Lactation Effect on Findings of Contrast-Enhanced and Diffusion-Weighted Magnetic Resonance Imaging in Patients with Idiopathic Granulomatous Mastitis

Introduction: Idiopathic granulomatous mastitis is a benign chronic inflammatory disease of the breast. Lactation is considered as one of the most important risk factors. We investigated the effect of lactation on the findings of contrast-enhanced and diffusionweighted magnetic resonance imaging (DW-MRI) in IGM and aimed to identify the most observed findings. Methods: Contrast-enhancedand DW-MRI of 40 patients with lactation history in the last 5 years and 35 patients reporting no lactation history had been reevaluated retrospectively. Morphological features, enhancement pattern and kinetics of lesions were assessed based on BI-RADS. The presence of diffusion restriction was evaluated and apparent diffusion coefficient (ADC) values were obtained. MRI findings with lactation status were compared. Results: Non-mass contrast enhancement (p<0.02), cluster pattern (p<0.008) and fistula formation (p<0.035) were more frequently seen in patients with a lactation history than in patients reporting no lactation history. On MRI most common observed findings were concomittantly mass and non-mass contrast enhancement and abscess formation. Diffusion restriction was present in all of the lesions and the mean ADC values were 0.93 ± 0.25 x 10-3 mm2/s. Conclusion: Lactation status increases the incidence of non-mass contrast enhancement, cluster pattern and fistula formation on MRI.


Introduction
Granulomatous diseases of the breast are a rare clinical condition in which the pathogenesis is still not fully understood; tuberculosis, sarcoidosis, foreign body reaction, mycotic and parasitic infections, autoimmune diseases, infection, trauma and oral contraceptive pill use, ethnicity, breast feeding, and prolactinemia are one of the describing risk factors [1,2]. Idiopathic granulomatous mastitis (IGM) categorized as granulomatous disease of breast is relatively rare, benign, chronic inflammatory breast lesion which clinically and radiographically could misdiagnose as a inflammatory breast carcinoma [3]. However cause of IGM is not clear, its frequency is increasing with lactation, pregnancy and hyperprolactinemia [4].
The sonography and mammography (MG) findings of IGM are widely described, and magnetic resonance imaging (MRI) findings have been reported in a limited number of patients. There are only few reports about diffusion-weighted magnetic resonance imaging (DW-MRI) findings, the apparent diffusion coefficient (ADC) values of the IGM and MRI findings [5].
To the best of our knowledge; the relationship between MRI findings and lactation which is an important risk factor has not been investigated. To evaluate the association between lactation status and MRI findings, we compared MRI findings of IGM patients with and without lactation history. We also aimed to describe contrast-enhanced MRI, DW-MRI findings of IGM, ADC values of IGM by revealing the radiological features of IGM in a large patient series. years or not. Group 1 consisted of 40 patients with lactation history; group 2 consisted of 35 patients without lactation history. Patients with spesific granulomatous diseases such as sarcoidosis, tuberculosis, cat stracth disease, breasts with implants, trauma and bacterial mastitis were excluded from the study. All of the patients had undergone tru-cut biopsy or segmental mastectomy and pathologically proven as IGM. The clinical and pathological findings obtained from local database of our hospital. Of the 75 patients with IGM; 26 of had undergone mammographic examination. MG examinations were performed using a Full Field Digital Mammography System (mammomat inspiration, Siemens, Erlangen, Germany). All of our patients had MRIs.

MR Protocol
All examinations were performed using dedicated 16-channel double breast coil with 1.5 Tesla MRI (Magnetom Aera; Siemens Healthcare, Erlangen, Germany) equipped with 45-mT/m gradients and the patient is placed in the prone position on the The all assessments were performed by a experienced radiologists (AA) in the field of breast imaging. The MRI and MG findings of the lesions were classified on the basis of the breast imaging reporting and data system (BI-RADS) lexicon established by the American college of radiology [6]. We defined type A and B breast pattern as dominantly lipomatous type 1, dominantly fibroglandulary type 2 was also included type C and D breast pattern. Mammographically, lesions were defined as asymmetric density, structural distortion, and mass lesion according to their appearance. Those without pathological findings were evaluated as normal. On MRI, enhancement patterns of lesions were categorized as mass and non mass enhancement (NME). Subsequently; morphological features, internal enhancement and kinetic curves were evaluated, presence of diffusion restriction was investigated and ADC values were obtained. Morphological features of masses were assessed according to their shapes (round, oval or irregular), margins (circumscribed or not). Internal enhancement features were classified as homogeneous, heterogeneous, or rim enhancement. The NME features were also described depending on their distribution as focal, linear, segmental, regional, multipl regional or diffuse and internal enhancement patterns defined as homogeneous, heterogeneous, clumped, or clustered ring. The timesignal intensity curve paterns determined as persistent (type 1), plateau (type 2) or washout (type 3). Diffusion restriction features of lesions were classified as single and multiple and ADC values were obtained via region of interest (ROI). After describing lesions features on MRI, the MRI and DWI findings of Groups 1 and 2 were compared and the effect of lactation on MRI findings were investigated. We also determined frequently encountered features of IGM by assessing MRI findings of all patients.

Statistical analysis
Frequency of MRI findings and presence of diffusion restriction and mean ADC values of all patients were demonstrated. The statistical analysis was performed with the SPSS 22 software (Copyright IBM Corporation and its licensors 1989, 2012). The results of the descriptive statistics were given as mean ± standard deviation (SD) and min-max. MRI findings based on lactation status were compared with Chi-square test. The association between ADC values and lactation status were investigated by using Kruskal Wallis and Mann Whitney-u test. p value less than 0.05 was considered as statistically significant.

Results
A total of 75 patients were reevaluated, retrospectively. Median age was 35 (ranging from 23 to 65). While 60% of our patients suffered from mastodinia, rest of them presented with palpable mass. 12% of our patients were breastfeeding, there was a lactation history within recent five years in 41% of them. The rest of them had no recent lactation history. The rate of right and left breast involvement was almost equal. While one third of our patients had breast pattern as dominantly lipomatous type 1, fibroglandulary dominantly type 2 pattern was noted in rest of them. Abnormal mammographic finding were not noted in 17% of our patients, asymmetrical density, distortion and mass lesion were seen in of 47.8%; 30.4 and 4.3; respectively (Table 1). Microcalcifications were not detected in any of our patients. When the MRI findings were compared between group 1 and group 2 according to the lactation status, NME and mass association, clustered-ring type enhancement as a non-mass internal contrast enhancement pattern and fistula formation co-occurrency were statistically significant ( Table 2) (Figures 1-3). There was no significant difference between the morphological features and enhancement patterns. On DWI, there was no significant difference between groups 1 and 2 in terms of diffusion restriction and mean ADC values.
When general describing features of lesions on MRI are summarized; mass lesion detected in 84% of all patients, NME also noted in 97% of all patients. The most frequently observed features were masses with a round shape (50.6%), a well circumscribed (58.6%), and rim enhancement pattern (69.3%). Of the NMEs,   (Figure 4 a,b,c).

4/6
Discussion IGM is a rarely encountered benign disease generally affecting women with childbearing age [7]. Patients may complain of a hard lump causing nipple retraction and fistulas on the skin [8]. The vast majority of patients have at least one live birth, history of lactation within last five years [9]. The association between IGM and lactation is explained by various hypotheses. It is hypothesized that proteinrich secretion accumulates in the ducts, especially due to hormonal changes in pregnancy and lactation, leading to dilatation of the ducts and leakage into the periductal area. Leakage into the periductal area may have an important role in development of IGM which is characterized with multiple abcess formation and accompanying suppurative necrosis by leading chemical mastitis [10,11]. Taylor et al. [2] assumed that leakage of the protein-rich secretion initiated a granulomatous response. In our research evaluated the effect of lactation on MRI findings; we found that the co-occurrence of clustered-ring enhancement pattern and fistula formation was more frequent. Clustered ring type enhancement is the pattern of contrast enhancement that defines the enhancement around the duct. The increase in enhancement of the periductal stroma may be seen secondary to infectious processes and malignancies [6]. Regarding the fact that this type of clustered ring pattern enhancement pattern is more frequent in our patients with lactation history has supported the hypothesis of chemical inflammation. Leakage of milk into the periductal area and leading chemical inflammation may be one of the most important factors in the development of IGM. Moreover; cooccurrence of fistula formation was more frequent in our lactating patients, this may be related to the fact that the inflammatory reaction is being provoked by lactation. In other words lactation might be the factor triggering the event, then contributing to the acceleration. Being aware of the increased possibility fistula development in IGM patients with lactation history could be guiding in follow ups and treatment choice; which does not yet have a established treatment algorithm.
Imaging features of IGM is atypical and characteristic features of IGM has not been radiologically described. A core needle biopsy is always necessary to obtain definitive diagnosis [12]. Mammographic appearance can be vary from asymmetrical density increasement to small, multiple, nodularities [13]. IGM also can be observed on MG as illdefined masses with spiculated contours; however microcalcifications generally do not accompany IGM different from the malignancy. Asymmetrical density increasement and parenchymal distortion without accompanying microcalcifications was observed in most of our patients as mammographic appearences corresponding to literature data [13,14].
Since IGM is rarely seen, there are very few studies in the literature that describe MRI findings. Our study has included one of the largest patient series, detailing the breast MRI findings of IGM. We investigated MRI findings of our patients as contrast-enhanced and diffusion-weighted imaging. Most important finding observed in most our patients was multiple numerous abcess formation. Most of them were well-circumcsribred round lesions with rim-shaped enhancement pattern, besides in 1 to 3 patients expressed irregular margins and shaped nonspesific lesions. Abcess formation was reported as very suggestive finding of IGM in literature data [14,15]. Shelfout et al. [16] found that nonspesific ill-defined masses lesions with multiple ring-like abscess formations is one of the most common appearance on dynamic MRI.
Regional asymmetrical enhancement , which was heterogeneously, was also one of the common findings in almost all of our patients on contrast-enhanced sequences. While Van ongeval et al. [17] described as one of most frequent MRI pattern of IGM as segmental heterogenous contrast enhancement, segmental enhancement pattern was rare in our series. In the research investigating enhancement patterns of IGM, reported as time-intensity curves may varied from patient to patient and from lesion to lesion [14]. Erozgen et al. [18] found that benign curves are more frequent. We mostly observed plato pattern, which is suitable for both benign and malign lesions. On diffusion weighted imaging (DAG); diffusion restriction can be frequently seen, since the inflammatory response in IGM leads to reduce in diffusion of water. The recent literature data has revealed that diffusion restriction is seen and ADC values is lower than normal breast parenchyme in patients with IGM [7,19]. A point to be emphasized; multiple, round-shaped hypointense areas were seen in more than half of our patients; which was characteristic appearence on ADC mapping. All patients had diffusion restriction and lower ADC values.

Conclusion
In conclusion, lactation status affects MRI findings, increases, clustered ring enhancement, fistula formation, co-ocurence of mass and NME. The most important contrast-enhanced and DWI MRI finding was parenchymal oedema, multiple abcesses and roundshaped hypointense areas on ADC mapping.