. LCIS was identified in 49 cases (50%), ALH was found to be present in 35 cases (35.7%), and both ALH and LCIS were noted in 14 cases (14.3%). Forty-five patients (46%) underwent a subsequent FSE and 53 patients (54%) were followed radiologically with no FSE performed. Of these 53 patients, subsequent clinical information was not available for 11 patients, who were classified as lost to follow-up. The remaining 42 patients with only radiologic follow-up (ranging from 1–8 years at the time of the current study, with a mean follow-up of 2.9 years) were all stable. In all the cases performed for the evaluation of calcifications, the calcifications were predominantly present in the surrounding fibrocystic changes and in only 7 cases (7.7%) were also present within the LN. Of these 7 cases of LN with calcifications, the vast majority of calcifications were present within the surrounding fibrocystic changes and involved the LN only focally. Careful radiologic-pathologic correlation in all these cases revealed the calcifications to correlate with fibrocystic changes in terms of the number of cores and the pattern of calcifications. In these cases, it was impossible to determine whether the calcifications present were because of the LN or whether they were already present and caused by fibrocystic changes with secondary involvement by the LN. Of the 7 cases performed for a mass lesion, 6 were found to be fibroadenomas and 1 was a benign cyst. Of these 7 cases, 6 were found to contain LN outside the mass lesion (5 fibroadenomas and 1 cyst). The remaining case of a fibroadenoma contained LN both within and outside the fibroadenoma. Of the 45 patients who underwent FSE, 42 underwent a needle-localized excision and 3 patients underwent a prophylactic mastectomy (1 with invasive carcinoma of the contralateral breast and 2 with intraductal carcinoma of the contralateral breast).

Table 1. Indications for Stereotactic Biopsy for Calcifications
Type of calcificationsNo. of casesLN with calcifications
  1. LN indicates lobular neoplasia; NOS, not otherwise specified.

NOS416
Clustered281
Punctate3 
Granular7 
Pleomorphic7 
Fine3 
Coarse and linear2 

Of the 45 patients who underwent FSE, 42 (93%) had either the same diagnosis on excision as on CNB or the ALH diagnosis was modified to LCIS (in 3 of the 42 cases). The remaining 3 cases were as follows. Case 1 had residual LCIS and a separate focus of microinvasive lobular carcinoma (measuring <1 mm), Case 2 had DCIS admixed with LCIS, and Case 3 had ADH. The original indication for biopsy in all 3 of these cases was the presence of calcifications (new clustered calcifications, low radiologic suspicion). In all these cases, the LCIS on the original CNB was not associated with calcifications. All calcifications in these cases were associated with fibrocystic changes and were correlated both in terms of number and pattern (ie, the target calcifications were present and identified on the histologic sections). Although the LN in these cases was determined to be incidental, because of a lack of standard of care in terms of treatment options, these calcifications were excised at the discretion of the radiologists and surgeons involved.

DISCUSSION

ALH and LCIS are often multifocal and not uncommonly present in the contralateral breast.4–10 They have generally been considered to be risk factors for the development of invasive carcinoma, with relative risk rates of 4 to 5 times for ALH and up to 8 to 10times for LCIS reported in the literature.4, 11–14 This is in contrast to ADH and DCIS, which are considered by most as direct precursor lesions for the development of malignancies.3, 10

Hypothesizing that LCIS represented a process capable of evolving into an invasive carcinoma, mastectomy was initially recommended by Foote and Stewart as definitive treatment after diagnosis.1 Currently, this practice has largely been replaced by careful radiologic follow-up with or without antiestrogenic hormonal therapy.4–6, 15, 16

To our knowledge to date, no consensus has been effectively reached outlining the best therapy for LN lesions detected on CNB. Although many subtypes of LCIS have been described (pleomorphic, histiocytoid, rhabdoid, signet ring cell, and apocrine) on the basis of cytologic criteria,3, 4, 10, 17–19 there is evidence in the literature that these variants are associated with a worse prognosis when compared with classic LCIS.4, 10 We restricted ourselves to assessing the follow-up of only those lesions meeting the strict criteria of LCIS with classic type A cells. In addition, pleomorphic LCIS is often associated with mammographic calcifications, and when invasive carcinoma does arise in association with pleomorphic LCIS it does so in its immediate vicinity.4, 20 These findings are more akin to the radiologic and clinical behavior of DCIS and therefore were excluded from our study cases because we always recommend excision of these lesions.

Determining a true incidence of LN is problematic because to our knowledge no reliable radiologic correlate currently exists, and it is likely that many asymptomatic women go undiagnosed. Even with the utilization of magnetic resonance imaging, a reliable diagnosis of LCIS may not be achievable because the enhancement of fibrocystic changes may mask any surrounding enhancement of LCIS. In addition, the true histologic correlate of “an area of enhancement” is difficult to investigate because the static histologic correlate of increased vascular flow is impossible to ascertain. However, the literature documents this rate as ranging anywhere from 0.5% to 8% in breast biopsies, with this discrepancy representing a gradual increase extending up to the present day.3–6, 10, 14, 20–22 This variation and the increased incidence noted in recent years is likely correlated with improvements in breast imaging and the subsequent increase in the frequency of CNB, as well as the greater amount of tissue acquired through the larger bore vacuum suction needles.

LN is still quite rare compared with ductal lesions of the breast. Therefore, many reports addressing the correlation between this entity and invasive carcinoma have suffered from a deficiency in the number of available cases. In addition, up to 50% of patients with LCIS have multifocal disease in the ipsilateral breast, and up to one-third have disease in the contralateral breast.3, 10 The increased risk that LN confers constitutes subsequent invasive carcinoma of both ductal and lobular type in either breast.3–6, 10, 20, 23

Classic LCIS is rarely associated with calcifications and does not present as an area of mammographic abnormality.3, 5, 10, 23 In those cases in which calcifications are identified, they are usually restricted to areas of fibrocystic change adjacent to the LCIS. In other words, it is an incidental finding. Although molecular genetic techniques have led to the discovery of certain similar gene truncations in both LCIS and invasive lobular carcinoma, to our knowledge, no causal relation has been established to date.3, 4, 10 Some recent studies have indicated that the increased risk LN was believed to confer bilaterally may in fact be greater (up to 3 times greater) in the ipsilateral breast, suggesting its possible role as a precursor lesion.24 Other studies have argued for the precursor nature of LCIS, basing this assertion on a meta-analysis of epidemiologic studies that demonstrated that 15% of patients diagnosed with LCIS developed invasive carcinoma in the ipsilateral breast.3, 10 One pertinent concern with these analyses is that no mention is made of whether the 2 lesions are spatially related (ie, whether the invasive carcinoma directly arose in the vicinity of the LCIS). Without this spatial information, it is difficult to reliably determine whether a direct precursor association exists as in DCIS, or if LCIS represents a risk factor for its development. This quagmire, along with the dilemma of multifocality combined with the rarity of the disease and its incidental nature, has resulted in disagreement with regard to the proper management of the lesion. The goal of the current study was to examine the follow-up data for those patients in which LN was incidental and not associated with a secondary lesion requiring surgical excision.

All patients from our LN cases who did not undergo surgical excision (N = 42) were found to be radiologically stable at the time of last follow-up (mean, 2.9 years [range, 1–8 years]). In addition, no other concurrent lesions or masses were found in the ipsilateral breast. Of those patients who underwent FSE (n = 45), 3 had their diagnoses modified, 2 significantly (2.3% [2 of 87 patients]). In each of these cases, the mammographic calcifications were of low suspicion. It is worthwhile to examine the details of each case.

Case 1 was a 75-year-old female with no significant past medical history or family history who presented for the evaluation of pleomorphic calcifications. The CNB demonstrated fibrocystic changes associated with calcifications and incidental ALH. On surgical excision, the biopsy site was identified as containing residual ALH and LCIS without calcifications. Separate from the biopsy site was a focus of invasive lobular carcinoma measuring <1 mm and not associated with any calcifications. On rereview of the mammographic films, no masses or areas of architectural distortion were identified. On rereview of the CNB slides, a small, monotonous proliferation of cells typical of type A or classic LCIS was observed (Fig. 1). Hence, this invasive carcinoma was clearly a finding incidental to the imaging studies.

Figure 1.

Example of a typical case with radiologically occult lobular neoplasia (LN). (a) Specimen radiograph demonstrating clustered pleomorphic calcifications of low radiologic suspicion (arrowheads). (b) Low-power view of the same case depicting incidental lobular carcinoma in situ (LCIS). Inset: the calcifications corresponding to those in the radiograph were identified within fibrocystic tissue away from the LN (H & E, × 100). (c) High-power view demonstrating the proliferation of small, uniform cells expanding the lobules, characteristics of classic LCIS (H & E, × 400).

Case 2 was a 47-year-old female without any significant past medical history who presented for the evaluation of indeterminate calcifications. Routine levels indicated fibrocystic changes with calcifications and LCIS without calcifications. FSE, which was performed at an outside institution, revealed biopsy site changes containing LCIS admixed with areas of DCIS (in situ carcinoma with mixed ductal and lobular features) that were not associated with any calcifications. The calcifications were present only in association with fibrocystic changes. Rereview of the original CNB specimen by 2 blinded pathologists revealed a lobular proliferation with foci of larger, atypical, and pleomorphic cells with mitoses (Fig. 2). The conclusion arrived at retrospectively was that the case would be better classified as type B LCIS/DCIS (ie, a case requiring surgical excision).

Figure 2.

A comparison of the radiologically occult case of lobular neoplasia (LN) shown in Figure 1 with Case 2. (a) Low-power view demonstrating a proliferation of cells expanding into the lobule (H & E, × 100). (b) High-power view of the same case demonstrating somewhat larger cells with mildly pleomorphic nuclei and mitoses (arrow), features of pleomorphic lobular carcinoma in situ not noted in the classic form depicted in Figure 1 (H & E, × 400).

Case 3 was a 68-year-old female without any significant medical history who presented for the evaluation of calcifications. A biopsy was performed, and after routine examination a diagnosis of fibrocystic changes associated with calcifications and ALH was rendered. FSE revealed biopsy site changes, ADH, and LCIS. No calcifications were noted in association with the ADH. This upgrade was made at an outside institution, and we were unable to acquire the slides for review and confirmation.

One of the most crucial elements in the management of LN is precise correlation between the radiologic image and the associated histologic findings on biopsy. In our study, radiologic-pathologic comparison was performed for all cases of LN at the time of the original CNB. This methodology was used to ensure that all factors influencing subsequent excisional biopsy were known to us. Other studies have shown higher rates of invasive carcinoma on excisional biopsy performed after a CNB with diagnosed LN. This disparity may be correlated in part with the retrospective nature of these studies, in which the purpose of FSE after CNB was not always known.20, 25 For example, a discrepancy between the radiologically and pathologically identified mass lesion/calcifications in a case may have resulted from 1 or both of the following reasons: 1) the mass lesion/calcifications, particularly those prompting the biopsy, may have been inadequately sampled; or 2) the mass lesion/calcifications of interest may have been left unsampled in the histologic block. If LN was observed in the CNB without another lesion mandating excision (ie, the mass lesion/calcifications noted on mammography), then one could mistakenly reason that the upstaging of the excisional biopsy was connected to the presence of LN. This point illustrates the profound importance of image correlation.

Although it has been widely accepted that ductal lesions on CNB convey a significantly increased likelihood of concomitant invasive carcinoma on excisional biopsy, it also appears from recent studies that this may be true, although to a lesser extent, with pLCIS.19, 26 Many previous reports have failed to discriminate and/or segregate cases of classic LCIS from those containing pLCIS, which may in part explain their calculated rates of lesional upstaging on CNB.5, 15, 20 Furthermore, pLCIS is significantly more likely to present radiologically with calcifications when compared with its classic counterpart. In fact, its radiologic and clinical behavior is more similar to that of DCIS. Hence, upstaged cases with microcalcifications within the LCIS may have demonstrated pleomorphic features on CNB that were not described.

One published report by Bauer et al. addressed both the issues of image correlation and the adoption of specific histologic criteria.6 The group's results, which closely parallel those of the current study, suggested that excisional biopsy need only be performed when LCIS is found in the presence of an accompanying significant lesion (ie, ADH, DCIS, radial scar, etc.). A study by Liberman et al. revealed similar findings with regard to cases with LCIS and detailed 3 situations necessitating FSE: 1) the presence of histologic features overlapping those of DCIS, 2) the identification of a coexisting high-risk lesion (as mentioned earlier), and 3) a discordance in findings on imaging compared with the corresponding histology.21

Because of the uncertainty regarding the biologic behavior of LN, controversies exist with regard to its proper management. Data from the current study indicate that, in the absence of radiologic discrepancies and excluding LN with larger cells and proliferative activity, the classic forms of LN can be managed with clinical surveillance without the need for surgical intervention. LN should be surgically excised if a radiologic-pathologic discordance exists (both in terms of the number and pattern of calcifications), if there is a second biologically aggressive lesion requiring surgical excision, or if the lobular proliferation demonstrates ductal features. Close radiologic and clinical surveillance to identify other preinvasive lesions or invasive carcinoma is a safe and adequate management guideline for LCIS until further molecular genetic techniques are developed that prove or disprove a direct causal relation between LN and invasive carcinoma.

Acknowledgements

We wish to thank Medical Imaging of Manhattan (Drs. Miriam Levy, Ulana Suprun, and Helene Tapper); Murray Hill Radiology and Mammography (Drs. Julie Mitnick, Barbara Baskin, Orna Hadar, Stacy Tashman, Stacey Vitiello, and Stephanie Zalasin); and Drs. George Hermann, Jolinda Mester, Susan Drossman, and Zeva Herman for their invaluable help.

Ancillary