Women s Imaging Original Research

Transcription

1 Women s Imaging Original Research Woodhams et al. Diffusion-Weighted Imaging of Breast Carcinoma Women s Imaging Original Research Reiko Woodhams 1,2,3,4 Satoko Kakita 1 Hirofumi Hata 1 Keiichi Iwabuchi 1 Shigeaki Umeoka 2 Carolyn E. Mountford 3 Hiroto Hatabu 2 Woodhams R, Kakita S, Hata H, et al. 1 Departments of Radiology, Pathology, and Surgery, Kitasato University School of Medicine and Kitasato University Hospital, Sagamihara, Kanagawa, Japan. 2 Department of Radiology, Brigham and Women s Hospital, Harvard Medical School, Boston, MA. 3 Centre for Clinical Spectroscopy, Department of Radiology, Brigham and Women s Hospital and Harvard Medical School, Boston, MA. 4 Present address: Department of Radiology, Kitasato University School of Medicine, , Kitasato, Sagamihara, Kanagawa, Japan. Address correspondence to R. Woodhams (reikow@hotmail.co.jp). AJR 2009; 193: WOMEN S IMAGING Keywords: ADC valve, breast, diffusion-weighted imaging, MRI, mucinous carcinoma of the breast DOI: /AJR Received August 15, 2008; accepted after revision January 9, X/09/ American Roentgen Ray Society Diffusion-Weighted Imaging of Mucinous Carcinoma of the Breast: Evaluation of Apparent Diffusion Coefficient and Signal Intensity in Correlation With Histologic Findings OBJECTIVE. The purposes of this study were to compare the apparent diffusion coefficient (ADC) of mucinous carcinoma of the breast with that of other breast tumors and to analyze correlations between signal intensity on diffusion-weighted images and the histologic features of mucinous carcinoma. SUBJECTS AND METHODS. Two hundred seventy-six patients with 277 lesions, including 15 mucinous carcinomas (13 pure type, two mixed type), 204 other malignant tumors, and 58 benign lesions, were examined with 1.5-T MRI at b values of 0 and 1,500 s/ mm 2. The correlations between cellularity and ADC, homogeneity of signal intensity on diffusion-weighted images, and histopathologic findings were analyzed. The difference was statistically significant (p < 0.05). RESULTS. The mean ADC of mucinous carcinoma (1.8 ± mm 2 /s) was statistically higher than that of benign lesions (1.3± mm 2 /s) and other malignant tumors (0.9 ± mm 2 /s) (p < 0.001). The ADC of pure type mucinous carcinoma (1.8 ± mm 2 /s) was higher than that of mixed type mucinous carcinoma (1.2 ± mm 2 /s) (p < 0.001) and other histologic types (p > 0.05). The correlation between mean cellularity and the ADC of mucinous carcinoma was significant (ρ s = 0.754; p = 0.001). The homogeneity of signal intensity on diffusion-weighted images correlated with the homogeneity of histologic structures of mucinous carcinoma (p < 0.001; κ = 0.826). CONCLUSION. Mucinous carcinoma can be clearly differentiated from other breast tumors on the basis of ADC. The low signal intensity of mucinous carcinoma on diffusionweighted images appears to reflect the presence of mucin and low cellularity. High signal intensity on diffusion-weighted images may reflect the presence of fibrovascular bundles, increased cell density, or a combination of these features. T he high sensitivity of diffusionweighted imaging (DWI) for malignant breast tumors is well recognized. According to previous reports [1, 2], the diffusivity of water molecules is restricted in environments of high cellularity, intracellular and extracellular edema, high viscosity, and a high degree of fibrosis because these conditions become barriers to the movement of water molecules. Malignant tumors, including breast cancer, are usually characterized by high cellularity. DWI depicts the restriction of water molecule diffusivity in malignant tissue as high signal intensity and yields a low apparent diffusion coefficient (ADC), which is a quantitative index of the diffusivity of water molecules [2 4]. Several studies have shown that benign and malignant tumors can be differentiated on the basis of ADC because a higher ADC may indicate benign lesions and a lower ADC, malignant lesions [1, 3, 4]. Studies [3, 4] of breast DWI, however, have shown that the ADC of mucinous carcinoma of the breast is considerably higher than that of other malignant tumors of the breast and may overlap the ADCs of benign tumors and normal breast tissue. Mucinous carcinoma of the breast is a relatively rare tumor that accounts for 1 7% of all cases of breast cancer. It is defined by a histologically distinctive pattern characterized by proliferation of clusters of generally uniform round cells floating in large amounts of extracellular mucus, as described in the World Health Organization international histologic classification of tumors [5]. There are two types of mucinous carcinoma. In the pure type, all of the tumor cells are completely surrounded by mucin, and the tumor 260 AJR:193, July 2009

2 Diffusion-Weighted Imaging of Breast Carcinoma does not have any invasive ductal components. In the mixed type, invasive ductal carcinoma is present but not embedded in extracellular mucin. Pathologic specimens of mucinous carcinoma of the breast vary in the amount of mucin and cells and are described as having the hypercellular and hypocellular properties of mucinous carcinoma of the breast [5]. We hypothesized that the ADC and DWI signal intensity of mucinous carcinoma of the breast vary depending on amount of mucin, cellularity, and fibrous stroma. This hypothesis was based on results of previous studies of tumors of various organs, including the brain, pancreas, and other soft tissues, that pointed to the influence of cellularity, fibrosis, and mucous matrix on ADC [6 9]. The aim of this study was to compare the ADC of mucinous carcinoma of the breast with the ADCs of other breast tumors. We also examined the correlation between signal intensity on diffusion-weighted images and the histopathologic features of mucinous carcinoma of the breast. Subjects and Methods Subjects This study was approved by the local institu tional review board, and written informed consent was obtained from all participants. At our institution, all patients who are scheduled for breast MRI undergo the same protocol except those unable to assume the prone position. Two hundred seventy-six patients (277 lesions) who underwent consecutive routine breast MRI examinations with the same protocol that included DWI from May 2005 to September 2007 participated in this study. All patients with mucinous carcinoma underwent ultrasound-guided core needle biopsy after breast MRI and then underwent surgery. Pathologic diagnoses of other malignant and benign lesions were made at core needle biopsy or surgery. Patients with pacemakers, an allergy to the MRI contrast agent, or severe renal or hepatic dysfunction had been excluded from the study. The group of 276 participants had 277 lesions. Fifteen patients (mean age, 59 years; range, years) had mucinous carcinoma (13 pure type, two mixed type). The other lesions were 204 malignant tumors and 58 benign tumors. MRI Routine breast MRI was performed with a 1.5-T MRI system (HDx, GE Healthcare) and a dedicated eight-channel breast-array coil on patients in the prone position. After the positioning acquisition, DWI of both breasts was performed in the axial plane with the single-shot echo-planar imaging sequence to reduce motion artifacts (TR/ TE, 9,500/89; section thickness, 4.0 mm; interslice gap, 0 mm; number of signals averaged, 8; field of view, 320 mm 2 ; matrix size, ; no parallel imaging because of the possibility of unfold failure and lower signal-to-noise ratio). Data were collected at b values of 0 and 1,500 s/ mm 2. The b value of 1,500 s/mm 2 was used to minimize the T2-weighted shine-through effect and to suppress the signal from normal breast parenchyma [10]. In addition, a frequencyselective radiofrequency pulse was used before the pulse sequence to suppress the strong signal from lipids, reducing chemical shift artifacts. The acquisition time for DWI was 5 minutes. After DWI, unilateral examination of the index breast consisted of a sagittal fast spin-echo T2- weighted sequence with fat suppression (4,000/90; section thickness, 5 mm; interslice gap, 2 mm; number of signals averaged, 2; field of view, 200 mm; matrix size, ) and a sagittal 3D T1-weighted fast gradient-echo sequence (16.3/2.1; flip angle, 15 ; section thickness, 2 mm; number of signals averaged, 1.5; field of view, 200 mm; matrix size, ) with chemical shift selective suppression technique for active suppression of the fat signal. For the dynamic contrast-enhanced portion of the study, a bolus injection of 0.10 mmol/kg of body weight of gadopentetate dimeglumine (Magnevist, Bayer HealthCare) was administered through the right antecubital vein with an automatic injector (Sonic Shot 50, Nemoto Kyorindo). The injection rate was 2 ml/s, and the contrast agent was followed by a 20-mL saline flush. Acquisitions were performed before and 90 seconds (early phase) and 300 seconds (delayed phase) after contrast injection. An ADC map was generated on a workstation (Advantage Windows 4.1, GE Healthcare) from diffusion-weighted images with b values of 0 and 1,500 s/mm 2. Assessment of ADC, Diffusion-Weighted Images, and Corresponding Histopathologic Structures Measurement of ADC of all tumors Calculation of the ADC of all tumors was by one radiologist with 6 years of experience in breast imaging, including MRI. This investigator was not aware of the histopathologic diagnosis in any case. The location of tumors on the diffusionweighted images and ADC maps was determined on fat-suppressed T2-weighted and contrastenhanced T1-weighted images. The region of interest covering the whole tumor was placed on the ADC map, and the ADC was recorded. If the presence of hemorrhage or necrosis was suspected on the basis of the findings on the unenhanced T1- and T2-weighted images, the region of interest did not include that area [11]. Regions of interest were placed inside the tumor on all slices that displayed the tumor and were averaged. Calculation of the ADC of each tumor was repeated, and the values were averaged. We disregarded satellite lesions coexisting with main tumors. The mean ADCs of mucinous carcinomas, other malignant breast tumors, and benign breast tumors were compared. Pathologic analysis of mucinous carcinoma Each resected specimen of mucinous carcinoma was obtained with four sutures of various lengths with the axilla, nipple side, cranial margin, and caudal margin for orientation. Surgical specimens were cut into 5-mm slices, fixed in 10% neutralbuffered formalin, and processed for histologic examination. The specimens were stained with H and E and evaluated by a pathologist expert in breast pathology. Mucinous carcinomas of the two subtypes, pure and mixed, were examined. The cellularity of mucinous carcinoma was analyzed according to the methods of previous studies [1] with National Institutes of Health Image J 1.40g software with a Windows operating system. The original magnification of the specimens was 200. A pathologist analyzed five sample views randomly chosen from the specimens for cellularity values. The mean cellularity of each tumor was calculated by averaging of these five values. The correlation between the mean cellularity value and ADC was statistically analyzed. Comparison between DWI and histopathologic findings of mucinous carcinoma We matched the tumor on the diffusion-weighted images and the specimens by referring to the pathologic diagram that displayed each orientation of each specimen in the resected breast tissue. On the basis of previous reports [1, 6, 9, 12] of the correlation between ADC and the presence of mucus, degree of cellularity, and the degree of fibrosis in other organs, we hypothesized that the pattern of signal intensity on diffusion-weighted images would reflect the histopathologic structure of mucinous carcinoma. In this study, signal intensity on diffusion-weighted images was considered inversely proportional to ADC because the effect of T2-weighted shine-through was considered to be at a minimum owing to the b value of 1,500 s/mm 2 [13]. Therefore, we hypothesized that the specimens with high DWI signal intensity would correlate with high-cellularity lesions or fibrovascular tissue and the specimens with low DWI signal intensity would correlate with lowcellularity, mucin-rich lesions. The homogeneity of the signal intensity pattern of mucinous carcinoma on diffusion-weighted images was categorized as homogeneously high AJR:193, July

3 Woodhams et al. TABLE 1: Comparison of Apparent Diffusion Coefficients of Mucinous Carcinoma, Other Malignant Tumors, and Benign Tumors Histopathologic Type No. Mean Apparent Diffusion Coefficient ( 10 3 mm 2 /s) Mucinous carcinoma ± 0.4 Pure type ± 0.3 Mixed type ± 0.2 Other malignant tumor ± 0.2 Ductal carcinoma in situ ± 0.2 Lobular carcinoma in situ ± 0.0 Invasive ductal carcinoma ± 0.2 Invasive lobular carcinoma ± 0.2 Other ± 0.3 Benign tumor ± 0.3 Phyllodes benign ± 0.1 Fibroadenoma ± 0.1 Fibrocystic disease ± 0.1 Intraductal papilloma ± 0.1 Other benign tumor ± 0.2 signal intensity, which was higher than the signal intensity of the rib or sternum; homogeneously low signal intensity, which was lower than or equal to the signal intensity of the rib or sternum; and heterogeneous signal intensity, a composite of high and low signal intensity. Each lesion was categorized by two radiologists with more than 6 years of experience in MRI, including breast MRI. If there was a discrepancy between the two readers, they discussed the findings and made a final decision. The homogeneity of histologic structures was categorized as homogeneously high cellularity, which consisted of uniform high-cellularity lesions with or without fibrovascular tissue; homogeneously low cellularity, which consisted of low cellularity, mucin-rich lesions not including fibrovascular tissue; and heterogeneous structures, which consisted of a mixture of low cellularity, mucin-rich lesions and high-cellularity lesions or fibrovascular tissue. Histologic assessment was per formed by a pathologist. The correlation between homogeneity of signal intensity on diffusion-weighted images and homogeneity of histologic structure was assessed. The lesions categorized as having heterogeneous signal intensity on diffusion-weighted images were analyzed in detail with regard to histopathologic characteristics. Statistical Analysis Statistical analysis was performed with JMP software (version 7.0, SAS Institute). The mean ADCs of mucinous carcinomas, other breast malig nant tumors, and benign breast tumors were com pared by the Student s t test. The mean ADCs of pure mucinous carcinoma and mixed mucinous carcinoma were compared by the Mann-Whitney U test. A value of p < 0.05 was considered significant. Spearman s correlation was used to evaluate the relation between the mean cellularity and the ADC of mucinous carcinoma. A value of p < 0.05 was considered to indicate a statistically significant dif ference. Correlation between signal intensity pattern on diffusion-weighted images and histologic homo geneity was validated with kappa statistics. Apparent Diffusion Coefficient ( 10 3 mm 2 /s) Benign tumor Mucinous carcinoma Other malignant tumor Fig. 1 Scatter diagram shows apparent diffusion coefficients of benign tumors (left), mucinous carcinoma (center), and other malignant tumors (right). Results The mean ADC of mucinous carcinoma was 1.8 ± 0.4 [SD] 10 3 mm 2 /s, of other malignant tumors was 0.9 ± mm 2 /s, and of benign tumors was 1.3 ± mm 2 /s. The mean ADC of mucinous carcinoma was statistically greater than that of other malignant tumors and that of benign tumors (p < 0.001) (Table 1) (Fig. 1). Clinical data, histologic subcategory, ADCs, and cellularity are summarized in Table 2. Thirteen of the 15 mucinous carcinomas were of the pure type, and two were of the mixed type. The mean ADC of pure mucinous carcinoma was 1.8 ± mm 2 /s, and that of mixed mucinous carcinoma was 1.2 ± mm 2 /s. The mean ADC of pure mucinous carcinoma was significantly greater than that of mixed mucinous carcinoma (p < 0.001) and of the other individual histologic types of breast tumors (p < 0.05). The ADC of mixed mucinous carcinoma was not significantly different from that of any other histologic type of breast tumor. Correlation between the mean cellularity and ADC of mucinous carcinoma was inverse and significant (ρ s = 0.754; p = 0.001) (Fig. 2). On diffusion-weighted images, mucinous carcinoma of the breast had homogeneously high signal intensity in one case, homogeneously low signal intensity in three patients, and heterogeneity in 11 patients (Table 3). In the comparison of the signal intensity pattern of mucinous carcinoma on diffusion-weighted images and the histologic homogeneity of the specimens, the only tumor with homogeneously high signal intensity on diffusionweighted images exhibited homogeneously high cellularity in the specimens (Fig. 3). Two of the three lesions with homogeneously low signal intensity on diffusion-weighted images were found to have homogeneously low cellularity and mucin-rich content (Fig. 4). In the third patient, homogeneous mucinrich content with thin fibrovascular tissue demarcating the border of the tumor was found. Specimens of all 11 tumors with heterogeneous signal intensity on diffusion-weighted images had a combination of low cellularity, mucin-rich content, and high-cellularity content or a combination of low cellularity, mucin-rich content, and fibrovascular tissue (Fig. 5). The correlation between the homogeneity of the signal-intensity patterns on diffusion-weighted images and the histologic homogeneity of the specimens was significant (p < 0.001; κ = 0.826). 262 AJR:193, July 2009

4 Diffusion-Weighted Imaging of Breast Carcinoma TABLE 2: Clinical and Histologic Details and Apparent Diffusion Coefficients of 15 Mucinous Carcinomas of the Breast Patient Age (y) Subtype Regarding the 11 tumors with heterogeneous signal intensity on diffusion-weighted images, we classified the pattern of the highsignal-intensity compartment into linear and nodular patterns to correlate these features with the histologic structure of the lesion. These 11 cases with heterogeneous signal intensity on diffusion-weighted images and the corresponding histologic structures are summarized in Table 4. Discussion To our knowledge, this study is the first evaluation of the correlation between the DWI and the histologic findings of mucinous carcinoma of the breast. Mucinous carcinoma of the breast had a much higher ADC than other malignant breast tumors. Abundant mucin and low cellularity may be responsible for this effect. This finding has been corroborated in studies [6, 7] of tumors in various organs, including the brain, pancreas, and other soft tissue, all findings pointing to the influence of the mucous matrix on ADC. Findings of studies [14, 15] showing an inverse correlation between cellularity and ADC in lesions of various organs, including breast tumors, suggest that the characteristically low cellularity of mucinous carcinoma of the breast contributes to the high ADC. The ADC of mucinous carcinoma also was higher than that of benign breast tumors. However, some benign breast tumors, Apparent Diffusion Coefficient ( 10 3 mm 2 /s) Mean Cellularity (%) Size (mm) 1 62 Pure 2.0 ± ± Pure 1.7 ± ± Pure 2.1 ± ± Pure 1.8 ± ± Pure 1.9 ± ± Pure 2.0 ± ± Pure 1.2 ± ± Pure 1.1 ± ± Pure 1.5 ± ± Pure 1.8 ± ± Pure 2.4 ± ± Pure 2.2 ± ± Pure 2.0 ± ± Mixed 1.3 ± ± Mixed 1.1 ± ± particularly fibroadenoma and benign phyllodes tumor (Table 1), have high ADCs overlapping those of mucinous carcinoma. Myxomatous or edematous stroma, which is sometimes found in fibroadenoma and phyllodes tumor on histopathologic images, may be the cause of high ADCs [16 18]. Articles [18, 19] have described the characteristics of mucinous carcinoma on conventional MRI. They mention the similarity between fibroadenoma and pure mucinous carcinoma on T2-weighted images and contrast-enhanced T1-weighted images. The comprehensive approach of DWI, ADC calculation, and conventional MRI to increase specificity is a subject for further study. Cysts were not included in our benign lesions. The ADCs of simple cysts can overlap those of pure mucinous carcinoma [20]. On diffusion-weighted images, the appearance of pure mucinous carcinoma, which has homogeneously low signal intensity, may Apparent Diffusion Coefficient ( 10 3 mm 2 /s) Cellularity (%) be similar to the that of cysts. Complicated cysts, such as highly condensed cysts and bloody cysts, can mimic the DWI findings and ADCs of mixed mucinous carcinoma and those of other malignant tumors because they have low ADCs [21, 22]. These issues need further study. The number of mixed mucinous carcinomas in our study was too small for reliable conclusions about ADCs, but the two mixed mucinous carcinomas had significantly lower ADCs than pure mucinous carcinoma. According to general pathologic findings [16, 17], the cellularity of mixed mucinous carcinoma tends to be greater than that of pure mucinous carcinoma. A previous study [23] showed that the proportion of extracellular mucin in mixed mucinous carcinoma was less than that of pure mucinous carcinoma. Our study also showed an inverse correlation between the cellularity of mucinous carcinoma of the breast and ADC (Fig. 2). Therefore, Fig. 2 Scatter diagram shows inverse correlation between tumor cellularity and apparent diffusion coefficient. Spearman s ρ = 0.754, p = TABLE 3: Correlation Between Homogeneity of Signal Intensity on Diffusion-Weighted Images and Histologic Homogeneity of Mucinous Carcinoma Specimens (n = 15) Histologic Homogeneity Homogeneity of Signal Intensity on Diffusion-Weighted Images Uniformly High Uniformly Low Heterogeneous Uniformly high cellularity Uniformly mucin-rich, low cellularity Heterogeneous structure Note κ = 0.826; p < AJR:193, July

5 Woodhams et al. A Fig year-old woman with mixed mucinous carcinoma of breast. A, Isotropic diffusion-weighted image with b value of 1,500 s/mm 2 shows tumor as high-signal-intensity nodule with homogeneous structure. B, Apparent diffusion coefficient map shows area with low coefficient of mm 2 /s (arrows). C, Photomicrograph of specimen shows dense tumor cell package with few surrounding mucous lakes. Intersecting thick fibrovascular bundles are evident. Mean cellularity is 67.0%. (H and E, 200) Fig year-old woman with pure mucinous carcinoma of breast. A, Isotropic diffusion-weighted image with b value of 1,500 s/mm 2 shows tumor with homogeneously low signal intensity (arrows) approximate to that of rib (arrowhead). Margin of tumor is indistinct. B, Apparent diffusion coefficient map shows tumor as distinctive red area (arrows) with high coefficient of mm 2 /s. C, Photomicrograph of histopathologic specimen shows sparse tumor cell distribution surrounded by abundant mucus (arrows). No fibrous stroma surrounds tumor. Mean cellularity of tumor is 4.0%. (H and E, 200) A B B it seems likely that the high cellularity and lower mucin matrix of mixed mucinous carcinoma contribute to its low ADC. If the tumor is already suspected of being mucinous carcinoma at MRI that includes DWI, a lower ADC of mucinous carcinoma may indicate the tumor is mixed mucinous carcinoma. It may not be possible to differentiate mixed and pure mucinous carcinoma with conventional MRI of the breast [24]. Mixed mucinous carcinoma has a worse prognosis and a higher incidence of axillary lymph node metastasis than does pure mucinous carcinoma [23]. Thus the DWI features and ADC of mucinous carcinoma may supply additional information for prognosis. The diagnosis of mixed mucinous carcinoma can be difficult, however, if there is no information suggesting the possibility of the presence of mucinous carcinoma. There was no significant difference between the ADC of mixed mucinous carcinoma and that of other histologic types of breast tumors except pure mucinous carcinoma. This issue needs further evaluation. Our study showed a significant correlation between signal intensity pattern on diffusion-weighted images and histologic features. The finding of a low-signal-intensity compartment is evidence of the presence of a mucin-rich, low cellularity compartment. The study also showed that not only high cellularity but also fibrous stroma is a cause of high signal intensity on diffusion-weighted images. This correlation is consistent with the findings in several previous reports [9, 12, 25] on the inverse correlation between ADC and degree of fibrosis. The high cellularity may narrow the free space for water diffusion, and fibrous tissue works as a C C 264 AJR:193, July 2009

6 Diffusion-Weighted Imaging of Breast Carcinoma Fig year-old woman with pure mucinous carcinoma of breast. A, Isotropic diffusion-weighted image with b value of 1,500 s/mm 2 shows tumor has heterogeneous signal intensity consisting of lobulated outline demarcated by high-signal-intensity margin and low signal intensity inside tumor. B, Apparent diffusion coefficient (ADC) map shows high ADC component inside tumor surrounded by low ADC component at periphery (arrows) corresponding to high signal intensity in A. ADC is mm 2 /s. C, Photomicrograph of representative portion of histopathologic specimen shows thick fibrotic lobulated border at periphery corresponding to high-signal-intensity outline of tumor in A. Inside of tumor shows mucin lake containing tumor clusters. Cellularity inside tumor is 27.5%. (H and E, 200) TABLE 4: Correlation Between Pattern of High Signal Intensity on Diffusion-Weighted Images and Histologic Structure in Tumors With Heterogeneous Signal Intensity (n = 11) Patient Subtype Pattern of High Signal Intensity 1 Pure Linear Fibrovascular tissue boundary and obstacle to the diffusibility of water molecules. To our knowledge, this study is the first to show the possible correlation between high signal intensity on diffusion-weighted images and the presence of fibrous stroma in breast tumors. Because there appears to be a correlation between the presence of a high-signalintensity lesion on diffusion-weighted images and its corresponding histologic structure, the morphologic features of a high-signal-intensity lesion may also represent the histologic structure. Linear high signal intensity is highly likely to be a sign of the presence of A Histologic Structure Accompanied by Mucin-Rich Area 2 Pure Linear, nodular Fibrovascular tissue, high cellularity 3 Pure Linear Fibrovascular tissue 4 Pure Linear Fibrovascular tissue 5 Pure Linear Fibrovascular tissue 6 Pure Linear Fibrovascular tissue 7 Pure Linear Fibrovascular tissue 8 Pure Linear Fibrovascular tissue 9 Pure Nodular Fibrovascular tissue 10 Pure Nodular High cellularity 14 Mixed Nodular Fibrovascular tissue, high cellularity B fibrous tissue, and nodular high signal intensity may be a sign of a the presence of either fibrous tissue or a high-cellularity lesion. It is not yet known whether the nature of the fibrovascular tissue on its own is a cause of high signal intensity on diffusion-weighted images. The neighboring structure, which is the mucin matrix in mucinous carcinoma, and infiltrating inflammatory cells and other factors may interact in concert with the fibrous tissue to cause high signal intensity. This issue needs further investigation. One of the limitations of this study was the premise of the minimum T2-weighted shinethrough effect at DWI with a b value of 1,500 s/mm 2. T2-weighted shine-through cannot be eliminated even with a higher b value. The contrast of the signal intensity between solid lesions, mucin lesions, and water lesions is larger with a higher b value than with a lower b value, as shown in a previous study of evaluation of pancreatic cystic tumors at a b value of 1,000s/mm 2 [21]. Another limitation was that the number of mixed mucinous carcinomas was too small for definitive comparison with pure mucinous carcinoma and other histologic types of breast tumors. The number of subjects with each benign tumor also was small. We need a larger number of subjects with each histologic type to reach a more reliable conclusion about comparison of ADCs between histologic types. We did not analyze the ADC of each signal component inside the mucinous carcinomas at DWI. Some of the tumors were small, and the ADCs of the components might have been inaccurate because of the partial volume effect. Pathologic specimens were not sliced according to the slice direction of DWI. Therefore, one-to-one correspondence between components on the tumor on diffusion-weighted images and in the histologic specimens was not strictly achieved. The location and outline of each signal-intensity component and those of the expected corresponding histologic structure were matched. We used bilateral DWI and unilateral conventional T1- and T2-weighted imaging and C AJR:193, July

7 Woodhams et al. contrast-enhanced imaging because at the time we were not satisfied with the quality of transverse bilateral breast dynamic images with volume imaging for breast assessment (Vibrant, GE Healthcare) compared with unilateral sagittal images. Therefore, we did not use bilateral imaging for dynamic studies. The quality of bilateral imaging has improved, and we are performing bilateral imaging for all sequences. DWI may be a helpful and supportive tool for conventional MRI, especially if integrated with standard metrics for lesions detected with breast MRI, such as signal intensity on T1- and T2-weighted images, morphologic features of lesions, and contrast enhancement kinetics. We suspect that better sensitivity and specificity can be achieved, but larger studies integrating ADC and DWI data with conventional MRI are required. We conclude that most mucinous carcinomas of the breast have a markedly higher ADC than other malignant breast tumors. The ADC and signal intensity pattern reflect the histologic characteristics of cellularity, fibrovascular tissue, and mucin-rich content. Thus, DWI is useful for evaluation of mucinous carcinoma of the breast. Acknowledgments We thank Peter Stanwell, physicist in the department of radiology, for technical advice and Alba Cid and Dale Woodhams for editorial assistance. References 1. Guo Y, Cai YQ, Cai ZL. Differentiation of clinically benign and malignant breast lesions using diffusion weighted imaging. J Magn Reson Imaging 2002; 16: Koh DM, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR 2007; 188: Woodhams R, Matsunaga K, Iwabuchi K, et al. Diffusion-weighted imaging of malignant breast FOR YOUR INFORMATION tumors: the usefulness of apparent diffusion coefficient (ADC) value and ADC map for the detection of malignant breast tumors and evaluation of cancer extension. J Comput Assist Tomogr 2005; 29: Woodhams R, Matsunaga K, Kan S. ADC mapping of benign and malignant breast tumors. Magn Reson Med Sci 2005; 4: Tavassoli FA, Devilee P, eds. Pathology and genetics: tumours of the breast and female genital organs. Lyon, France: IARC Press, Irie H, Honda H, Kuroiwa T, et al. Measurement of the apparent diffusion coefficient in intraductal mucin-producing tumor of the pancreas by diffusion-weighted echo-planar MR imaging. Abdom Imaging 2002; 27: Maeda M, Matsumine A, Kato H, et al. Soft-tissue tumors evaluated by line-scan diffusion-weighted imaging: influence of myxoid matrix on the apparent diffusion coefficient. J Magn Reson Imaging 2007; 25: Hayashida Y, Hirai T, Morishita S, et al. Diffusion-weighted imaging of metastatic brain tumors: comparison with histologic type and tumor cellularity. Am J Neuroradiol 2006; 27: Annet L, Peeters F, Abarca-Quinones J, Leclercq I, Moulin P, Van Beers BE. Assessment of diffusion-weighted MR imaging in liver fibrosis. J Magn Reson Imaging 2007; 25: Kuroki Y, Nasu K. Advances in breast MRI: diffusion-weighted imaging of the breast. Breast Cancer 2008; 15: Atlas SW, DuBois P, Singer MB, Lu D. Diffusion measurements in intracranial hematomas: implications for MR imaging of acute stroke. Am J Neuroradiol 2000; 21: Boulanger Y, Amara M, Lepanto L, et al. Diffusion-weighted MR imaging of the liver of hepatitis C patients. NMR Biomed 2003; 16: Geijer B, Sundgren PC, Lindgren A, Brockstedt S, Stahlberg F, Holtas S. The value of b required to avoid T2 shine-through from old lacunar infarcts in diffusion-weighted imaging. Neuroradiology 2001; 43: Guo AC, Cummings TJ, Dash RC, Provenzale JM. Lymphomas and high-grade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 2002; 224: Hatakenaka M, Soeda H, Yabuuchi H, et al. Apparent diffusion coefficients of breast tumors: clinical application. Magn Reson Med Sci 2008; 7: Tavassoli FA. Mucinous carcinoma: colloid type. In: Pathology of the breast. Stamford, CT: Appleton & Lange, 1999: Rosen PP. Fibroepithelial neoplasms. In: Rosen s breast pathology. Philadelphia, PA: Lippincott Williams & Wilkins, 2001: 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: Okafuji T, Yabuuchi H, Sakai S, et al. MR imaging features of pure mucinous carcinoma of the breast. Eur J Radiol 2006; 60: Marini C, Iacconi C, Giannelli M, Cilotti A, Moretti M, Bartolozzi C. Quantitative diffusionweighted MR imaging in the differential diagnosis of breast lesion. Eur Radiol 2007; 17: Inan N, Arslan A, Akansel G, Anik Y, Demirci A. Diffusion-weighted imaging in the differential diagnosis of cystic lesions of the pancreas. AJR 2008; 191: Fischbein NJ, Roberts TP, Dillon WP. Bleed or stroke? Diffusion measurements in intracranial hematomas. Am J Neuroradiol 2000; 21: Komaki K, Sakamoto G, Sugano H, Morimoto T, Monden Y. Mucinous carcinoma of the breast in Japan: a prognostic analysis based on morphologic features. Cancer 1988; 61: Kawashima M, Tamaki Y, Nonaka T, et al. MR imaging of mucinous carcinoma of the breast. AJR 2002; 179: Muraoka N, Uematsu H, Kimura H, et al. Apparent diffusion coefficient in pancreatic cancer: characterization and histopathological correlations. J Magn Reson Imaging 2008; 27: Got a few minutes for a visit? Stop by the American Roentgen Ray Society s online women s imaging community, which features exclusive content and resources by and for women s imaging specialists, including articles, electronic exhibits, webcasts, current news, and job and meeting listings. ARRS provides these valuable resources as a free service just for you. Visit us at AJR:193, July 2009