Article Text

Download PDFPDF

Granulomatous mastitis: the histological differentials
  1. Maribel Lacambra1,
  2. Tu Anh Thai2,
  3. Christopher C F Lam1,
  4. Alex M C Yu1,
  5. Huong Thien Pham2,
  6. Phuong Viet The Tran3,
  7. Bonita K B Law4,
  8. Thanh Van Nguyen2,
  9. Dung Xuan Pham2,
  10. Gary M Tse1
  1. 1Department of Anatomical and Cellular Pathology, Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong, China
  2. 2Department of Pathology, Ho Chi Minh City Oncology Hospital, Ho Chi Minh City, Vietnam
  3. 3Department of Breast Surgery, Ho Chi Minh City Oncology Hospital, Ho Chi Minh City, Vietnam
  4. 4The Breast Center, Union Hospital, Hong Kong, China
  1. Correspondence to Dr Gary M Tse, Department of Anatomical and Cellular Pathology, Prince of Wales Hospital, Ngan Shing Street, Shatin, NT, Hong Kong SAR, China; garytse{at}


Background The management of granulomatous mastitis depends on the causative factor, and accurate diagnosis in distinguishing between idiopathic granulomatous mastitis (IGM) and tuberculous mastitis (TBM) is indispensable. This is particularly problematic in the cases of granulomatous mastitis in which the microbiological studies are negative. In this study, in a large cohort, the histological features for IGM and TBM were compared.

Methods The histopathology files from the two participating hospitals were searched for cases of granulomatous inflammation of the breast over an 8-year period. The parameters assessed included age of patient, lesional size, systemic and local symptoms, and histological findings of inflammatory cells, granulomas, necrosis, multinucleated giant cells, fibrosis and calcifications.

Results 29 cases of IGM and 33 cases of TBM were included in this study. A significant difference was seen between the two groups with regard to patient age (t=2.52, p<0.05) and lesional size (t=−5.56, p<0.01). TBM occurred in a significantly younger population, and demonstrated larger lesional sizes than IGM. There was no difference between the number of cases showing mass, local and systemic symptoms. Comparing the different histological features, the TBM group showed significantly more fibrosis, eosinophils and necrosis, whereas the IGM group showed significantly more plasma cells. Taking all the cases together as one group to evaluate the relationship between the histological parameters, there was significant positive correlation between eosinophils and fibrosis (rs=0.39, p<0.01), and negative correlation between vague and well-formed granulomas (rs=−0.38, p<0.01).

Conclusion TBM was more likely to occur in younger patients, with a larger clinical mass at presentation. Histologically, TBM tends to show more eosinophils and necrosis, and IGM is associated with more plasma cells. The characteristics of the granulomas and giant cells were not distinguishing features.

  • Breast cancer
  • breast pathology
  • core needle biopsy
  • fine needle aspiration cytology
  • granulomatous mastitis
  • idiopathic granulomatous mastitis (IGM)
  • tuberculosis mastitis (TBM)

Statistics from


Granulomatous mastitis is a rare clinical entity that often mimics breast cancer. The clinical presentation is variable, usually affecting women of childbearing age. Patients may present with progressively enlarging breast lumps of variable size that are firm, tender and unilateral, and can be fixed to the skin causing nipple retraction, or fixed to underlying pectoralis muscle.1 2 The radiological finding of granulomatous mastitis is of an ill-defined mass that is difficult to distinguish from that of carcinoma and other granulomatous diseases. The frequent findings on mammogram and ultrasound are asymmetric, diffuse, increased density, hypoechoic mass lesions, or nodular structures.3 The aetiology is multiple and can be infectious or non-infectious. Idiopathic granulomatous mastitis (IGM, or lobular granulomatous mastitis, granulomatous lobulitis) is a rare non-infectious chronic granulomatous inflammatory process that was first described in 1972.4–6 The diagnosis of IGM is difficult clinically, and the exclusion of other possible causes of granulomatous inflammation is essential. Among the infectious causes, tuberculous mastitis (TBM) caused by Mycobacterium tuberculosis is probably the most common. It is considered more likely that there is secondary infection of the breasts in most cases, even when the presumed primary focus is clinically inapparent.7 8 Histologically, irrespective of whether it is infectious or not in aetiology, granulomatous mastitis shows epithelioid histiocytes, plasma cells, lymphocytes, eosinophils, neutrophils, multinucleated giant cells and necrosis.2 9 The management of granulomatous mastitis depends on the causative factor and, in the case of IGM, steroid therapy can be effective; on the other hand, steroid therapy can exacerbate TBM.10 Hence, accurate diagnosis to distinguish between IGM and TBM is indispensable. This is particularly problematic in cases of granulomatous mastitis in which microbiological studies are negative. Some authors consider that demonstration of acid-fast bacilli is not mandatory in the diagnosis of TBM,11 because in cases of pre-operative fine needle aspiration cytology or core needle biopsy, the demonstration of acid-fast bacilli requires material containing a concentration of 10 000–100 000 bacteria/ml.12 In cases without demonstrable acid-fast bacilli, whether the patients should be treated for TBM or IGM poses a difficult management dilemma. Would the histopathological features be useful in this differentiation? In the literature, little has been published comparing infectious and non-infectious granulomatous mastitis with respect to their individual histopathological characteristics. In this study, we compare a large cohort of patients with IGM and TBM to evaluate the potential differentiating histological features.

Materials and methods

The histopathology files from the two participating hospitals (Prince of Wales Hospital, Hong Kong, and Ho Chi Minh City Oncology Hospital) were searched for granulomatous inflammation of the breast over an 8-year period. All specimens were formalin fixed and routinely processed, and H&E-stained sections (4 μm thick) were examined. Ziehl Neelsen stain was also performed for all cases, and selected cases were stained with Wade-Fite stain, periodic acid–Schiff with (PASD) and without (PAS) diastase, Gram stain and Grocott's methenamine silver (GMS) for fungi and bacteria. The inclusion criterion was cases in which all these microbiological studies on the histological sections were negative and diagnosis was confirmed by treatment outcome.

All the slides were analysed by three of the authors in a semiquantitative manner, blinded to the diagnosis. Any discrepancies were discussed and resolved by using a multihead microscope. The parameters assessed included the following criteria. (1) Inflammatory cells including lymphocytes, plasma cells, neutrophils, eosinophils and histiocytes. All inflammatory cells were scored as 0, 1, 2 or 3 based on the following criteria of the relative percentages among all inflammatory cells: (a) for histiocytes, lymphocytes, neutrophils and plasma cells: 0, none; 1, <20%; 2, 20–50%; and 3, >50%; (b) for eosinophils: 0, none; 1, <15%; 2, 15–30%; 3, >30%. (2) Granulomas were classed as either vague (or poorly defined, evolving granulomas) or well-formed (mature granulomas, clearly demarcated from the adjacent background) (figures 1 and 2). (3) Necrosis was observed as either absent or present. (4) Multinucleated giant cells were classed as either foreign body type (individual nuclei being arranged haphazardly) or Langhans type (individual nuclei being arranged in a horseshoe pattern). (5) Fibrosis was scored from 0 to 3 (0, no fibrosis; 1, focal fibrosis; 2, moderate fibrosis; 3, extensive fibrosis). (6) Calcifications were observed as either present or absent.

Figure 1

Photomicrograph showing a vague granuloma (H&E, ×20).

Figure 2

Photomicrograph showing a well-formed granuloma (H&E, ×20).

The medical records were retrieved to assess the clinical presentation (the presence of mass lesion, fever and other systemic symptoms), microbiological studies, treatment and co-existing morbidities, including pulmonary tuberculosis. A diagnosis of TBM was made by clinical response to antituberculosis therapy. Those cases that subsided without specific antituberculosis treatment were diagnosed as IGM.

Statistical analysis

Mann–Whitney U test was used to differentiate the histological parameters between the IGM group and the TBM group. Spearman's correlation test was used to analyse the relative correlations between each histological parameter. All statistical tests were done using SPSS 16.0 (SPSS, Chicago, Illinois, USA) and statistical significance was established at p<0.05.


A total of 62 cases from 62 patients of granulomatous mastitis were included in this study, including 29 cases of tuberculosis and 33 non-infectious cases. All patients were female, with an overall age range of 21–56 years (mean 35.49 years). Thirty-seven breast lesions were found on the right side and 25 lesions were found on the left side; the lesional sizes ranged from 1 to 15 cm (mean 4.42 cm) (table 1). In all cases, there were no systemic vasculitic changes, and no clinical evidence or history of other forms of granulomatous inflammation. Special stains for organisms including PAS, PASD, GMS, Wade-Fite and Ziehl-Neelsen were all negative.

Table 1

Summary of medical details from patients with idiopathic granulomatous mastitis and tuberculous mastitis

In the 29 cases of TBM, the age range of the patients was 21–56 years (mean 32.71 years). The lesional sizes ranged from 2 to 15 cm (mean 6.04 cm). Systemic symptoms included five cases of fever (17%), while 24 (83%) cases demonstrated no symptoms. Local symptoms included mass (n=22, 77%), pain alone (n=1, 3%), mass, pain and fistula (n=2, 8%), mass, pain, fistula and itching (n=1, 3%), mass, pain and erythema (n=1, 3%), and ulcer and pain (n=1, 3%), while one case (3%) showed no local symptoms. Concomitant disease included tuberculous sinusitis (n=2, 8%), hypertension (n=1, 3%), and psoriasis and diabetes (n=1, 3%), whereas 25 cases (86%) showed no concomitant disease (table 1).

For the 33 patients with IGM, 19 patients were not given any treatment, nine were given antibiotics (but not specific for tuberculosis), three had their lesions excised, and two were given steroids. In all cases the lesions subsided and did not recur. The age range of the patients was 29–56 years (mean 37.85 years). The lesional sizes ranged from 1 to 5 cm (mean 2.53 cm). Systemic symptoms included two cases of cough (6%), while 31 (94%) cases demonstrated no symptoms. Local symptoms included abscess (n=3, 9%), mass (n=22, 67%), and pain alone (n=4, 12%), whereas four cases (12%) showed no local symptoms. No concomitant diseases were demonstrated in this group (table 1).

Significant differences were seen between the two groups for age (t=2.52, p<0.05) and lesional size (t=−5.56, p<0.01). TBM occurred in a significantly younger population, and demonstrated larger lesional sizes than IGM. There was no difference between the number of cases showing mass, local and systemic symptoms.

The histological characteristics of tuberculous and non-infectious granulomatous mastitis are presented in table 2. Comparing the different histological features, the TBM group showed significantly more fibrosis, eosinophils, and necrosis, whereas the IGM group showed significantly more plasma cells.

Table 2

Results of histological parameter analysis for idiopathic granulomatous mastitis and tuberculous mastitis

Further analysis was performed to evaluate the relationship between the histological parameters for all the cases taken together as one group. There was significant positive correlation between eosinophils and fibrosis (rs=0.39, p<0.01), and negative correlation between vague and well-formed granulomas (rs=−0.38, p<0.01) (table 3).

Table 3

Correlation between the various histological parameters analysed for non-tuberculous and tuberculous granulomatous mastitis cases


Granuloma is defined as a compact (organised) collection of mature mononuclear phagocytes (macrophages and/or epithelioid cells) that may or may not be accompanied by accessory features such as necrosis or the infiltration of other inflammatory leucocytes.1 It is a non-specific type of inflammatory response that may be triggered by diverse antigenic agents including infectious agents and inert foreign materials causing activation of the cellular immune system (T lymphocytes and macrophages).

In IGM, granulomas occur around the lobules and ducts in the breast in the absence of specific infectious agents, trauma or foreign material. The variability in clinical presentation and duration of symptoms reflects the heterogeneity of IGM. Presence of a palpable mass is the most common clinical finding, with or without associated pain and/or abscess. In the current series, 67% of patients showed mass lesions, and 12% presented with pain. Histologically, IGM is characterised by non-caseating granuloma with epithelioid histiocytes, presence or absence of multinucleated giant cells, predominantly plasma cell infiltrates with attendant lymphocytes and neutrophils, absence of necrosis, and negative microbiological examination (figures 3 and 4). Microabscess formation is occasionally seen. Necrosis and fibrosis are less salient features. Not all cases have typical well-formed granuloma, but all cases have epithelioid histiocytes, which can be seen in cytological preparations.9 Mammographic findings are non-specific, and sonography discloses a hypoechoic mass.13 14 An autoimmune phenomenon has been postulated as the underlying aetiology, as evidenced by the use of steroids in this disease along with consistent findings of negative cultures.8 Other conditions that have been associated with increased risk of IGM include pregnancy, breastfeeding, hyperprolactinaemia with galactorrhoea, and α-1-antitrypsin deficiency.14–17 Infection with Gram-positive bacilli, particularly Corynebacterium kroppenstedtii, has also been reported in a series of cases, but this remains to be elucidated.18 Most of the reported series of IGM have originated in developing countries, and it is postulated that there may be an underdiagnosis of TBM.16 17 19 20 The treatment of IGM is as moot as its aetiology. The available treatment options include close follow-up, immunosuppressive drugs and surgical excision. Wilson et al21 proposed a treatment algorithm indicating that patients with localised disease and symptoms should be observed with close clinical and radiographical follow-up; patients being followed with change in clinical findings should prompt additional biopsies and may be treated with localised excision; patients with localised disease may be treated with local excision; and for patients with more extensive disease and more severe symptoms of pain and sinus formation, trial of steroids is warranted once all infectious causes have been excluded.

Figure 3

Photomicrograph showing an idiopathic granulomatous mastitis granuloma occurring within and around the lobules (H&E, ×20).

Figure 4

Photomicrograph showing an idiopathic granulomatous mastitis granuloma with predominantly plasma cell infiltrates (H&E, ×40).

TBM, on the other hand, is characterised by the presence of aggregates of epithelioid histiocytes, more pronounced eosinophilic infiltrates and surrounding fibrosis, presence or absence of necrosis and Langhans-type giant cells, and scattered lymphocytes, plasma cells and neutrophils (figures 5 and 6). The granulomatous lesions are more associated with ducts than with lobules. Microabscess formation is seldom seen. Microbiological examination, including special stains (Ziehl-Neelsen), is variable, and can be negative. The most common presentation of TBM is a solitary breast lump followed by breast abscess and painful sensation.22 In the current series, 91% presented with mass and 20% with pain, either with or without other associated symptoms. TBM is often mistaken for carcinoma, especially if well-defined clinical features are absent. Many reported cases have had unnecessary mastectomies because of its clinical and radiological resemblance to carcinoma.23–25 Historically, TBM has been classified into five pathological varieties: nodular, diffuse, sclerosing, tuberculous mastitis obliterans and acute miliary tuberculous mastitis, based on clinical presentation.26 Tuberculous involvement of the breast occurs either by direct inoculation of the bacilli through abrasions in the nipple (this is rare), or more commonly via lymphatic, haematogenous or contiguous seeding. The lymphatic route is the most likely route of breast involvement, and occurs by retrograde extension from the axillary lymph node.27–32 Treatment of TBM is by specific antimycobacterial therapy.

Figure 5

Photomicrograph showing a tuberculous mastitis granuloma with surrounding fibrosis (H&E, ×20).

Figure 6

Photomicrograph showing a tuberculous mastitis granuloma with Langhans-type giant cells and more pronounced eosinophilic infiltrates (H&E, ×40).

Although IGM and TBM possess different aetiology, the clinical, radiological and histological features are highly similar. As the treatments for IGM and TBM are different, giving inappropriate treatment (eg, steroid therapy given for TBM misdiagnosed as IGM) would cause significant morbidity. This highlights the significance of correct diagnosis. Routine microbiological diagnosis does not provide high sensitivity, and in the areas endemic for tuberculosis, molecular testing for mycobacterial DNA by PCR may not be readily available. As such, attempts to differentiate these entities based on histology may be worthwhile, to guide further investigations, if not treatment.

In this study, the clinical features of TBM and IGM were similar. Histologically, TBM showed more fibrosis, eosinophils and necrosis, whereas IGM showed a higher plasma cell concentration. These observations may possibly be explained by the differences in aetiology. In TBM, the mycobacteria may incite chronic and granulomatous inflammatory responses more akin to those of a fungus rather than pyogenic bacteria, and one may postulate that eosinophils would play a more significant role in the cellular immune response. Conversely for IGM, the current belief is that the aetiology is more likely to be autoimmune, which would call into play humoral responses, and hence more plasma cells. Caseation necrosis as a result of tissue damage is a hallmark of tuberculosis, and can also be seen in TBM; in the current series, it was present in 41% of the TBM cases. However, necrosis is not specific to TBM, as it was also present in IGM, despite in a much lower number of cases (7%). In TBM, the associated tissue damage, necrosis and inflammatory response may result in more pathological accumulation of extracellular matrix in the chronic and organisational phase of the disease process, and hence more fibrosis.

Other histological factors do not appear to differ significantly between TBM and IGM. Of interest were the characteristics of the granulomas and the giant cells. In TBM and IGM, the granulomas could be vague or well formed, and the giant cells could be either Langhans or non-Langhans type. As Langhans giant cells are traditionally considered specific for tuberculosis, the presence of these in IGM did raise the possibility of whether this could be related to mycobacterial infection, at least in some of the cases.

Other rare causes of inflammatory breast lesions of the breast include: non-infectious causes such as sarcoidosis, Wegener's granulomatosis (WG) and foreign body reaction; infectious causes such as parasitic infestation or mycotic infection; and vasculitides such as giant cell arteritis and polyarteritis nodosum.16 Breast sarcoidosis could simulate breast carcinoma on ultrasound or mammography, and can present as a round, well-defined mass reflecting intramammary node involvement.33 34 In most cases of breast sarcoidosis, patients are highly likely to have subclinical pulmonary disease or other organ involvement. Pulmonary function testing, serum lysozyme and serum ACE levels are helpful in assessing activity of the disease, and directing therapy.35 36 In the current series, none of the patients had any history or systemic evidence of sarcoidosis.

WG usually presents with systemic manifestation and rarely with breast involvement. It is characterised by small-sized and medium-sized vasculitis in the lungs and kidney.37 Most patients will have a positive anti-neutrophil cytoplasmic autoantibody (ANCA). Serological studies include p-ANCA, c-ANCA, and the recently introduced PR-3 ANCA ELISA.38 39 In an extensive series of breast WG reported over the last 3 decades, most cases have systemic manifestations as well as serological/histological findings in which the histology revealed vasculitis of small-sized to medium-sized vessels in conjunction with granuloma.40 Other vasculitides that have been reported in the in the breast are polyarteritis nodosa and giant cell arteritis, with small-sized to medium-sized blood vessels being affected from a lymphocytic infiltrates within the external or internal elastic lamina or adventitia to a more pronounced destruction of the vessel wall with thrombosis and multinucleated giant cells.41 The absence of vasculitis in the current series precluded these vasculitides.

Tissue reaction to a foreign body, including a silicone implant, may also produce a granulomatous reaction.42 Other infectious organisms that incite a reaction producing a mass-like effect are parasites or fungal organisms. Demonstration of the fungus or parasite can be made by using H&E however, in most instances the organisms are unobtrusive. In the current series of cases, special stains such as PAS, PASD and GMS were all negative.

In the TBM group, the patients either had pulmonary symptoms, previous pulmonary tuberculosis, or strong exposure to patients infected with tuberculosis. In the literature, TBM may be secondary to bloodstream dissemination or be an extension of an adjacent tuberculous process.43 The clinicopathological pattern of tuberculous infection is multifaceted. Primary infection is the form of disease that develops in a previously unexposed unsensitised person, with only about 5% of patients developing a clinically significant disease. Secondary infection is a pattern that arises in a previously sensitised host, and this commonly occurs after reactivation of a dormant primary lesion. Reactivation of tuberculosis is more common in low-prevalence areas, while reinfection plays an important role in regions of high contagion.44 In the current series, most of the cases from the TBM group were from an endemic area with a high prevalence rate.45 It was likely that some of the patients with pulmonary symptoms represented reactivation of infection in the breast.

Given the somewhat similar histological picture of IGM and TBM, and in view of the significant difference in aetiology, it is imperative that, when one is dealing with a granulomatous mastitis, the presence of significant eosinophilic infiltrate, necrosis or fibrosis should prompt a diligent search for evidence of mycobacterial infection. Failure to respond to conventional or conservative therapy despite negative microbiological investigations should also make one seriously consider tuberculosis treatment.

The role of PCR for TBM has not been extensively investigated.43 In a recent study,44 it was shown that PCR detected mycobacteria in 50% of the mammary granulomatous reactions associated with breast cancer. These cases were all acid-fast stain negative, and there was no correlation with culture or treatment results. Cases with and without necrosis showed a similar PCR-positive rate of 50%. Whether any of these cases represented false-positive results remains an unresolved issue.

In summary, in this large cohort of clinical TBM and IGM, it was demonstrated that TBM was more likely to occur in younger patients, with a larger clinical mass at presentation. Histologically, TBM showed more eosinophils and necrosis, whereas IGM was associated with more plasma cells. The characteristics of the granulomas and giant cells were not distinguishable features. Given the relatively low sensitivity for TBM in microbiological tests, this diagnosis should be considered in the appropriate settings with the clinical and histological characteristics, even with negative microbiological results.

Take-home messages

  • Granulomatous mastitis can have infectious or non-infectious causes, respectively tuberculous mastitis (TBM) or idiopathic granulomatous mastitis (IGM), with both requiring different treatments.

  • Microbiological tests on histological or cytological samples have low sensitivity.

  • TBM tends to occur in younger patients with larger clinical mass, and histologically shows more eosinophils and necrosis.

  • IGM tends to show more plasma cells histologically.

  • The characteristics of the granulomas or giant cells are similar in TBM and IGM.

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 The authors report no competing interests. The authors alone are responsible for the content and writing of the paper.

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

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.