Annals of Surgical Oncology

, Volume 19, Issue 10, pp 3131–3138

Long-Term Follow-up of Lobular Neoplasia (Atypical Lobular Hyperplasia/Lobular Carcinoma In Situ) Diagnosed on Core Needle Biopsy

Authors

  • Miraj G. Shah-Khan
    • Department of SurgeryMayo Clinic
  • Xochiquetzal J. Geiger
    • Department of Laboratory MedicineMayo Clinic
  • Carol Reynolds
    • Department of Laboratory Medicine and PathologyMayo Clinic
  • James W. Jakub
    • Department of SurgeryMayo Clinic
  • Elizabeth R. DePeri
    • Department of RadiologyMayo Clinic
    • Department of RadiologyMayo Clinic
Breast Oncology

DOI: 10.1245/s10434-012-2534-9

Cite this article as:
Shah-Khan, M.G., Geiger, X.J., Reynolds, C. et al. Ann Surg Oncol (2012) 19: 3131. doi:10.1245/s10434-012-2534-9

Abstract

Background

Lobular neoplasia (LN) includes atypical lobular hyperplasia (ALH) and lobular carcinoma in situ (LCIS). LN often is an incidental finding on breast core needle biopsy (CNBx) and management remains controversial. Our objective was to define the incidence of malignancy in women diagnosed with pure LN on CNBx, and identify a subset of patients that may be observed.

Methods

Patients diagnosed with LN on CNB between January 1993 and December 2010 were identified. Patients with an associated high-risk lesion or ipsilateral malignancy at time of diagnosis were excluded. All cases were reviewed by dedicated breast pathologists and breast imagers for pathologic classification and radiologic concordance, respectively.

Results

The study cohort was comprised of 184 (1.3 %) cases of pure LN (147 ALH, 37 LCIS) from 180 patients. Pathologic–radiologic concordance was achieved in 171 (93 %) cases. Excision was performed in 101 (55 %) cases and 83 (45 %) were observed. Mean follow-up was 50.3 (range, 6–212) months. Of cases excised, 1 of 81 (1.2 %) ALH and 1 of 20 (5 %) LCIS cases were upstaged to ductal carcinoma in situ (DCIS) and invasive lobular carcinoma (ILC), respectively. Only 1 of 101 (1 %) concordant lesions was upstaged on excision. Of the cases observed, 4 of 65 (6.2 %) developed ipsilateral cancer during follow-up: 1 of 51 (2 %) case of ALH and 3 of 14 (21.4 %) cases with LCIS (2 ILC, 2 DCIS). During follow-up, 2.9 % (4/138) patients with excised or observed LN developed a contralateral cancer.

Conclusions

These data support that not all patients with LN diagnosed on CNB require surgical excision. Patients with pure ALH, demonstrating radiologic–pathologic concordance, may be safely observed.

Lobular neoplasia (LN) refers to a spectrum of proliferation consisting of atypical lobular hyperplasia (ALH) and lobular carcinoma in situ (LCIS).1 Traditionally thought of as a risk factor for the development of invasive ductal or lobular cancer in either breast, more recent studies have suggested that LN may actually be a nonobligate precursor to malignancy.25 This has contributed to significant controversy regarding the management of LN when found on core needle biopsy (CNBx).6 Current National Comprehensive Cancer Network (NCCN) guidelines recommend excision of LCIS when identified on CNBx to rule out an associated malignancy.7 The guidelines do not specify the management of ALH, and a lack of consensus exists in the literature regarding the rate of upstaging to malignancy of LN.

The objective of this study was to identify the incidence of malignancy in women diagnosed with LN on CNBx and to identify a subset of patients who can be safely observed. Our study attempts to define the natural history of LN by evaluating a large group of patients with pure LN, not associated with a high-risk lesion, with long-term follow-up.

Materials and Methods

Patients

Following Institutional Review Board approval, we identified all cases with a diagnosis of ALH or LCIS on CNBx performed at our institution (Mayo Clinic, MN and Mayo Clinic, FL, USA) between January 1993 and December 2010. Cases of LN associated with ipsilateral invasive carcinoma and/or ductal carcinoma in situ (DCIS) were excluded. Cases with pleomorphic LCIS (pLCIS) or an associated high-risk lesion, including atypical ductal hyperplasia (ADH), papilloma, radial scar, or flat epithelial atypia (FEA) also were excluded. Patients were classified into two groups, surgical excision or observation, based on a 6-month cutoff from the time of CNBx to surgical excision. Clinical data were compiled from medical records to identify variables, such as age at LN diagnosis, personal history of breast cancer, and family history. Follow-up data were obtained from the medical record to determine the development of ipsilateral or contralateral breast cancer, subsequent breast surgery, and use of chemoprevention. Length of follow-up was determined by most recent documentation of clinical breast evaluation or imaging.

Imaging

All imaging was re-reviewed by a dedicated fellowship-trained breast radiologist (KG, ED). The imaging indication that led to CNBx was recorded as calcifications, mass, or MRI enhancement. Imaging modality used for biopsy was noted as stereotactic, ultrasound-guided, or MRI-guided. Information regarding the number of cores and needle gauge used for CNBx was recorded. All lesions were reviewed to determine radiologic–pathologic concordance.

Pathology

All CNBx and surgical specimens were re-reviewed by a dedicated breast pathologist (CR, XG). ALH and LCIS have identical cytologic features and were defined histologically by the differing degrees of distention of the acini within the terminal duct lobular unit (TDLU). LCIS was diagnosed if the following criteria were met: (1) the TDLU demonstrated a dyshesive, uniform population of cells filling the involved acini, and (2) greater than 50 % of the acini within the TDLU were greatly distended or expanded. “Greatly distended” was classified as more than 8–10 cells across an acini. TDLUs that lack these characteristic histologic findings were classified as ALH (Fig. 1). CNBx were examined for the presence and location of calcifications (within LN, benign acini, or stroma). The percentage of TDLUs involved with LN was recorded. An upgrade was defined as a diagnosis of invasive carcinoma and/or DCIS within the final excised specimen.
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-012-2534-9/MediaObjects/10434_2012_2534_Fig1_HTML.jpg
Fig. 1

Histopathologic appearance of lobular neoplasia a lobular carcinoma in situ. Note that the acini within the lobular unit are filled and show significant distension by the neoplastic proliferation. b Atypical lobular hyperplasia. Cytologically similar to LCIS, with dyshesive and uniform cells, but minimal distension of the acini

Statistical Analysis

Statistical analysis was performed by using chi-squared or Fisher’s exact test. Comparison of means was performed by using t test.

Results

During the study period, 13,765 image-guided CNBx were performed at Mayo Clinic, MN; 411 (3 %) cases of LN were identified. Seventy-three cases were excluded due to an associated high-risk lesion (ADH, papilloma, FEA, radial scar, or pLCIS), and 163 were excluded due to a known ipsilateral invasive cancer or DCIS at the time of diagnosis. Eight patients refused research authorization. The remaining 167 cases were included in the study and represent 1.2 % of CNBx and 40.6 % of LNs during this study period. An additional 17 cases that met the study criteria during this time period were included from Mayo Clinic, FL, USA. Thus, 184 cases of pure LN from 180 patients were identified and comprise the study cohort included for analysis. Mean age of patients was 56.5 (range, 34–87) years (Table 1).
Table 1

Characteristics with and without subsequent excision

Total CNBx

Group (n = 184) (%)

Excision (n = 101) (%)

Observation (n = 83) (%)

P

Mean age at diagnosis (year)

56.5 (34–87)

53.5 (34–76)

60.2 (42–87)

<0.0001

ALH

147 (79.9)

81(80.2)

66 (79.5)

0.91

LCIS

37 (20.1)

20 (19.8)

17 (20.5)

Follow-up time (mo)

50.3 (6–212)

47.9 (6–212)

53.0(6–135)

0.19

Imaging finding

 Calcifications

136 (74)

67 (66.3)

69 (83.1)

 

 Mass

39 (21)

27 (26.7)

12 (14.5)

 

 MR enhancement

9 (5)

7 (6.9)

2 (2.4)

 

Biopsy modality

 Stereotactic

138 (75)

67 (66.3)

71 (85.5)

 

 US guidance

34 (18.5)

26 (25.7)

8 (9.6)

 

 MR guidance

12 (6.5)

8 (7.9)

4 (4.8)

 

Needle gauge

 9

40 (21.7)

25 (24.8)

15 (18)

 

 11

103 (56)

49 (48.5)

54 (65)

 

 14

31 (16.8)

23 (22.8)

8 (9.6)

 

 16

1 (0.5)

 

1 (1.2)

 

 18

2 (1.1)

1 (1)

1 (1.2)

 

 Not specified

7 (3.8)

   

Location of calcifications

 LN only

5 (3)

3 (3)

2 (2.4)

 

 LN and benign acini/stroma

53 (29)

32 (31.7)

21 (25.3)

 

 Benign acini and stroma

96 (52)

44 (43.6)

52 (62.7)

 

% Lobular units in CNBx

 1–25 %

130 (70)

70 (69)

59 (71)

 

 26–50 %

40 (21.5)

23 (22.5)

17 (20)

 

 51–75 %

7 (3.8)

4 (3.9)

3 (3.6)

 

 76–100 %

9 (4.8)

5 (4.9)

4 (4.8)

 

Concordance

 Yes

171 (93)

91 (90)

80 (96)

0.1

 No

13 (7)

10 (10)

3 (3.6)

Contralateral cancer at diagnosis

18 (10)

14 (14)

4 (4.8)

0.002

History of contralateral cancer

9 (5)

3 (3)

6 (7.3)

0.35

History of ipsilateral cancer

5 (2.7)

3 (3)

2 (2.4)

0.4

Bold shows statistical significance

CNBx core needle biopsy, ALH atypical lobular hyperplasia, LCIS lobular carcinoma in situ, LN lobular neoplasia

Imaging

Overall, 74 % of CNBx were performed for calcifications, 21 % for a mass or nodule, and 5 % for MRI enhancement (Table 1). This corresponded with the method of biopsy; predominantly stereotactic (75 %, n = 138), followed by ultrasound-guided (18.5 %, n = 34), and MRI-guided (6.5 %, n = 12). Of the stereotactic-guided CNBx performed, 74 % were via vacuum-assisted devices using 11-gauge needles (n = 102), and 16 % (n = 23) were via 9-gauge needles (Table 2). Of the lesions biopsied for calcifications, the number of lesions excised and observed was similar (49.3 vs. 50.7 %). However, a greater percentage of CNBx performed for mass lesions (69 %) and MRI enhancement (78 %) were excised. Of the 27 mass lesions that underwent excision, 16 were deemed discordant by the initial radiologist, the consulting surgeon, or the dedicated breast radiologist reviewing the imaging for this study.
Table 2

Biopsy modality and needle gauge utilized

 

Ultrasound (n = 34)

Stereotactic (n = 138)

MR (n = 12)

Needle gauge

 9

6 (18 %)

23 (16 %)

11 (92 %)

 11

 

102 (74 %)

1 (8 %)

 14

23 (68 %)

8 (6 %)

 

 16

1 (3 %)

  

 18

2 (6 %)

  

Not specified

2 (6 %)

5 (4 %)

 

Pathology

Of the 184 (147 ALH, 37 LCIS) cases of pure LN identified, 101 cases (55 %) underwent excisional biopsy; 81 (55 %) of the ALH cases and 20 (54 %) of the LCIS cases. Calcifications were solely associated with LN in only 3 % of CNBx cases (n = 5). Calcifications were identified in benign ductules and/or stroma in 52 % (n = 96) of cases and present in LN and benign ductules and/or stroma in 29 % (n = 53) of cases. The involvement of TDLUs by LN was less than 25 % in the majority of cases (70 %, n = 130).

Excision Group

The total number of CNBx that underwent surgical excision was 101 (54.8 %) sites in 99 patients. Radiologic–pathologic concordance was demonstrated in 90 % (91/101) of cases. The rate of upstage to cancer on surgical excisions was 2 % (2/101). One of 81 (1.2 %) ALH cases and 1 of 20 (5 %) LCIS cases were upstaged. Of the concordant lesions, only 1 of 101 (1 %) was upstaged to cancer at surgical excision (Table 3).
Table 3

Follow-up events by subtype of lobular neoplasia

 

ALH (%)

LCIS (%)

P

Upstaged at excision

1/81 (1.5)

1/20 (5)

0.36

 Concordant cases only

0/73 (0)

1/18 (5.6)

0.2

Developed ipsilateral cancer

2/112 (1.8)

3/26 (11.5)

0.04

 Observation group

1/51 (2)

3/14 (21)

0.03

 Excision group

1/61 (1.6)

0/12 (0)

0.65

Development of contralateral cancer

3/112 (2.7)

1/26 (3.8)

0.57

Bold shows statistical significance

Case Details of Excision Group Upgraded Cases

Case 1

Diagnostic mammogram demonstrated a 5.0-mm cluster of mildly pleomorphic calcifications, which had increased in number over 1 year at the site of a prior benign surgical biopsy for calcifications. CNBx identified LCIS in 15 % of the TLDUs. These findings were concordant on radiologic–pathologic review. Upon surgical excision, the patient had a 0.9-cm, moderately differentiated, invasive lobular carcinoma (ILC) at the CNBx site. The history is significant for a synchronous contralateral invasive ductal carcinoma (IDC) identified at the time of diagnosis.

Case 2

Screening mammogram demonstrated a focal density with irregular margins. ALH was identified in 1 % of the TDLUs in the CNBx and was discordant at radiologic–pathologic review. A 4.0-mm, low-grade DCIS was identified within the surgical specimen away from the CNBx site (Table 4).
Table 4

Characteristics of upgrades and follow-up events

 

Age

Imaging finding

Biopsy modality

CNB diagnosis

RPC

Event grade

Same quadrant

Family history

Upgrades

 1

54

Calcs on screening

11 g VAB ST

LCIS

Y

9 mm ILC-II on excision

N/a

Nb

 2

51

Focal density with irregular margins

ST, size not known

ALHa

N

4 mm DCIS-I on excision

N/a

N

Ipsilateral cancer

 1

51

Pleomorphic calcs on diagnostic

11 g VAB ST

ALH, excised

Y

9 mm ILC-I at 81 mo

No

Y

 2

81

amorphous calcs on diagnostic

9 g VAB ST

ALH, observed

Y

DCIS-II at 7 mo

Yes

unknown

 3

69

amorphous calcs on screening

11 g VAB ST

LCIS, observed

Y

12 mm DCIS-I at 78 mo

No

Y

 4

55

linear calcs on screening

11 g VAB ST

LCIS, observed

Y

14 mm ILC-II at 37 mo

No

N

 5

63

amorphous calcs on diagnostic

11 g VAB ST

LCIS, observed

Y

16 mm ILC-II at 70 mo

No

N

Contralateral cancer

 1

47

Abnormal diagnostic

11 g VAB ST

ALH, excised

Y

17 mm IDC-II at 7 mo

N/a

Y

 2

64

Calcs on abnormal screening

11 g VAB ST

ALH, observed

Y

IDC-II at 24 mo

N/a

Y

 3

70

Branching calcs on screening

11 g VAB ST

ALH, observed

Y

IDC-I at 25 mo

N/a

unknown

 4

61

Abnormal screening

11 g VAB ST

LCIS, excised

Y

3 mm IDC-II at 36 mo

N/a

N

Family history, positive if at least one first-degree relative or two second-degree relatives with breast and/or ovarian cancer

CNBx core needle biopsy, VAB vacuum-assisted biopsy, ST stereotactic, RPC radiologic–pathologic concordance, ALH atypical lobular hyperplasia, LCIS lobular carcinoma in situ, ILC invasive lobular carcinoma, IDC invasive ductal carcinoma, DCIS ductal carcinoma in situ, calcs calcifications, LN lobular neoplasia, TDLU terminal ductal lobular unit

aOnly 1 % of TDLU involved lobular neoplasia

bSynchronous contralateral breast cancer at time of diagnosis

Observation Group

There were 83 total cases observed without surgical excision (66 ALH, 17 LCIS) in 82 patients. Radiologic–pathologic concordance was demonstrated in 96 % (80/83) of cases. All three patients with discordant lesions were recommended surgical excision. One patient underwent delayed excision in the follow-up period with benign findings, one pursued excision at another institution and was lost to follow-up, and the third patient continued observation and had a stable mammogram at 7 years. The number of patients with follow up data available was 65 (80 %) with a mean follow-up time of 53 months.

Development of Ipsilateral Cancer in all Patients

To evaluate the events in the follow-up period for all patients, the following patients were excluded from analysis: patients who underwent bilateral mastectomy at the time of diagnosis (n = 16), patients upgraded to carcinoma on excision (n = 2), and patients with less than 6 months of follow-up (n = 25). Of the bilateral mastectomy group, six were prophylactic and ten were secondary to a synchronous contralateral breast cancer. Five of the remaining 138 (3.6 %) patients developed ipsilateral breast cancer during the follow-up period. This included 1.4 % (1/73) of patients that had undergone excision compared with 6.2 % (4/65) of patients who underwent observation (P = 0.19). In patients with ALH, 2 % (1/51) in the observation group and 1.6 % (1/61) in the excision group developed an ipsilateral malignancy. In patients with LCIS, 21.4 % (3/14) of patients observed, and no patients who underwent surgical excision, developed an ipsilateral malignancy (Table 3).

Case Details of Development of Ipsilateral Cancer

One patient in the excision group (case 1) and four patients in the observation group (cases 2–5) developed an ipsilateral breast cancer. All five patients demonstrated radiologic–pathologic concordance on CNBx (Table 3).

Case 1

Indication for CNBx was pleomorphic calcifications on screening mammogram and pathology revealed ALH. At surgical excision, no higher risk lesion was identified. At 81 months, the patient developed an ipsilateral ILC in a different quadrant than her initial CNBx.

Case 2

Patient was diagnosed with ALH on CNBx obtained for amorphous calcifications. Calcifications were in benign ducts only. Six-month follow-up mammogram demonstrated an increase in the size and a change in morphology of the calcifications. Surgical excision at 7 months revealed DCIS.

Cases 3–5

The remaining three patients had LCIS on CNBx for calcifications. Two of these patients had calcifications in benign ducts only. One patient developed DCIS at 78 months, and two patients developed ILC at 37 and 70 months. All events were in a different quadrant than initial CNBx.

Development of Contralateral Cancer in all Patients

Of the 138 patients, 4 (2.9 %) patients developed contralateral invasive breast cancer (Table 3). Three of 112 (2.7 %) patients with ALH developed contralateral (IDC) at 7, 24, and 25 months, respectively. One of 26 (3.8 %) patients diagnosed with LCIS on CNBx developed a contralateral IDC at 36 months.

Chemoprevention

Patients not eligible for chemoprevention included patients who had undergone bilateral prophylactic mastectomy at diagnosis (n = 6), contralateral cancer at time of diagnosis (n = 21), upgrade to carcinoma (n = 2), and patients with a history of breast cancer and prior use of endocrine therapy. Of the eligible patients, 30 of 120 (25 %) had documented use of chemoprevention. The proportion of patients on chemoprevention in the observation group versus the excision group was lower (21 vs. 28 %, P = 0.37). No patients managed with chemoprevention after CNBx developed an ipsilateral or contralateral malignancy during the follow-up period.

Discussion

Our data support observation in select patients diagnosed with pure LN on CNBx. Particularly, in patients whose CNBx demonstrate pure ALH without any associated high-risk lesions or ipsilateral malignancy at the time of diagnosis with radiologic–pathologic concordance. LN is usually an incidental pathologic finding at CNBx and lacks distinctive features on physical examination or imaging.810 Reported incidences range from 0.02–3.8 % of biopsy specimens.1114 The incidence of LN in CNBx in our series was 3 %. The most frequent mammographic abnormality prompting biopsy is reported to be calcifications, which is consistent with our findings, although calcifications are rarely associated with LN alone.1519

The management of LN demonstrated on CNBx continues to be controversial. The rates of upstage on surgical excision range from 2–40 %.11,17,1923 Much of this variation arises from small case numbers, heterogeneity of lesions included, and lack of careful radiologic–pathologic concordance. Many studies suffer from selection bias by simply reporting on the group of patients with LN who underwent excisional biopsy and do not report outcomes of lower risk lesions that are observed. In our unique subset of patients with pure LN, the rate of upstage to cancer on surgical excision was 2 %. After excluding the one discordant lesion, the rate of upstage was 1 %. This is consistent with previous studies that have evaluated pure LN with meticulous review of radiologic–pathologic concordance.19,20,24 Hwang et al.20 reported 136 excisional biopsies with LN. After excluding cases that were discordant and those associated with high risk or nonclassic pathology, the upgrade rate for pure LN was 1 %. During a mean follow-up period of 49 months, the development of ipsilateral and contralateral malignancy was 2 and 0.7 %, respectively, in a group of 148 cases observed without surgical excision. Nagi et al.19 reported a series of 98 patients with pure LN. Of the 45 patients who underwent surgical excision, 2 (4 %) were upstaged to malignancy (1 patient had a separate minute focus of ILC, the other with DCIS).

The risk of the development of an ipsilateral or contralateral breast cancer is reported to be four- to fivefold elevated in patients diagnosed with ALH, and eight- to tenfold elevated in patients with LCIS.2527 In our cohort of patients with long-term follow-up, 3.7 % (5/138) patients developed an ipsilateral malignancy and 2.9 % (4/138) of patients developed a contralateral malignancy. The low rate of events in the follow-up period may be partially accounted for by the use of chemoprevention in 25 % of the eligible patients. Although we have demonstrated a low rate of events in our cohort, all patients with biopsy proven LN have a well-recognized risk of subsequent cancer in either breast and should be followed carefully.9,20,28

The natural history of LN is poorly understood. The evidence that LCIS is a precursor lesion for invasive cancer arises from studies that have demonstrated that patients with a diagnosis of LCIS have a significantly greater risk of developing ipsilateral breast cancer, and often in the same quadrant.29,30 In addition, shared molecular alterations have been demonstrated between LCIS and synchronous ILC.3134 In this study, the incidence of developing an ipsilateral cancer in all patients was 3.6 %, and four of five (80 %) cancers occurred in a different quadrant. The low number of events and lack of molecular phenotyping do not allow us to draw conclusions regarding the pathway of LN; however, this clinical outcome data do not appear to be consistent with an obligate precursor model.

Many authors have tried to identify imaging, histopathologic features, and CNBx variables to predict which lesions should be surgically excised.12,17,35,36 Rendi et al.36 reported an upstage rate of 4.4 % in a series of 68 cases of pure LN that underwent surgical excision. Upstaged patients were clinically high-risk, had discordant imaging, or extensive LN. They suggested that the extent of LN on the CNBx correlated with the risk of upgrade and indicated that patients with a CNBx of less than four foci of pure LN that are not otherwise at high risk may be safely observed. On their review of literature including 17 prior studies,11,12,1820,2224,3544 the upgrade rate ranged from 1–38 %. Of the 103 patients that were upgraded, 63 (61 %) had masses, nonclassic LCIS, or radiologic–pathologic discordance. One of the most critical elements in the management of LN on CNBx is careful correlation between the radiologic image and associated histopathologic findings. The overall radiologic–pathologic concordance in our study was 93 %. The observation group contained a greater percentage of concordant lesions (96 vs. 90 %).

In summary, our study provides information regarding the management of LN from a large number of cases, with careful radiologic–pathologic review, and long-term follow-up. Although the retrospective nature and lower number of patients with LCIS remains a limitation of this study, these data support that patients with pure ALH, demonstrating radiographic–pathologic concordance and no associated high-risk lesions found on CNBx can be safely observed. There is likely a subset of patients with pure LCIS that can be observed; however, we cannot make definite conclusions in this regard. Prospective studies are necessary to further elucidate the clinical and molecular behavior of these lesions. We do recommend that LN should be excised if there is radiologic–pathologic discordance, an associated high-risk lesion, or nonclassic pathology. Chemoprevention should be considered and careful surveillance is imperative in all patients, because they harbor an increased risk of development of cancer in either breast.

Copyright information

© Society of Surgical Oncology 2012