Idiopathic granulomatous mastitis: magnetic resonance imaging findings with diffusion MRI

Abstract

Background

Idiopathic granulomatous mastitis (IGM) is a rare benign breast disease with unknown etiology which can mimic breast carcinoma, both clinically and radiologically. Magnetic resonance imaging (MRI) findings of IGM have been previously described; however there is no study evaluating diffusion-weighted MRI findings of IGM.

Purpose

To analyze conventional, dynamic contrast-enhanced, and diffusion-weighted MRI signal characteristics of IGM by comparing it with the contralateral normal breast parenchyma.

Material and Methods

A total of 39 patients were included in the study. On dynamic contrast-enhanced MRI, the distribution and enhancement patterns of the lesions were evaluated. We also detected the frequencies of involving quadrants, retroareolar involvement, accompanying abscess, and skin edema. T2-weighted (T2W) and STIR signal intensities and both mean and minimum apparent diffusion coefficient (ADC) values were compared with the contralateral normal parenchyma.

Results

IGM showed significantly lower mean and minimum ADC values when compared with the normal parenchyma. Signal intensities on T2W and STIR sequences of the lesion were significantly higher than the normal parenchyma. On dynamic contrast-enhanced MRI, 7.7% of the patients had mass-like contrast enhancement, 92.3% of the patients had non-mass-like contrast enhancement. Abscess was positive in 33.3% of the patients.

Conclusion

As a result, IGM showed commonly non-mass-like lesions with restricted diffusion. Although it is a benign pathology, it may show clustered ring-like enhancement like malignant lesions.

Keywords

Granulomatous mastitis magnetic resonance imaging (MRI) diffusion MRI apparent diffusion coefficient (ADC)breast

Introduction

Idiopathic granulomatous mastitis (IGM) is a rare benign breast disease with unknown etiology mimicking breast carcinoma both clinically and radiologically (1). Ultrasound and mammography findings of IGM have been previously reported (2,3). However, there are limited numbers of studies evaluating magnetic resonance imaging (MRI). Diffusion-weighted magnetic resonance imaging (DW-MRI) findings of IGM have not yet been clearly described (2 –4).

In this study, we aimed to analyze MRI signal characteristics of IGM and compare the T2-weighted (T2W), STIR signal intensities (SI), and apparent diffusion coefficient (ADC) values of IGM with the normal appearing parenchyma of the contralateral breast. We also detected the frequency of the involved quadrants, retroareolar involvement, accompanying abscess, and skin edema.

Material and Methods

Patients

This retrospective study was approved by Baskent University Institutional Review Board (project number: KA: 14/303). Between the dates of January 2011 to February 2015, 43 patients who had pathologically proven granulomatous mastitis underwent breast MRI in our department. Four of the patients who had accompanying lobular intraepithelial neoplasm, contralateral breast carcinoma, granulomatous mastitis due to vasculitis, and bilateral diffuse granulomatous mastitis without normal appearing parenchyma were excluded. Diagnosis was confirmed by either ultrasound-guided core needle biopsy (n = 23) or surgical specimen (n = 16). Polymerase chain reaction DNA analyses for tuberculosis were negative for all patients. As a result we included 39 female patients (mean age, 40.33 years; age range, 24–58 years) who had pathologically proven idiopathic granulomatous mastitis. Control measurements were done from the contralateral normal appearing breast parenchyma of the same patients.

MRI examination

MRI studies were performed on 1.5 Tesla System (Avanto, Siemens, Erlangen Germany). Standard protocol for breast imaging including: (i) STIR (TR/TE/TI, 4120/99/160 ms); matrix, 320 × 85; NEX, 2; slice thickness, 3 mm; and (ii) T2W (TR/TE: 3960/105 ms); matrix, 512 × 85; NEX, 1; slice thickness, 3 mm, in axial planes was performed with breast coil. Field of view (FOV) was 330 × 90 mm for axial planes.

Echo-planar images were obtained with diffusion gradients in the x, y, z planes at b values of 0, 200, 500, and 800 (TR/TE, 6500/108 ms; matrix, 192 × 100; NEX, 2; slice thickness, 4 mm). ADC maps were automatically created.

The scanning parameters for dynamic contrast-enhanced MRI were as follows: TR/TE, 5.17/2.39 ms; matrix, 384 × 80; slice thickness, 1.1 mm; flip angle, 10°; FOV, 47.7 × 33.7 cm.

MRI analysis

Two radiologists evaluated all the images on consensus at a workstation (Synapse version 4.0, Fujifilm Medical Systems Inc., Stamford, CT, USA) for qualitative and quantitative evaluation. Both radiologists determined the involving quadrant or quadrants, presence of retro-mammary space extension, and skin edema with consensus. T2 and STIR SI were measured and compared with the contralateral normal appearing breast parenchyma.

The lesion and the normal appearing parenchyma were spatially labeled, which allowed getting measurements from the same area on all sequences. A region of interest (ROI) of 300–450 mm² was placed on the lesion site and the contralateral normal appearing parenchyma. When measuring ADCs, a large ROI (range, 300–450 mm²) was used to get mean ADC values and a small ROI (range, 20–40 mm²) was set to get minimum ADC on the labeled area of both breasts (Fig. 1). If the skin thickness increased focal or diffuse on STIR images, a diagnosis of skin edema was accepted. Retroareolar involvement was defined as the extension of lesion closer than 2 cm to the nipple.

Fig. 1.

A 27-year-old woman with IGM. (a) Subtracted contrast-enhanced MRI depicts non-mass-like regional heterogenous enhancing lesion in the right breast. The lesion and contralateral normal parenchyma were spatially labeled on all sequences. (b) Diffusion-weighted MRI at b value of 800 showed increased SI of the lesion. (c) ADC map depict lower mean and minimum SI of the lesion compared with the contralateral normal parenchyma.

On dynamic contrast-enhanced MRI, the distribution and enhancement patterns of the lesions were evaluated according to the American College of Radiology BI-RADS Atlas 5th edition (5).

Statistical analysis

The SPSS 16.0 software package was used for statistical analysis (SPSS Inc., Chicago, IL, USA). By using paired t-test, all SI and ADC values were compared. P < 0.05 was considered statistically significant. The involving frequencies of the each quadrant were also analyzed.

Results

Twenty-four of the 39 patients had IGM on the left breast (61.5%). The frequency of the involving quadrants were: upper outer, 59%; upper inner, 56.4%; lower outer, 56.4%; lower inner, 51.3%. Twenty-nine of the patients (74.4%) had accompanying skin edema and 32 patients (82.1%) had retroareolar space involvement.

The mean SI on T2W and STIR sequences are tabulated in Table 1 and calculated ADC values are shown in Table 2. The STIR, T2 SI, and ADC values of each patient are summarized in Table 3.

Table 1.

The T2 and STIR signal intensities.

	Lesion mean ± SD	Control mean ± SD	P
T2 SI	916.897 ± 226.45	832.1 ± 257.7	0.026
STIR SI	444.462 ± 146.64	208.308 ± 68.27	<0.001

SD, standard deviation.

Table 2.

The ADC values of the lesion and normal parenchyma.

	Lesion (mm²/s)	Control (mm²/s)	P
Mean ADC	1.234 × 10⁻³± 0.19	1.409 × 10⁻³± 0.3	0.002
Minimum ADC	0.903 × 10⁻³± 0.19	1.233 × 10⁻³± 0.29	<0.001

Table 3.

The STIR, T2 SI, and ADC values of each patient.

Case	STIR SI	T2 SI	Minimum ADC ( × 10⁻³ mm²/s)	Mean ADC ( × 10⁻³ mm²/s)
1	144	480	1.007	1.161
2	474	960	1.115	1.474
3	459	992	0.692	1.249
4	354	544	0.876	0.964
5	303	676	0.997	1.103
6	312	1372	0.808	1.108
7	422	880	0.904	1.127
8	535	1153	1.199	1.495
9	367	876	0.97	1.389
10	382	754	1.003	1.168
11	556	1025	0.849	1.241
12	686	1077	0.753	1.292
13	508	1105	1.448	1.693
14	294	737	0.817	1.294
15	384	654	1.139	1.494
16	556	940	1.115	1.452
17	267	877	0.956	1.136
18	635	947	0.584	0.874
19	536	830	0.945	1.332
20	431	691	0.838	1.372
21	462	794	0.997	1.231
22	426	785	1.058	1.171
23	653	1415	0.649	0.899
24	396	774	0.962	1.358
25	693	1201	0.659	1.048
26	417	711	0.835	1.181
27	453	803	1.167	1.522
28	317	1076	1.026	1.281
29	354	841	1.057	1.426
30	369	954	1.039	1.409
31	371	776	0.932	1.227
32	320	727	0.786	1.172
33	438	710	0.65	0.987
34	924	1295	0.551	1.302
35	553	1060	0.432	0.757
36	264	681	0.779	1.092
37	596	1187	0.924	1.344
38	348	1150	0.929	1.148
39	375	1249	0.777	1.161

SI on T2W and STIR sequences of the lesion were significantly higher than normal parenchyma. Both mean and minimum ADC values from the lesion were lower than normal parenchyma (P value = 0.002 and < 0.001, respectively).

The BI-RADS MRI scores of the lesions were: BI-RADS 2 (n = 1), BI-RADS 3 (n = 4), BIRADS 4 (n = 30), and BI-RADS 5 (n = 4).

On dynamic contrast-enhanced MRI, three (7.7 %) of the patients had mass-like contrast enhancement, 36 (92.3%) of the patients had non-mass-like enhancement (Figs 2 and 3). Shape and enhancement patterns of the mass-like lesions were: irregular homogenous, irregular rim-like, oval shaped lesion including black internal septations. Non-mass-like enhancing lesions showed diffuse (n = 9) and regional (n = 27) distribution. Enhancement patterns of these lesions were: clustering rings (n = 14), heterogenous enhancement (n = 22). An abscess was found in 13 of the 39 patients (33.3%) (Fig. 2).

Fig. 2.

A 34-year-old woman with IGM. (a) T2W MRI showed hyperintense lesion in the left breast with surrounding hypointense halo (black arrow). (b) Diffusion-weighted MRI at b value of 800 showed, hyperintense SI of the lesion. (c) The lesion was hypointense on ADC map, consistent with restricted diffusion (arrowhead). (d) Subtracted contrast-enhanced MRI depicts non-mass-like enhancing lesion with clustered rings. The lesion showed peripheral enhancement consistent with abscess.

Fig. 3.

A 35-year-old woman with IGM. (a) Subtracted contrast-enhanced MRI depicts non-mass-like enhancing lesion with clustered ring pattern on left breast. (b) Diffusion-weighted MRI at b value of 800 showed, hyperintense SI of the lesions. (c) ADC map showed hypointense SI of the lesions consistent with restricted diffusion.

Discussion

In this study we aimed to analyze conventional, dynamic contrast-enhanced, and diffusion-weighted MRI signal characteristics of IGM by comparing it with the contralateral normal breast parenchyma. IGM showed lower mean and minimum ADC values and higher T2 and STIR SI when compared with the normal parenchyma.

IGM is a rare chronic inflammatory benign breast disease with unknown etiology, which can mimic breast carcinoma (1). Mammography and ultrasound findings of IGM have been previously reported, however, diffusion-weighted MRI findings of IGM have rarely been described (2,3). By far, this is the largest series evaluating diffusion-weighted MRI findings of IGM.

In the literature non-mass-like enhancing lesions in IGM were more frequent than mass-like lesions, concordant to our study (2,6). Non-mass-like enhancement may be seen in both benign and malignant lesions including hormonal stimulation, inflammatory causes, ductal carcinoma in situ, and invasive lobular carcinoma. Imamura et al. showed that adding ADC value criteria to dynamic contrast-enhanced MRI increased sensitivity of differentiating malignant non-mass-like enhancing breast lesions from benign non-mass-like enhancing lesions while decreasing specificity (7). Low specificity is the main problem of breast MRI (8 –10). Different cutoff ADC values ranging from 1.1 × 10⁻³ to 1.3 × 10⁻³ mm²/s were reported to differentiate benign and malignant non-mass-like lesions (7,11). Partridge et al. suggested an ADC threshold of 1.81 × 10⁻³ mm²/s (12). As the cutoff value of ADC increases, the biopsy frequency of benign lesions may increase. Although IGM is a benign condition, in our study it had lower mean and minimum ADC values than previously reported ADC threshold values. Although we did not compare the ADC values of IGM with malignant breast lesions, diffusion MRI may not help to differentiate IGM from breast carcinoma.

Tozaki et al. studied the incidence of clustered ring like enhancement pattern among non-mass-like enhancing lesions (13). Sixty-three percent of malignant and 4% of benign lesions showed clustered ring-like enhancement in that study. Clustered ring-like enhancement was found in breast carcinomas, duct ectasia, and fibrocystic disease (13). Although IGM is a benign condition, in our study 14 of 36 (28.8%) non-mass-like enhancing lesions showed clustered ring-like enhancement pattern. When evaluating this pattern, presenting symptoms, physical examination, and patient’s history should be considered.

Retroareolar space involvement was present in 82.1% of the patients in our study. In the literature, inflammation of periductal areas in IGM has been previously shown pathologically (14,15). High SI on T2W and STIR sequences are perhaps the result of increased water and edema in the inflammation area. The clustered ring-like enhancement may be due to the enhancing inflamed periductal parenchyma and restricted diffusion may be due to intraductal pus.

Breast carcinomas typically show low SI on T2W MRI. In our study IGM showed high T2 SI in contrast to breast carcinomas. Mucinous carcinomas, malignant phyllodes tumors, and invasive ductal carcinoma may show high T2 SI, soT2 SI characteristics alone may not be a good indicator to differentiate benign and malignant lesions (16).

Our study has several limitations. Inter- and intra-observer variability were not assessed. The retrospective design of the study was another limitation.

In conclusion, IGM showed lower mean and minimum ADC values and higher T2 and STIR SI when compared with normal breast parenchyma. Although it is a benign condition, IGM show restricted diffusion and clustered ring-like enhancement pattern like malignant lesions. Non-mass-like enhancement pattern was more frequent than mass-like enhancement.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support with respect to the research, authorship, and/or publication of this article.

References

Milward

Gough

. Granulomatous lesions in the breast presenting as carcinoma. Surg Gynecol Obstet 1970; 130: 478–482.

Dursun

Yilmaz

Yahyayev

. Multimodality imaging features of idiopathic granulomatous mastitis: outcome of 12 years of experience. Radiol Med 2012; 117: 529−538–529−538.

Van Ongeval

Schraepen

Van Steen

. Idiopathic granulomatous mastitis. Eur Radiol 1997; 7: 1010−1012–1010−1012.

Al-Khawari

Al-Manfouhi

Madda

. Radiologic features of granulomatous mastitis. Breast J 2011; 17: 645–650.

American College of Radiology. Breast imaging reporting and data system (BI-RADS), 5th ed. Reston, VA: American College of Radiology, 2013.

Ozturk

Mavili

Kahriman

. Granulomatous mastitis: radiological findings. Acta Radiol 2007; 48: 150–155.

Imamura

Isomoto

Sueyoshi

. Diagnostic performance of ADC for non-mass-like breast lesions on MR imaging. Magn Reson Med Sci 2010; 9: 217–225.

Orel

Schnall

LiVolsi

. Suspicious breast lesions: MR imaging with radiologic-pathologic correlation. Radiology 1994; 190: 485–493.

Fischer

Kopka

Grabbe

. Breast carcinoma: effect of preoperative contrast-enhanced MR imaging on the therapeutic approach. Radiology 1999; 213: 881–888.

10.

Guo

Cai

. Differentiation of clinically benign and malignant breast lesions using diffusion-weighted imaging. J Magn Reson Imaging 2002; 16: 172–178.

11.

Yabuuchi

Matsuo

Kamitani

. Non-mass-like enhancement on contrast-enhanced breast MR imaging: lesion characterization using combination of dynamic contrast-enhanced and diffusion-weighted MR images. Eur J Radiol 2010; 75: 126–132.

12.

Partridge

Mullins

Kurland

. Apparent diffusion coefficient values for discriminating benign and malignant breast MRI lesions: effects of lesion type and size. Am J Roentgenol 2010; 194: 1664–1673.

13.

Tozaki

Igarashi

Fukuda

. Breast MRI using the VIBE sequence: clustered ring enhancement in the differential diagnosis of lesions showing non-masslike enhancement. Am J Roentgenol 2006; 187: 313–321.

14.

Erhan

Veral

Kara

. A clinicopathologic study of a rare clinical entity mimicking breast carcinoma: idiopathic granulomatous mastitis. Breast 2000; 9: 52–56.

15.

Heer

Shrimankar

Griffith

. Granulomatous mastitis can mimic breast cancer on clinical, radiological or cytological examination: a cautionary tale. Breast 2003; 12: 283–286.

16.

Santamaría

Velasco

Bargalló

. Radiologic and pathologic findings in breast tumors with high signal intensity on T2-weighted MR images. Radiographics 2010; 30: 533–548.