Original Research JOURNAL OF MAGNETIC RESONANCE IMAGING 32: (2010)

Transcription

1 JOURNAL OF MAGNETIC RESONANCE IMAGING 32: (2010) Original Research Peripheral Hyperintense Pattern on T2-Weighted Magnetic Resonance Imaging (MRI) in Breast Carcinoma: Correlation With Early Peripheral Enhancement on Dynamic MRI and Histopathologic Findings Hiroko Kawashima, MD, 1 * Miki Kobayashi-Yoshida, MD, 2 Osamu Matsui, MD, 2 Yoh Zen, MD, 3 Masayuki Suzuki, MD, 1 and Masafumi Inokuchi, MD 4 Purpose: To investigate the correlation between the peripheral hyperintense pattern of breast carcinoma on T2- weighted images (T2WI) and the early peripheral enhancement (EPE) on dynamic magnetic resonance imaging (MRI) and to examine the histological characteristics involved in the causes thereof. Materials and Methods: The signal intensity of 131 breast carcinomas of common type on T2WI was classified into four types and the correlations between the signal intensity and the histopathological subtypes were evaluated. We investigated correlation with the peripheral hyperintense pattern on a T2WI and other findings on MR images including EPE. In addition, we also investigated the histopathological characteristics in the lesions with a peripheral hyperintense pattern. Results: The signal intensities of the lesions on T2WI demonstrated hyperintensity, isointensity, hypointensity, and a peripheral hyperintense pattern in 26 (20%), 66 (50%), 19 (15%), and 20 (15%) lesions, respectively. A peripheral hyperintense pattern correlated with the appearance of EPE (P < ) and the degree of central fibrosis (P < ). Conclusion: A peripheral hyperintense pattern on T2WI is a characteristic finding observed in breast carcinomas and correlated with the appearance of EPE. The appearance of a peripheral hyperintense pattern was affected by 1 Department of Quantum Medical Technology, Graduate School of Medical Science, Kanazawa University, Kanazawa, Japan. 2 Department of Radiology, Graduate School of Medical Science, Kanazawa University, Kanazawa, Japan. 3 Division of Pathology, Kanazawa University Hospital, Kanazawa, Japan. 4 Department of Gastroenterologic and Breast Surgery, Graduate School of Medical Science, Kanazawa University, Kanazawa, Japan. *Address reprint requests to: H.K., Department of Quantum Medical Technology, Graduate School of Medical Science, Kanazawa University, Kodatsuno, Kanazawa, , Japan. hirokok@med.m.kanazawa-u.ac.jp Received March 2, 2010; Accepted June 3, DOI /jmri View this article online at wileyonlinelibrary.com. various histopathological features in the same manner as that observed for the formation of EPE. Key Words: peripheral hyperintense pattern; T2-weighted MR images; breast carcinoma; early peripheral enhancement; dynamic MR imaging J. Magn. Reson. Imaging 2010;32: VC 2010 Wiley-Liss, Inc. BREAST MAGNETIC RESONANCE IMAGING (MRI) is currently widely used with the main purpose of diagnosing the progress level before surgery for breast carcinoma. The examination mainly comprises a dynamic study using a contrast agent, and it is therefore rare to perform breast MRI for patients for whom contrast agents cannot be used. T2-weighted imaging is almost essential in MRI for all regions, and the objective of such imaging is to evaluate internal properties in order to obtain a qualitative diagnosis of a lesion. Although T2-weighted imaging is not as important as dynamic MRI during breast MRI, there are many institutions where T2-weighted imaging is conducted for the purpose of evaluating the internal properties of a lesion. Breast lesions that present strong high signals on a T2-weighted image include cysts, mucinous carcinomas, and mucinous fibroadenomas (1 3). It has been reported that the signal intensity of a breast carcinoma on T2-weighted images presents medium to low signals compared to the adjacent breast parenchyma in most cases, whereas fibroadenomas tend to present higher signals than breast carcinomas (4). In recent years, it has been reported that there are also cases involving special types of breast carcinoma other than mucinous carcinoma in which strong high signals are presented on T2-weighted images, and the involvement of abundant cytoplasm, edematous stroma, and hemorrhage have been pointed out as causal factors (5). Kobayashi et al (6) previously reported two types of ring-like enhancement observed in dynamic MR VC 2010 Wiley-Liss, Inc. 1117

2 1118 Kawashima et al. images in a study limited to common types of breast carcinoma and analyzed the cause. In the observations of T2-weighted images during routine reading of MRI we noted the presence of breast carcinoma that presented a peripheral T2 high-signal area (peripheral hyperintense pattern on T2-weighted images) that was similar in shape to Kobayashi et al, called early peripheral enhancement (EPE) on dynamic MRI (6). The purpose of the present study was to investigate the correlation between the peripheral hyperintense pattern of breast carcinoma on T2-weighted images and the EPE on dynamic MRI and to examine the histological characteristics involved in the causes thereof. MATERIALS AND METHODS Between January 1995 and December 2004, 518 patients with breast carcinoma underwent dynamic MRI at our institution. Of these 518 patients, 207 patients with the following cancers were excluded: special types of invasive carcinomas, 25; diffusely enhanced lesions, 34; advanced cancers extending into the entire breast, 16; entirely cystic masses, 7; cases that had received preoperative neoadjuvant chemotherapy, 23; those who had undergone an excisional biopsy, 67; recurrent breast carcinoma, 14; and other, 21. Of the remaining 311 patients, T2- weighted images were not obtained in 180 patients. Ultimately, 131 patients (age years; mean age 56 years) with breast carcinomas were selected for this study. Among these 131 cancers, 112 had no special type of invasive ductal carcinomas and 19 were ductal carcinoma in situ. Disease grades were 19 Stage 0, 50 Stage I, 41 Stage IIA, 17 Stage IIB, 1 Stage IIIA, 2 Stage IIIC, and 1 Stage IV. All of these patients underwent an operation after undergoing breast MRI. Ninety breasts received breast-conserving surgery and 41 received mastectomy. The size of the lesions ranged from 6 54 mm (mean, 18 mm). Upon considering the need to protect personal information, no institutional approval was needed for this kind of retrospective analysis of imaging findings at our institution. MRI MRI was performed on a 1.5T magnet (Signa Horizon; GE Medical Systems, Milwaukee, WI) using a dedicated breast coil (GE Medical Systems). The patients underwent imaging in the prone position. T1-weighted images were obtained in the sagittal plane using a spin-echo sequence (TR/TE [repetition time msec / echo time msec] 500/8 12, field of view [FOV] 16 cm, section thickness 5 8 mm, gap 1 2 mm, matrix or , one signal acquired). T2-weighted images with fat suppression were then obtained in the coronal plane using a fast spin-echo sequence (TR/TE 4000/100, FOV 16 cm, section thickness mm, gap 0.6 mm, matrix , three signals acquired, 3 minutes and 12 seconds acquisition time). The dynamic study was obtained in the coronal plane in all patients. A bolus of gadopentetate dimeglumine (Gd-DTPA) (Magnevist; Schering, Berlin, Germany) was intravenously injected at a dose of 0.1 mmol/kg by hand, followed by a 20-mL saline flush. Between January 1995 and June 1997, a fat-saturated 2D fast spoiled gradient-echo (SPGR) sequence (TR/TE 175/3.9 msec, flip angle 60, FOV 16 cm, section thickness 5 mm, gap 2.5 mm, matrix , two signals acquired, 1 minute and 7 seconds acquisition time) was used in 56 patients. The images were obtained before and every 1.5 minutes after the administration of Gd-DTPA for 10.5 minutes. Between July 1997 and June 1999 a fat-saturated 3D enhanced fast gradient echo (EFGRE) sequence (TR/ TE 7.6/2.6 msec, flip angle 40, FOV 16 cm, section thickness mm, gap 0 mm, matrix , one signal acquired, 1 minute and 5 seconds acquisition time) was used in 67 patients. The images were obtained before, and 1.5, 3, 4.5, 7.5, and 10.5 minutes after the administration of Gd-DTPA. Between July 1999 and December 2004 a fat-saturated 2D SPGR sequence (TR/TE 200/2.6 msec, flip angle 90, FOV 16 cm, section thickness 3 4 mm, gap 0 mm, interleave, matrix , three signals acquired, 1 minute and 57 seconds acquisition time) was used in 188 patients. The images were obtained before and after the first 2-minute scan, the next 2- minute scan (time 2 4 minutes), and the fourth 2-minute scan (time 6 8 minutes) the administration of Gd-DTPA. Image Analysis While referring to the dynamic MR images, the lesions were identified on T2-weighted images. The signal intensity of the lesions on T2-weighted images was compared with the signal intensity of the adjacent breast parenchyma in order to classify the signals into three types: hyperintensity, isointensity, and hypointensity. As there were several cases in which the peripheral portion of the lesion demonstrated a high signal and the central portion presented a low signal, this pattern was referred to as a peripheral hyperintense pattern (Fig. 1) and a classification including a total of four types of signals was obtained. In the postcontrast images the lesion shape and border were evaluated. The overall shape was classified as either round, oval, lobulated, or irregular. The border of the lesion was classified as either smooth, ill-defined, irregular, or spiculated. As previously reported, there are two types of ringlike enhancement of breast carcinomas observed on dynamic MRI that have different sites of appearance and different times of appearance (6). EPE was defined as a thick doughnut-like enhancement in the peripheral portion of the tumor, which was stronger than that in the central portion. EPE is seen in the early (within 2 minutes following the injection of Gd-DTPA) phase. A delayed rim enhancement (DRE) was defined as thin enhancement surrounding the mass which is seen in the intermediate (time 2 4 minutes) or delayed (time 6 8 minutes) phase. In this study we investigated whether the presence of a

3 Peripheral Hyperintense Pattern on T2WI 1119 case a single representative block was chosen. The thickness of the tissue sections was 5 mm. CD34 (1:200 dilution, DakoCytomation, Japan) was used for blood microvessel quantification. First, the slice was observed at a 100 magnification to find the five regions with the highest vascular density, which was the hot-spot. Second, the blood vessels were counted in these five fields at 200 magnification (0.785 mm 2 ) in the central and peripheral portions of each lesion. The mean number of vessels in each of five fields (blood vessel density) was calculated and the ratios of the peripheral to central blood vessel density were calculated for each lesion. Statistical Analysis Figure 1. A 79-year-old female with invasive ductal carcinoma. A spiculated mass shows a peripheral hyperintense pattern on the fat-suppressed-t2 weighted image. peripheral hyperintense pattern on a T2-weighted image correlated with the presence of EPE. All images were newly evaluated retrospectively by two radiologists together until they reached a consensus. All images were displayed on a PACS workstation. A peripheral hyperintense pattern on a T2- weighted image and EPE were assessed on all lesioncontaining slices. Histopathologic Analysis Surgical specimens were cut into 5-mm slices. Using slices stained with hematoxylin-eosin, invasive ductal carcinoma (IDC) was classified into three subtypes (IDC with abundant stroma, IDC with moderate stroma, and IDC with scanty stroma) based on differences in the size of the cancer nests and the amount of stroma, as determined by one pathologist. Adding ductal carcinoma in situ (DCIS), including microinvasive ductal carcinoma, to these, the histopathological subtypes were classified as one of four kinds. When two or more histopathological patterns were seen, then the pathologist classified the case as the predominant one. The existence of fibrosis in the central portion (the longest diameter of the area with fibrosis in more than half of the longest diameter of the tumor) was noted, except for three unqualified lesions because of the removal of the central portion for intraoperative verification. Immunohistochemical Analysis Immunohistochemistry was performed in 17 of the IDCs. The cases showing a comparatively clear EPE, or its absence, by MRI with a section suitable for immunohistochemical analysis were selected. For each To investigate the correlation between signal intensity on T2-weighted images and histopathological subtypes and between peripheral hyperintense pattern on T2-weighted images and other findings on MR images, we used the chi-square test. Fisher s exact test was used to verify the significant difference between the two groups. To investigate the correlation between the existence of peripheral hyperintense pattern on T2- weighted image and the ratio of the peripheral to central blood vessel density, the median values and quartiles were used and the Wilcoxon rank sum test was performed. Statistical analysis was performed using two statistical software packages (SAS-JMP; SAS Institute, Cary, NC. and the R-program). P-values less than 0.05 were considered statistically significant. RESULTS The signal intensities of the breast carcinomas on T2- weighted images were hyperintensity, isointensity, hypointensity, and the peripheral hyperintense pattern for 26 (20%), 66 (50%), 19 (15%), and 20 (15%) lesions, respectively. The correlation between signal intensity on T2-weighted images and the histopathological subtypes are shown in Table 1. No statistical significant difference was observed in the ratio between the signal intensity on T2-weighted images and any of the histopathological subtypes (P ¼ ), but the group with peripheral hyperintense patterns exhibited a significantly higher rate of IDC with abundant stroma and exhibited a significantly low rate of IDC with scanty stroma compared to the hyperintensity group (P ¼ ). Moreover, the group with peripheral hyperintense patterns exhibited a significantly higher rate of IDC with abundant stroma and exhibited a low rate of IDC with scanty stroma compared to the isointensity group (P ¼ ). The correlation between peripheral hyperintense pattern on T2-weighted image and other findings on MR images are shown in Table 2. There were no cases in which a peripheral hyperintense pattern was presented in a mass less than 10 mm in size. No significant differences were observed in the rate of appearance between the presence of a peripheral

4 1120 Kawashima et al. Table 1 Signal Intensity on T2-Weighted Image and Histopathological Subtypes (n¼131) Histopathological subtypes Hyperintensity (n¼26) Signal intensity on T2-weighted image a Isointensity (n¼66) Hypointensity (n¼19) Peripheral hyperintensity b (n¼20) IDC with abundant stroma (n¼70) 9 (34.6) 34 (51.5) 11 (57.9) 16 (80.0) IDC with moderate stroma (n¼20) 7 (26.9) 7 (10.6) 3 (15.8) 3 (15.0) IDC with scanty stroma (n¼22) 7 (26.9) 13 (19.7) 2 (10.5) 0 (0.0) DCIS (n¼19) 3 (11.5) 12 (18.2) 3 (15.8) 1 (5.0) a Signal intensity of lesion relative to that of surrounding breast parenchyma b Hyperintensity in peripheral portion and hypointensity in central portion of the mass IDC, invasive ductal carcinoma; DCIS, ductal carcinoma in situ. Data are observed numbers of lesions and data in parentheses are percentages. hyperintense pattern on a T2-weighted image and the size and shape of the mass (P ¼ , P ¼ ), but a statistically significant difference was observed in the rate of appearance in the periphery of the masses (P ¼ ). In the study between the two groups, the smooth group showed a significantly higher appearance rate of the peripheral hyperintense pattern compared to the ill-defined or irregular group (P ¼ ), and the spiculated group also exhibited a significantly higher appearance rate of the peripheral hyperintense pattern in comparison to the illdefined or irregular group (P < ). In lesions with a peripheral hyperintense pattern on the T2- weighted images the appearance of EPE was significantly higher (P < ) (Fig. 2). In lesions with a peripheral hyperintense pattern on the T2-weighted images, the rates of central fibrosis were significantly higher (P < ) (Table 3, Fig. 2). Of the 17 lesions examined by immunohistochemical analysis, four demonstrated a peripheral hyperintense pattern on T2-weighted images. The correlation of the existence of a peripheral hyperintense pattern on the T2-weighted images and the ratio of the peripheral to central blood vessel density is shown in Table 4. No statistically significant difference was observed between the presence of a peripheral hyperintense pattern on a T2-weighted images and the ratio of the peripheral to central blood vessel density (P ¼ ). DISCUSSION T2-weighted imaging is a basic imaging procedure for MRI that is suitable for inferring the internal properties of a lesion. However, breast MRI is often performed for patients who have already been diagnosed with breast carcinoma, and as a result, T2-weighted imaging, which does not greatly contribute to diagnosis of progress level, tends to be valued less than dynamic MRI. Mucinous carcinoma is one special type of breast carcinoma that exhibits strong high signals on T2-weighted images (1). Yuen et al (5) reported that subtypes of breast carcinoma other than mucinous carcinoma, such as apocrine carcinoma and invasive micropapillary carcinoma, also have a tendency to exhibit strong high signals in the internal area on T2-weighted images, reflecting pathologies such as abundant cytoplasm, edematous stroma, and hemorrhage. However, in their study it was rare for any common type of invasive ductal breast carcinoma to have strong high signals within the tumor on a T2- weighted image, and it is almost impossible to make a diagnosis of a common type of invasive ductal breast carcinoma with confidence based only on the signal intensity on T2-weighted images. Peripheral ring-like enhancement on dynamic contrast enhanced MR imaging is known to be a specific finding for breast carcinoma (3,7 9). However, definitions of the terms and aspects of events have varied between researchers, creating confusion for a long period of time (8 15). Kobayashi et al (6) defined two different types of ring-like enhancement demonstrated by dynamic MRI as follows. One was the stronger enhancement in the peripheral portion of the tumor than that in the central portion, which appeared within 2 minutes after the administration of Gd-DTPA (early phase) and had a certain degree of thickness, and was defined as early peripheral enhancement (EPE). The other was the thin linear enhancement Table 2 Correlation Peripheral Hyperintensity on T2-Weighted Image and Other Findings on MR Images (n¼131) Peripheral hyperintensity on T2-weighted image Presence (n¼20) ] ] ] Absence (n¼111) P-value Size (mm) 0.17 <10 0 (0.0) 17 (15.3) 10, <20 12 (60.0) 58 (52.3) 20 8 (40.0) 36 (32.4) Shape round or oval 7 (35.0) 34 (30.6) lobulated 9 (45.0) 44 (39.6) irregular 4 (20.0) 33 (29.7) Border smooth 10 (50.0) * 49 (44.1) ill-defined or irregular 2 (10.0) 49 (44.1) spiculated 8 (40.0) * 13 (11.7) Early peripheral enhancement < Presence (n¼29) 15 (75.0) 14 (12.6) * Absence (n¼102) 5 (25.0) 97 (87.4) Data are observed numbers of lesions and data in parentheses are percentages. *Significant.

5 Peripheral Hyperintense Pattern on T2WI 1121 Figure 2. A 54-year-old female with invasive ductal carcinoma. a: A spiculated mass shows a peripheral hyperintense pattern on the fat-suppressed T2-weighted image. b: This mass also shows EPE on the fat-suppressed T1-weighted image. c: A photomicrograph shows a star-shaped mass. Cancer tissue is located in the periphery of the mass and the central area is occupied by fibrous tissue. The histopathological subtype was invasive ductal carcinoma with abundant stroma (hematoxylin-eosin stain, 1). rimming the tumor, which was visualized at more than 2 minutes after the administration of Gd-DTPA (intermediate or delayed phase), and was defined as delayed rim enhancement (DRE). In the observations of T2-weighted images during routine reading of MRI, we noted there were cases in which the peripheral portion of the tumor presented high signals of a shape similar to EPE while the central portion presented low signals, and we referred to this pattern as a peripheral hyperintense pattern. In this study, common types of IDC presented isointensity compared to the surrounding mammary gland on T2-weighted images in 50% of the subjects, which was the highest rate. The peripheral hyperintense pattern, which was a focus of this study, was observed at Table 3 Correlation Peripheral Hyperintensity on T2-Weighted Image and the Existence of Central Fibrosis Peripheral hyperintensity on T2-weighted image Presence (n¼19) Absence (n¼109) P-value Central fibrosis < Presence (n¼45) 16 (84.2) 29 (26.6) ] * Absence (n¼83) 3 (15.8) 80 (73.4) Data are observed numbers of lesions and data in parentheses are percentages. *Significant. a relatively high rate of 15%, indicating the existence and distribution of fibrosis with its appearance. Moreover, it was discovered that the properties of the periphery of the masses were also involved with the appearance of the peripheral hyperintense pattern and the frequency of appearance is high in smooth borders and spiculated borders while the frequency of appearance is low in ill-defined borders and irregular borders. These characteristics are almost identical to the images and pathological characteristics of the cases presenting EPE as previously reported (6), so it can be said that the fact that EPE was observed at a very high rate in the subjects with the peripheral Table 4 Correlation Peripheral Hyperintensity on T2-Weighted Image and Blood Vessel Density Peripheral hyperintensity on T2-weighted image Presence (n¼4) Median [25%, 75%] Absence (n¼13) Median [25%, 75%] P-value Blood vessel density [1.245, 3.203] [1.095, 1.385] Ratio of peripheral to central Data are presented as median (percentile) values

6 1122 Kawashima et al. hyperintense pattern in this study was a natural result. Histologically, the appearance of a peripheral hyperintense pattern on T2-weighted images significantly correlated with central fibrosis in the same manner as for the appearance of EPE. Therefore, similar to EPE, it was believed that the presence of central fibrosis was one of the causes of the peripheral hyperintense pattern observed on the T2- weighted images. In general, IDC with abundant stroma has a high degree of malignancy among IDCs and significant invasive growth to the surrounding areas. Moreover, IDC with abundant stroma exhibits nonuniform histological internal properties compared to other types of IDC. In other words, the peripheral portion of the tumor features abundant cellular elements and active blood vessel proliferation. On the other hand, the central portion of the tumor has abundant fibrillary elements, and necrosis is sometimes present as well. In this study, no significant difference was observed in the ratio of the peripheral to central blood vessel density between cases with and without the presence of the peripheral hyperintense pattern on the T2- weighted images. In the previous study, the ratio of the peripheral to central blood vessel density was significantly higher in the cases with EPE than in those without EPE (6). As the number of subjects in this study was low, it cannot be said that a sufficient investigation was conducted, and if a study is conducted with a higher number of subjects, it may be able to demonstrate that the degree of acceleration of neovascularization plays some role in the appearance of a peripheral hyperintense pattern in the same manner as that for the appearance of EPE. It was speculated that no subject with a mass measuring less than 10 mm in diameter presented a peripheral hyperintense pattern because nonuniformity of the internal properties is not obvious when the tumor is small. Moreover, most of the masses with illdefined or irregular borders had distorted shapes and large differences between the long and short diameters, and it was therefore believed that it would be difficult to distinguish based on differences between the histological properties of the periphery and the center. It was speculated that this was the reason for the low rate of appearance of the peripheral hyperintense pattern. Yamaguchi et al (16) focused on the presence of IDC with a large central acellular zone and reported that all seven subjects who underwent MRI exhibited ringlike enhancement on dynamic MRI while in five of the seven subjects the central portions of the masses presented high to very high signal intensity on T2- weighted images. This is interesting because the signal pattern is an inversion of our peripheral hyperintense pattern. In that report, the central portions consisted of an acellular zone with myxomatous matrix, matching the cause of the T2-weighted image high signals of Yuen et al. A combination of ring-like enhancement and high signals in the central portion on a T2-weighted image is a finding observed in cases of breast carcinoma with obvious necrosis in the central portion containing a squamous cell carcinoma or a matrix-producing carcinoma that produces a cartilage matrix. It is hoped that the combination of ringlike enhancement and high signals in the central portion on a T2-weighted image of an IDC as reported by Yamaguchi et al will be discussed separately from the image findings of special types of breast carcinoma containing necrosis and cartilage matrices. In addition, it is hoped that the EPE and the peripheral hyperintense pattern on T2-weighted images that we have defined will also be recognized as characteristic findings that appear in common types of IDC at a constant frequency. There are some limitations to this study. First, only the common types of breast carcinoma were retrospectively evaluated in our study. As a result, we cannot evaluate its usefulness in the differential diagnosis of various kinds of breast mass lesions, including benign lesions. Second, an immunohistochemical analysis could not be performed in all cases because of the lack of suitable surgical specimens for immunohistochemical analysis. In conclusion, the peripheral hyperintense pattern on a T2-weighted image is a characteristic finding observed in common types of IDC, and its appearance strongly correlated with the appearance of EPE. It was thus confirmed that the appearance was affected by various morphologic and histological features including intratumoral fibrosis in the same manner as that observed for the formation of EPE. REFERENCES 1. Kawashima M, Tamaki Y, Nonaka T, et al. MR imaging of mucinous carcinoma of the breast. Am J Roentgenol 2002;179: Hochman MG, Orel SG, Powell CM, et al. Fibroadenomas: MR imaging appearances with radiologic-histopathologic correlation. Radiology 1997;204: Orel SG, Schnall MD, LiVolsi VA, et al. Suspicious breast lesions: MR imaging with radiologic-pathologic correlation. Radiology 1994;190: Kuhl CK, Klaschik S, Mielcarek P, Gieseke J, Wardelmann E, Schild HH. Do T2-weighted pulse sequences help with the differential diagnosis of enhancing lesions in dynamic breast MRI? J Magn Reson Imaging 1999;9: Yuen S, Uematsu T, Kasami M, et al. Breast carcinomas with strong high-signal intensity on T2-weighted MR images: pathological characteristics and differential diagnosis. J Magn Reson Imaging 2007;25: Kobayashi M, Kawashima H, Matsui O, et al. Two different types of ring-like enhancement on dynamic MR imaging in breast cancer: correlation with the histopathologic findings. J Magn Reson Imaging 2008;28: Buadu LD, Murakami J, Murayama S, et al. Breast lesions: correlation of contrast medium enhancement patterns on MR images with histopathologic findings and tumor angiogenesis. Radiology 1996;200: Hachiya J, Seki T, Okada M, Nitatori T, Korenaga T, Furuya Y. MR imaging of the breast with Gd-DTPA enhancement: comparison with mammography and ultrasonography. Radiat Med 1991; 9: Schnall MD, Blume J, Bluemke DA, et al. Diagnostic architectural and dynamic features at breast MR imaging: multicenter study. Radiology 2006;238: Matsubayashi R, Matsuo Y, Edakuni G, Satoh T, Tokunaga O, Kudo S. Breast masses with peripheral rim enhancement on dynamic contrast-enhanced MR images: correlation of MR findings with histologic features and expression of growth factors. Radiology 2000;217:

7 Peripheral Hyperintense Pattern on T2WI Buadu LD, Murakami J, Murayama S, et al. Patterns of peripheral enhancement in breast masses: correlation of findings on contrast medium enhanced MRI with histologic features and tumor angiogenesis. J Comput Assist Tomogr 1997;21: Rubens D, Totterman S, Chacko AK, et al. Gadopentetate dimeglumine-enhanced chemical-shift MR imaging of the breast. Am J Roentgenol 1991;157: Mussurakis S, Gibbs P, Horsman A. Peripheral enhancement and spatial contrast uptake heterogeneity of primary breast tumors: quantitative assessment with dynamic MRI. J Comput Assist Tomogr 1998;22: Sherif H, Mahfouz AE, Oellinger H, et al. Peripheral washout sign on contrast-enhanced MR images of the breast. Radiology 1997; 205: Teifke A, Behr O, Schmidt M, et al. Dynamic MR imaging of breast lesions: correlation with microvessel distribution pattern and histologic characteristics of prognosis. Radiology 2006;239: Yamaguchi R, Furusawa H, Nakahara H, et al. Clinicopathological study of invasive ductal carcinoma with large central acellular zone: special reference to magnetic resonance imaging findings. Pathology International 2008;58:26 30.