A retrospective analysis was conducted of the medical records of patients with surgically confirmed ILC treated at the Gachon University of Medicine and Science between 2009 and 2014. We identified 71 ILC patients based on pathological data. Two patients who underwent excisional biopsy before breast MRI and six patients without preoperative breast MRI were excluded; hence, data from 63 patients were available. One patient had bilateral cancer; therefore, there were a total of 64 cancers from 64 affected breasts.

Further clinical information was also reviewed, including the patient age, symptoms or signs (tumor palpability and nipple discharge), and family history of breast cancer.

The MRI study was performed with patients in the prone position with 3.0 Tesla machines (Skyra and Verio, Siemens Medical Solutions, Erlangen, Germany) equipped with a breast coil. The imaging protocol consisted of a fat-suppressed T1-weighted, T2-weighted, axial or sagittal, 3-dimensional, gradient echo sequence, and dynamic contrast-enhanced (DCE) images. Imaging with the 3.0-T Skyra scanner covered both breasts with a minimum repetition time and echo time (4.79/1.83 ms for the dynamic images), a 10 flip angle, a 34 cm field of view, 1.5 mm section with no gap, a 512 × 512 matrix, and a scan time of 1-2 min. Imaging with the 3.0-T Verio scanner covered both breasts with a minimum repetition time and echo time (4.7/1.78 ms for the dynamic images), a 10 flip angle, a 35 cm field of view, 1.5 mm section with no gap, a 512 × 512 matrix, and a scan time of 1-2 min. For dynamic contrast enhancement, 0.1 mmol/kg body weight gadobutrol (Gadovist; Schering AG, Berlin, Germany) was injected, followed by a 20-mL saline flush. After the examination, two subtraction images were made automatically on a pixel-by-pixel basis; the unenhanced images were subtracted from the early post-contrast images (standard subtraction), and the last post-contrast image was subtracted from the early post-contrast images (reverse subtraction). The reformatted images with a maximum intensity projection were then created from the standard and reverse subtraction images. The breast MR images were interpreted by one of two faculty radiologists present at that time.

The images were interpreted independently by two breast radiologists (Y.E.Y., N.S.Y., with 6-7 years of experience in breast imaging) who were blinded to the clinical and radiologic information of all patients. The image manipulations did not involve computer-aided detection. Interpretations were performed by consensus review to resolve differences after their review of images.

The morphology, DCE kinetics, and multiplicity were analyzed based on the American College of Radiology Breast Imaging Reporting and Data System MRI lexicon, 2nd edition (12). The lesions were classified as mass or non-mass enhancement. The masses were evaluated for shape (oval, round, irregular), margin (circumscribed, irregular, spiculated), and internal enhancement characteristics (homogeneous, heterogeneous, rim enhancement). Furthermore, there were also analyses of the lesion distribution (focal, linear, segmental, regional, multiple regions, diffuse) and internal enhancement pattern (homogeneous, heterogeneous, clumped, clustered ring) of non-mass enhancement lesions. The DCE kinetic time-intensity curve was evaluated based on initial and delayed phases. The initial phase was described as slow, medium, and fast. The delayed phase was categorized into three types: type 1-persistent, type 2-plateau, or type 3-washout. Multiplicity was defined as the presence of more than one unconnected lesion in a breast. When there were multiple lesions in one breast, only the largest lesion was analyzed.

We reviewed the pathological reports for the following parameters: tumor size, multiplicity, nodal involvement, number of metastatic lymph nodes, and ER, PR, and HER2 statuses. The positive ER and PR statuses were defined by an Allred score greater than or equal to 3, based on immunohistochemical staining. The HER2 status was considered positive if the immunohistochemical stain results indicated an Allred score of 3+ or 2+, with confirmation of HER2 gene amplification by silver in situ hybridization.

Statistical analysis of the patients' ages and tumor sizes was performed using the Student t-test and Mann-Whitney U test, respectively. Multiplicity, lymph node metastasis, and the number of lymph node metastases were compared using the chi-square test. Differences in MRI features according to ER, PR, and HER2 statuses were investigated using the chi-square test or the Fisher exact test. All statistical analyses were performed using SPSS software version 19.0 (SPSS Inc., Chicago, IL, USA); and a P value < 0.05 was considered statistically significant.

The mean (± standard deviation) age of the patients was 49.78 (± 1.02) years. Of the 64 ILCs, 59 (92.2%), 54 (84.4%), and 49 (76.6%) were ER positive, PR positive, and HER2 negative, respectively. Significant differences in clinicohistological features correlated only with PR status. A comparison of the clinicohistological characteristics between the PR positive and PR negative cancers is shown in Table 1. PR negativity was associated with larger tumor sizes (P = 0.046). A total of 20 PR positive patients (37.0%) and 6 PR negative patients (60%) had lymph node metastases, which was a statistically insignificant difference (P = 0.157); however, the number of positive metastatic lymph nodes was significantly different between PR positive and PR negative patients (P = 0.010). There was no correlation between ER or HER2 statuses and clinicohistological features in ILCs (P > 0.05).

The mean tumor size on MRI was 2.96 ± 0.26 cm. Tumor measurements determined by MRI were not significantly different from those recorded in the pathological reports (3.48 ± 0.33 cm). A significant difference in MRI features was only noted according to the PR status of the tumor. Table 2 summarizes the comparison of analyzed MRI features from 54 PR positive cancers (the data are from 53 patients; one had bilateral cancer) and 10 PR negative cancers. A total of 49 of 54 PR positive cancers (90.7%) showed mass lesions (Fig. 1) and five (9.3%) showed non-mass enhancement. A total of six of the 10 PR negative cancers (60%) showed mass lesions and four (40%) showed non-mass enhancement (Fig. 2). Therefore, PR negative cancers were more likely to show non-mass enhancement (4/10 vs. 5/54, P = 0.027). Of the non-mass enhancement cancers, PR negative cancer was more likely to show clustered ring internal enhancement pattern on MRI than PR positive cancer (P = 0.056), but the difference was not statistically significant. There was no significant difference in MRI features according to ER or HER2 statuses (Table 3).