Article Text

Download PDFPDF

Lymph-node metastases in invasive lobular carcinoma are different from those in ductal carcinoma of the breast
  1. Beatriz Fernández,
  2. E Claire Paish,
  3. Andrew R Green,
  4. Andrew H S Lee,
  5. R Douglas Macmillan,
  6. Ian O Ellis,
  7. Emad A Rakha
  1. Departments of Histopathology and Surgery, Nottingham University Hospitals NHS Trust, The University of Nottingham, Nottingham, UK
  1. Correspondence to Dr Emad A Rakha, Department of Histopathology, Nottingham University Hospital NHS Trust, City Hospital Campus, Hucknall Road, Nottingham NG5 1PB, UK; emadrakha{at}


Aim Invasive lobular carcinoma (ILC) and invasive ductal carcinoma (IDC) of the breast are distinct morphological entities with different biological features and clinical behaviour. In the present study, the authors compare the axillary-lymph-node (ALN) status of patients with grade-matched ILC (no=426) and IDC (no=820). The pattern of nodal metastatic deposits (nodular, sinusoidal and diffuse) and the proportion of involved nodes were also analysed in a selected group of 246 tumours, which were associated with a single positive ALN.

Results Compared with grade-matched IDC, ILC was associated with a higher nodal stage (13.1% vs 4.5% of ILC and IDC were stage 3), higher absolute number of positive nodes and higher ratio of positive nodes (0.46±0.30 and 0.33±0.23 in ILC and IDC respectively). These differences were maintained in the different size subgroups. The most common metastatic morphological pattern was nodular in both types of carcinomas. A sinusoidal pattern was more frequent in IDC, and the diffuse pattern was more frequent in ILC. Despite these differences, ILC and grade-matched IDC exhibited similar rates of regional recurrences (RR) and breast-cancer survival.

Conclusion This study provides clinical evidence which further demonstrates that ILC and IDC are biologically distinct entities with different lymph-node involvement patterns and ILC having a tendency to metastasise to more nodes than IDC. However, this difference was not associated with a significant impact on patient outcome.

  • Breast
  • molecular pathology
  • molecular biology
  • endocrinology
  • breast cancer
  • steroid receptors
  • tamoxifen
  • PCR
  • histopathology
  • immunocytochemistry
  • tumour angiogenesis
  • breast pathology
  • angiogenesis
  • cytology
  • diagnosis
  • cancer
  • genetics
  • oncology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Breast cancer is a heterogeneous disease with varied morphology, behaviour and response to therapy. The most frequently observed types of invasive breast cancer are infiltrating ductal carcinoma (IDC) and infiltrating lobular carcinoma (ILC), accounting for more than 90% of all invasive breast carcinomas, with an incidence of 72–80% and 5–15% respectively.1–5

Several studies have reported that ILC is a distinct entity of breast cancer, with distinct biological features. Compared with IDC, ILC is associated with older age, lower histological grade and less vascular invasion, and it is more frequently hormone-receptor-positive and HER2-negative. ILC is associated with a larger tumour size and shows a higher rate of multiple metastases with a distinct pattern of involvement of distant sites.1 3 4 6–15

Although most studies reported no differences between ILC and IDC carcinomas regarding axillary-lymph-node (ALN) status,3 4 8 14 16–18 some authors have reported an association between ILC and less frequent ALN positivity,1 and some studies have also demonstrated that ILC is more likely to be ALN-positive than IDC.15 19 This inconsistency has been attributed to variations in ALN staging techniques, which have yielded different numbers of dissected ALN.20 It is well documented that ALN involvement and the number of involved ALN are among the most important prognostic factors in early-stage breast cancer.21 22 Patients with four or more positive ALN (tumour-node-metastasis (TNM) nodal stage pN2 and pN322 or Nottingham Prognostic Index (NPI) nodal stage 323) have a worse prognosis than those with fewer than four positive nodes.24 25 However, the biological rationale of this cut-off of four nodes remains unclear.26 The AJCC TNM staging system for breast cancer has confirmed that the absolute number of ALN involved by metastatic cancer is one of the most important prognostic factors in breast-cancer staging.22 27 Several series reported that the ratio of positive ALN (ratio of involved nodes divided by the total number of removed nodes; LNR) may be a useful prognostic factor for predicting overall survival, disease-free survival, risk of loco-regional recurrence (RR) and development of metastatic disease.28–30 In these studies, an LNR higher than 0.20–0.30 was associated with a worse prognosis. These findings support the suggestion that not only the ALN stage or the absolute number of positive ALN but also the ratio of positive ALN should be considered in specific adjuvant treatment decisions.20 26 31–34 Despite this observation, the potential differences between ILC and IDC regarding the percentage of positive ALN have not been previously studied.

In a previous proof-of-principle study,35 we demonstrated that metastatic growth in ALN occurs in three distinct patterns: sinusoidal, nodular and diffuse. The preliminary results suggest that metastatic patterns may have prognostic significance with the sinusoidal pattern being related to a poor prognosis and the diffuse pattern with a good prognosis. However, no reliable conclusions were provided regarding the most common patterns in ILC and IDC tumours, owing to the small number of cases.

The aim of this study was to perform a comprehensive comparison of ALN status in ILC and IDC tumours, taking into account the absolute number, the percentage and the metastatic morphological pattern of positive ALN in both types of carcinomas.

Materials and methods

The study population was derived from the Nottingham Tenovus Primary Breast Carcinoma Series from women aged 70 years or less who presented with primary operable invasive breast carcinomas (Stage I and II) between 1989 and 2004. This is a well-characterised series of patients who have been treated in a single institution.4 36 37 All patients received standard surgical treatment of either mastectomy or wide local excision with radiotherapy. Most patients underwent initial axillary-node sampling (1–5 nodes) with or without completion axillary clearance. Patients with one to three positive lymph nodes on axillary node sampling were offered irradiation of the axilla as a standard protocol (in addition to appropriate systemic therapy). Cases with four positive lymph nodes were offered axillary-node clearance. Of the whole series, 41% underwent axillary clearance.

Only patients with a histological diagnosis of pure ILC or pure invasive ductal carcinoma (IDC) of no special type were included. From the whole series (4620 cases), 426 ILC cases were identified. In addition, a group of grade-matched IDC cases with ALN status available (820) were included (table 1). Tumour characteristics including primary operation type and number and type of ALN surgery, primary tumour size, histological tumour type,38 histological grade,37 39 vascular invasion (VI),40 LN status, oestrogen receptor and progesterone receptor were obtained from the database. Data on ALN status included LN stage (TNM pathological nodal stage; pN0=negative, pN1=1–3 positive regional nodes, pN2=4–9 positive nodes and pN3=10 or more nodes are positive22 or the NPI nodal staging; stage 1=negative, stage 2=1–3 positive nodes, stage 3=4 or more positive nodes23), the total number of ALN removed, the absolute number of positive ALN and the percentage of positive ALN. ILC was not further subtyped in this study. Clinical follow-up data including RR, survival time and disease-free interval were available in a subset of cases (1107 patients).4 36 37 41 Patients were followed up at 3-month intervals initially, then 6-monthly and annually for a median period of 83 (range 1–369 months). RR was defined as a tumour arising in the draining lymph nodes. Breast-cancer-specific survival (BCSS) was defined as the interval between the operation and death from breast cancer; death being scored as an event, and patients who died from other causes or were still alive were censored at the time of last follow-up.

Table 1

Pathological features of invasive lobular carcinomas compared with grade-matched invasive ductal carcinoma of no special type

In addition, for assessment of ALN metastasis patterns, a series of 246 cases including 219 IDC and 27 ILC, known to contain metastases (macrometastatic type) in a single sentinel lymph node (SLN) only, were included. None of these cases showed extracapsular nodal spread. This cohort included T1 and T2 tumour (size ranging from 0.5 to 5 cm, median 2 cm). The immunohistochemistry and bright-field microscopic analysis of the metastatic pattern were performed as previously described.35 42 Briefly, from each of the 246 cases, all formalin-fixed, paraffin-embedded (FFPE) SLN blocks were sectioned at 4 μm onto charged slides (X-tra Adhesive micro slides, Surgipath Europe Ltd), dried overnight at 37°C and then stained using standard Avidin–Biotin Complex (ABC; Dako Ltd, Denmark). The primary antibody used was the broad spectrum cytokeratin mouse monoclonal antibody AE1/AE3 (Dako Ltd, Denmark), optimally diluted at 1/100 in normal swine serum. Biotinylated secondary antibody and preformed ABC preceded visualisation using the chromogen 3′3-diaminobenzidine (DAB) solution.

Each IHC-stained section was analysed using bright-field microscopy, in order to assess the metastatic growth patterns present. Three different growth patterns were noted within the LNs; sinusoidal, where the tumour cells were present within the subcapsular sinus of the ALN; nodular, where the tumour cells formed a clump, or nodule, of tightly packed cells in the cortex and and/or medulla of the node; and diffuse, where the tumour cells were dispersed through the cortex and/or medulla of the node in a diffuse spreading pattern. In some sections, only one growth pattern was seen, whereas in others they were seen in combination.35 If a pattern combination was present, the predominant pattern was noted. In addition to recording which metastatic patterns were present, the proportion of each section involved with a tumour (tumour burden (TB)) was noted. Results were recorded in categories of increasing TB: <10%, 10–25%, 26–50%, 51–75%, 76–90% and >90% respectively.

Statistical analysis

A statistical analysis was performed using PASW 18.0 statistical software (SPSS, Chicago, Illinois). To assess differences in tumour and ALN pathological factors between ILC and IDC contingency tables, χ2 tests and Mann–Whitney tests were used. BCSS and RR curves were drawn using Kaplan–Meier estimates and were compared using logrank tests. Statistical significance was defined as a p value of <0.05.


Table 1 summarises the pathological features of ILC (426 cases) compared with grade-matched IDC (820 cases). ILC is associated with a larger size but less vascular invasion and hormone receptor positivity than with grade-matched IDC (p<0.0001). ALN status was available in all patients. Despite the lower frequency of vascular invasion, ILC showed more positive ALN (39.2%) than grade-matched IDC (33.0%) (p=0.031). When ALN was compared as three stages (stage 1, 2 and 3) the difference was highly significant (p<0.0001), mainly due to the high frequency of stage 3 ALN disease in ILC when compared with IDC (13.1% and 4.5% respectively).

The number of ALN sampled was slightly higher in ILC compared with IDC (p=0.043). Among the node-positive patients, the median number of involved nodes was two (range 1–38) in ILC tumours compared with one (range 1–20) in IDC (p<0.0001). The difference in the absolute number of positive nodes was more obvious in stage three tumours than in stage 2. In stage 3 tumours, the median number of positive nodes was seven nodes in ILC (range 4–38) compared with five nodes in IDC (range 4–20), while in stage 2 the median was one node (range 1–3) in both ILC and IDC. A total of 4.2% of ILC showed ≥10 positive nodes (TNM stage pN322) compared with 0.2% in IDC (p<0.0001).

In this study, as the standard axillary procedure was a node sample, some cases showed a low number of ALN examined. Therefore, the analyses were repeated after exclusion of cases with fewer than four ALN sampled (251 cases; 85 ILC and 166 IDC). Similar associations were obtained between ILC and higher LN stage (χ2=30.4, df=2, p<0.0001) and absolute number of ALN (Z=3.2, p=0.001). The ratio of positive nodes to the total number of removed nodes was significantly higher in ILC than in IDC (mean ratio 0.46±0.30 and 0.33±0.23 in ILC and IDC respectively) (p<0.001). Importantly, these ALN differences between both tumour types were maintained when they were matched according to size subgroups (pT1 (<2 cm) and pT2 (2–5 cm)). The difference between ILC and IDC with regard to the number of removed uninvolved ALN showed an association of borderline significance (Z=1.9, p=0.063) between ILC and lower number of uninvolved nodes (mean 4.5±4.1).

Analysis of the nodal metastatic patterns showed that the sinusoidal pattern was the most frequent (93%) followed by the nodular pattern (89%), whereas the diffuse pattern was the least frequently seen (38%). When the predominant pattern was considered, the nodular was the most frequent (73% as the main predominant pattern and 8% in combination with other patterns). There was an association between primary tumour size and pattern of nodal metastasis (χ2=11.5, df=3, p=0.009) in which a larger size is associated with nodular, nodular and diffuse patterns, whereas a smaller size is associated with the predominantly diffuse pattern. No association was found between metastatic pattern and tumour histological grade or presence of vascular invasion (p>0.05).

ILC showed a predominant diffuse metastatic pattern or predominant diffuse and nodular pattern more frequently than IDC. While none of the ILC cases showed a predominantly sinusoidal pattern (0%), this was seen in 11.9% in IDC (p=0.028). Although a predominantly nodular pattern was the most frequent in both tumours, it was seen less in ILC (66.7% and 74.0% in ILC and IDC respectively) (table 2). ILC metastatic nodal deposits tend more frequently to be a combination of the three patterns (nodular, diffuse and sinusoidal; SND) (70.4%), while IDC deposits tend more frequently to be sinusoidal and nodular in combination (39.7%). Of note, six of seven cases of ILC with definite vascular invasion showed a nodular metastatic pattern, and the remaining case showed a nodular and diffuse pattern. No such association was identified in IDC.

Table 2

Overall predominant pattern in invasive lobular carcinomas and invasive ductal carcinoma subtypes

In addition to the pattern of metastatic deposits, the proportion of nodal involvement or metastatic TB was assessed (Table 3). The majority of cases (61%) had a low TB (<50%), and 28% showed a TB of <10%. There was an association between TB and metastatic tumour patterns (p<0.001). IDC showed a lower (<10%) TB (29.7% and 14.8% in IDC and ILC respectively), while ILC was associated with a higher (>50%) TB (37.9% and 48.1% in IDC and ILC respectively). However, these differences were not statistically significant (p=0.317).

Table 3

Percentage of tumour metastatic deposits to the whole size of the involved lymph node (tumour burden) in both invasive lobular carcinomas and invasive ductal carcinoma subtypes


In this study, 33 patients developed RR in the lymph node (3.1%) during a median period of 70 months (range 9–226 months). In the whole series, there was no association between development of RR and ALN status when analysed as ALN stages (3.1%, 2.7% and 5.7% of patients with NPI nodal stage 1, 2 and 3 developed RR respectively), absolute number or ratio (p>0.05). However, there was a strong association between ALN status and BCSS when analysed as ALN stage (χ2=59.2, p<0.0001), absolute number or ratio (p<0.001).

When ILC was compared with grade-matched IDC, no statistical difference was identified regarding the rate of development of RR in the whole series (χ2=2.43, p=0.119) or the ALN-negative or -positive subgroups (p>0.05). Similarly, no difference was identified between ILC and IDC with regard to BCSS in the whole series or in the different nodal subgroups (p>0.05).

In the whole series, there was an association between pattern of metastasis and BCSS (χ2=12.68, p=0.005) in which both a nodular and nodular and diffuse pattern were associated with a shorter survival, while the diffuse pattern was associated with a longer survival (figure 1). Although this association was retained in IDC (χ2=12.40, p=0.006), it was not observed in ILC (χ2=1.17, p=0.555), which may be related to the few events occurring in ILC during the follow-up period (three events); and similarly an association between TB and BCSS in the whole series (χ2=13.85, p=0.017) and in IDC (χ2=14.82, p=0.017) but not in ILC (χ2=4.79, p=0.441) in which only three events occurred during the period of follow-up.

Figure 1

Association between different metastatic patterns in the lymph node and breast cancer and specific survival (BCSS) in the whole series (χ2=12.68, p=0.005).


The single most important predictor of outcome for woman with early-stage breast carcinoma is the status of ALN,22 43 44 but the impact of the histological primary tumour type on the risk and nature of ALN involvement remains uncertain.17 18 Previous studies have shown conflicting results regarding the clinical and biological difference between ILC and IDC.1 3 4 8 14–19 45 46 Although most studies did not find any significant differences in ALN status between ILC and IDC, some authors reported a higher rate of ALN positivity15 19 and false-negative ALN by ultrasound examination47 or histological examination48 in ILC. Moreover, ILC has special characteristic features that are associated with cell dyshesion, motility and metastasis, particularly loss of function of the cell adhesion molecule E-cadherin.49 50 Loss of expression of E-cadherin is associated with reduced cell adhesiveness and facilitates the infiltration in regional nodes and distant sites.1 51 52

Previous comparative studies of ILC have typically included unselected series of IDC which include >50% grade 3 which contrasts with the low frequency (10%) of grade 3 disease in ILC.4 37 In addition, the difference between ILC and IDC regarding the absolute number of positive ALN, ratio of positive nodes, pattern of nodal metastasis and development of tumour recurrence in the regional nodes remains unclear. Therefore, the aims of the current study were to determine the ALN status in ILC compared with a grade-matched group of IDC. Our findings demonstrate that the histological subtype of the primary tumour can influence the likelihood of ALN metastasis and is associated with the pattern of metastatic deposits.

This study showed that when ILC metastasises to ALN, it tends to involve more nodes at the time of initial surgery. The ratio of positive nodes as compared with the total number of removed nodes (ALN ratio) is also higher in ILC than IDC. This effect was maintained when both tumour types were matched for size. Compared with IDC, ILC is associated with distinct patterns of nodal metastasis; it was more likely to produce a diffuse pattern of metastasis with or without associated nodular pattern, but less likely to produce a sinusoidal pattern. These differences are consistent with previous studies which demonstrated a difference between ILC and IDC at the molecular and genetic levels and that they are biologically distinct tumours.3 10 53 54 In addition to loss of E-cadherin function, ILC has a lower proliferative activity3 10 and shows distinct genomic and expression profiles with significant difference in the expression of other genes involved in cell growth, cell adhesion/motility and immune response54 55 compared with IDC. Furthermore, ILC is associated with a higher rate of multiple distant metastases11 and with a distinct pattern of involvement of distant sites.1 6 56 57

The indolent nature of ILC might possibly lead to a longer period of growth before detection with an increased likelihood of involving more nodes. The lack of associated desmoplastic reaction may also make ILC more difficult to detect than other forms of breast cancer, which could partly explain the involvement of more nodes at the time of diagnosis than IDC.58 The current study, in agreement with previous reports,3 4 14 59 demonstrates that the tumour size is larger in ILC than in grade-matched IDC. It has been demonstrated that there is an association between increasing tumour size and increasing number of positive ALN,18 24 and that the frequency of sentinel node involvement increases with increasing tumour size in ILC.60 However, the nodal difference between ILC and IDC observed in this study was maintained in the size-matched groups indicating that size is not the main reason for this difference. Moreover, despite the higher degree of nodal involvement, vascular invasion was less frequently seen in ILC. A similar difference in vascular invasion has been described in a non-grade-matched series.61 62 A possible explanation for this difference may be that tumour emboli are smaller and more difficult to identify in ILC.

In this study, the absence of association between ALN status and RR is observed in the whole series, and the lack of differences between ILC and IDC may be related to the low frequency of RR (3%) and the use of local radiotherapy as well as systemic therapy. More ILC patients (55%) received hormone therapy than did IDC patients (43%). Although, in this study, the biological difference between ILC and IDC regarding ALN status did not translate into a clinically meaningful difference in terms of RR or BCSS, this may be related to the small number of ALN-positive ILC with follow-up data. Larger and more detailed studies are required to confirm these results and to determine any associations with clinical outcome.

This study also confirms the results of our previous proof-of-principle three-dimensional reconstruction study of ALN metastasis in a small number of cases.35 An association between pattern of metastasis and patient outcome in terms of BCSS was identified. Both nodular and nodular and diffuse patterns were associated with a shorter survival, while the diffuse pattern was associated with a longer survival. The nodular metastatic growth pattern was the most frequent pattern in both tumour types either as the predominant pattern or associated with other patterns. Furthermore, we observed that the sinusoidal pattern is more common in IDC and that the diffuse pattern is more common in ILC. These findings demonstrate that the ALN metastatic morphological pattern for ILC differs from that of IDC. However, as the selection criteria used to collect cases to compare metastatic patterns have resulted in a limited number of ILC cases, these results may be interpreted with caution. Analysis of the effect of size showed an association between a smaller size and the diffuse pattern of nodal metastasis, which is the pattern seen more frequently in ILC. This may argue against size as a confounding factor for the association between tumour type and pattern of nodal metastasis. There was an association between TB and outcome in the whole series and in IDC, but this association was not found in ILC, which could be related to the small number of ILC with one positive node and the few events that took place during the period of follow-up (3/27).

In conclusion, our findings add further evidence to indicate that ILC and IDC are biologically distinct entities with different ALN growth patterns. Moreover, ILC tends to metastasise to more nodes than IDC, but this does not translate into a difference in breast-cancer survival.

Take-home messages

  • ILC is a distinct entity of breast cancer that is different from the most common IDC at the genetic, molecular, histological and immunophenotypic pattern levels.

  • Previous studies have shown controversial results regarding the clinicopathological difference between ILC and IDC.

  • This study shows that when metastasised to lymph nodes, ILC tends to involve larger number of nodes and produce a distinct pattern of nodal metastasis compared with grade-matched IDC. However, this biological difference is not associated with a significant impact on outcome.

Interactive multiple choice questions

This JCP article has an accompanying set of multiple choice questions (MCQs). To access the questions, click on BMJ Learning: take this module on BMJ Learning from the content box at the top right and bottom left of the online article. For more information please go to: Please note: the MCQs are hosted on BMJ Learning the best available learning website for medical professionals from the BMJ Group. If prompted, subscribers must sign into JCP with their journal's username and password. All users must also complete a onetime registration on BMJ Learning and subsequently log in (with a BMJ Learning username and password) on every visit.



  • Competing interests None.

  • Ethics approval Ethics approval was provided by the Nottingham Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.