HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2413051470v1 241/3/702    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Yabuuchi, H. Articles by Honda, H. Search for Related Content PubMed PubMed Citation Articles by Yabuuchi, H. Articles by Honda, H.

Phyllodes Tumor of the Breast: Correlation between MR Findings and Histologic Grade¹

¹ From the Departments of Clinical Radiology (H.Y., H.S., Y.M., T.O., M.H., H.H.), Anatomic Pathology (T.E.), Surgery and Oncology (S.K.), and Surgery and Science (E.T.), Graduate School of Medical Sciences; and Department of Health Sciences, School of Medicine (S.S.), Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan; and Departments of Breast Surgery (S.O.) and Pathology (K.N.), National Kyushu Cancer Center, Fukuoka, Japan. Received September 1, 2005; revision requested November 4; revision received February 6, 2006; final version accepted April 7. Address correspondence to H.Y. (e-mail: yabuuchi{at}radiol.med.kyushu-u.ac.jp).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Purpose: To retrospectively evaluate the magnetic resonance (MR) imaging findings of phyllodes tumor of the breast and to compare these findings with the histologic grade.

Materials and Methods: Institutional review board approval and informed consent were obtained. The authors reviewed the MR findings in 30 female patients aged 16–73 years (mean, 40.2 years) with surgically confirmed phyllodes tumors. Analyzed MR findings included tumor shape, margin, internal enhancement, and size; signal intensity (SI) of tumor higher than that of normal breast tissue on T1-weighted images; SI of tumor lower than or equal to that of normal tissue on T2-weighted images; cyst wall appearance; kinetic curve assessment; and apparent diffusion coefficient (ADC). The MR findings and histologic grade were statistically analyzed to determine whether any correlations existed. Significant MR findings were compared with histopathologic findings.

Results: Nineteen benign, six intermediate (characterized by five to nine cell reproductions at 10 high-power fields, pushing or infiltrative margins, moderate stromal cellularity, and atypia and overgrowth), and five malignant phyllodes tumors were assessed. Irregular cyst wall (P = .003), tumor SI lower than or equal to normal tissue SI on T2-weighted images (P = .005), and low ADC (P = .001) correlated significantly with histologic grade. Tumor SI higher than normal tissue SI on T1-weighted images was more frequent in the malignant (in three of five tumors) and intermediate (in three of six tumors) groups than in the benign group (in two of 19 tumors); however, it was not a significant finding (P = .024). Tumor SI higher than normal tissue SI on T1-weighted images and irregular cyst wall corresponded histopathologically to hemorrhagic infarction and necrosis, respectively. Tumor SI lower than or equal to normal tissue SI on T2-weighted images and low ADC corresponded histopathologically to stromal hypercellularity. Other findings were not significant.

Conclusion: Several MR findings can be used to help determine the histologic grade of phyllodes breast tumors.

© RSNA, 2006

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Phyllodes tumor of the breast is a rare tumor that accounts for less than 1% of all breast tumors and 2%–3% of fibroepithelial neoplasms (1). These tumors are composed of epithelial elements and a connective tissue stroma. Approximately 20%–50% of phyllodes tumors are reported to be malignant (2).

Treatment of both benign and malignant phyllodes breast tumors requires complete surgical excision with wide margins owing to the high recurrence rate in patients with resection margins of less than 1 cm around the primary tumor. This recurrence rate applies to both benign and malignant phyllodes tumors (3). Buchanan (4) proposed an algorithm for the proper management of phyllodes tumors, which consists of needle biopsy, tumor excision with confirmation that the margin is histopathologically negative for tumors cells, and simple mastectomy for tumors larger than 5 cm in diameter and those of any size found to be malignant or intermediate (characterized by five to nine cell reproductions at 10 high-power fields, pushing or infiltrative margins, moderate stromal cellularity, and atypia and overgrowth). Therefore, information regarding the histologic grade of a phyllodes tumor influences the surgical course of treatment.

Fine-needle aspiration cytology may not be helpful for differentiating benign and malignant phyllodes tumors because this tumor is frequently composed of various changes in the stroma within the same tumor (5). Moreover, fine-needle aspiration cytology cannot be used to distinguish a fibroadenoma from a benign phyllodes tumor (6).

According to previous reports (7,8), mammography and ultrasonography (US) also are not useful for differentiating benign and malignant phyllodes tumors. Moreover, only limited information on the magnetic resonance (MR) imaging characteristics of phyllodes tumors is available. Farria and colleagues (9) reported the MR features of benign phyllodes tumors of the breast, and all four tumors analyzed showed a rapid increase in signal intensity on dynamic contrast material–enhanced images. Kinoshita and co-workers (10) analyzed the MR findings of eight benign phyllodes tumors of the breast. They noted that smooth margins, lobular or polygonal shape, and occasional cystic changes or internal septa were characteristic of benign phyllodes breast tumors.

To our knowledge, however, no study had been conducted to investigate the possible correlations between the MR findings and histologic grades of phyllodes tumors of the breast. If it were possible to predict the histologic malignancy grade of phyllodes tumors on the basis of MR findings, then the decision regarding the optimal location for biopsy to achieve an adequate sample of the tumor would be greatly facilitated. Thus, the purpose of our study was to retrospectively evaluate the MR findings of phyllodes tumor of the breast and to compare these findings with the histologic grade.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Patients
In a computer search of the surgery records at two institutes (Kyushu University Hospital and National Kyushu Cancer Center), a total of 35 patients with a phyllodes tumor of the breast between September 1996 and October 2004 were identified. The data on five patients were excluded from the final analysis because they did not undergo MR imaging; thus, the data on a total of 30 patients were included in this study. All patients were female and between the ages of 16 and 73 years (mean, 40.2 years). All of the lesions were examined by using fine-needle aspiration cytology before the MR examination, and in 14 cases, core-needle biopsy also was performed. However, all core-needle biopsies were performed after the MR examination to avoid artifacts such as hemorrhage. The time from fine-needle aspiration cytology to MR examination ranged from 5 to 23 days (mean, 14.5 days). The time from MR examination to core-needle biopsy ranged from 2 to 10 days (mean, 6.8 days). The diagnosis of phyllodes tumor of the breast was histopathologically confirmed by means of examination of the specimens obtained at surgery. The study protocol was approved by the institutional review boards of Kyushu University Hospital and National Kyushu Cancer Center. All patients gave written informed consent for the review of their medical records, files, and images.

MR Imaging Technique
The MR examinations were performed within 2 weeks before surgery in all patients. A 1.5-T MR unit (Magnetom Symphony; Siemens, Erlangen, Germany) was used with commercially available breast coils at both institutions. All breast MR examinations were performed with the patient in the prone position. Initially, transverse short inversion time inversion-recovery T2-weighted images (repetition time msec/echo time msec/inversion time msec, 5970/103/150; matrix, 320 x 256; section thickness, 5 mm; intersection gap, 1 mm; field of view, 35–40 cm; one signal acquired) of both breasts in their entirety were obtained in all 30 patients. In 10 patients, after the short inversion time inversion-recovery images were acquired, diffusion-weighted MR images of the affected breast were obtained by using a multisection spin-echo single-shot echo-planar sequence in the sagittal plane. Sensitizing diffusion gradients were applied sequentially in the x, y, and z directions with b values of 0 and 1000 sec/mm². The diffusion-weighted images were obtained by using the following parameters: repetition time msec/echo time msec, 10 000/100; receiver bandwidth, 1698 Hz/pixel; field of view, 25 x 25 cm; matrix, 128 x 128; section thickness, 6 mm; intersection gap, 1 mm; six signals acquired; parallel imaging with sensitivity encoding, with use of a parallel imaging reduction factor of two; and acquisition time, 93 seconds. The diffusion-weighted images were acquired by using a two-channel phased-array body coil at the patient's back. Apparent diffusion coefficient (ADC) maps were generated from the diffusion-weighted images.

Subsequently, coronal three-dimensional dynamic T1-weighted MR images were obtained by using a fat-saturated fast low-angle shot sequence (6.4/2.5; flip angle, 25°; matrix, 256 x 256; section thickness, 1 mm; field of view, 35–40 cm; one signal acquired) in all patients. For dynamic MR imaging, gadopentetate dimeglumine (Magnevist; Nihon Schering, Osaka, Japan) was administered intravenously at a rate of 2 mL/sec (total dose, 0.1 mmol per kilogram of body weight) by using a power injector (Sonic Shot 50; Nemoto Kyorindo, Tokyo, Japan); a 20-mL saline flush followed. MR images were sequentially obtained before and 0, 60, 120, 180, 240, 300, and 360 seconds after the administration of gadopentetate dimeglumine. Subtraction images were obtained by subtracting the nonenhanced images from the contrast-enhanced peak-phase images on a pixel-by-pixel basis.

Interpretation of MR Findings
The breast MR images were retrospectively reviewed independently by two radiologists experienced in breast MR imaging (H.Y. and T.O., with 10 and 6 years experience, respectively); a consensus interpretation was reached in cases of disagreement. A consensus decision was needed only once—when we analyzed cyst wall appearance. Interpretation of the breast MR imaging findings was based on the following three characteristics according to American College of Radiology Breast Imaging Reporting and Data System MR imaging criteria (11): shape (round, oval, lobular, or irregular), margin (smooth, irregular, or spiculated), and internal enhancement (homogeneous, heterogeneous, rim enhancement, dark internal septa, enhancing internal septa, central enhancement, or no enhancement) of the tumor mass. In addition, we analyzed the findings seen on the MR images, including tumor size, tumor signal intensity higher than normal breast tissue signal intensity on T1-weighted images, tumor signal intensity lower than or equal to normal breast tissue signal intensity on T2-weighted images, cystic change with irregular wall, cystic change with smooth wall, time–signal intensity curve pattern on dynamic contrast-enhanced images, and ADC derived from the ADC map.

To analyze the T1-weighted images, we used the nonenhanced fat-saturated fast low-angle shot images. When determining whether the tumor had higher signal intensity than the normal breast tissue on the T1-weighted images or lower or equal signal intensity compared with the normal tissue on T2-weighted images, we judged the results to be positive when the area of corresponding signal intensity was visible in more than 25% of the tumor in the largest section. We analyzed the contrast-enhanced subtraction images for the presence of cystic change with irregular wall or cystic change with smooth wall. For the signal intensity measurements, one of the authors (H.Y.) placed regions of interest (ROIs) to evaluate the enhancement pattern that was demonstrative of the highest visual enhancement. Thereafter, time–signal intensity curves were constructed. Kinetic analysis was performed according to Breast Imaging Reporting and Data System MR imaging guidelines (11).

Time–signal intensity curve patterns were categorized into three types on the images obtained during the last four phases of contrast-enhanced dynamic imaging: the persistent pattern, in which the signal intensity continues to increase over time; the plateau pattern, in which the signal intensity does not change over time after its initial increase during the delayed phase; and the washout pattern, in which the signal intensity decreases after reaching the highest point of its initial increase during the delayed phase. For the ADC measurements, one of the authors (H.Y.) also manually placed ROIs based on five pixels within a given lesion. Five ROIs were placed within the area corresponding to the area in which an ROI was placed for kinetic analysis. To place the ROI for both time–signal intensity curve construction and ADC measurement in the same area within a tumor, we used the same elements: sagittal reconstructed contrast-enhanced images and sagittal ADC maps. ROIs for time–signal intensity curve construction and ADC measurement in which obvious cystic areas would be avoided also were chosen. Thereafter, average ADCs were calculated. In addition, ADCs were calculated by using the formula described in previous reports (12): ADC = [–1/(b₂–b₁)] · ln (S₂/S₁), where S₁ and S₂ are the signal intensities of the ROIs with two different gradient factors (b₁ and b₂, respectively).

Statistical Analyses
The MR findings and histologic grades were statistically analyzed to determine whether any correlations existed between the two factors. We performed Spearman correlation coefficient analysis to determine whether any significant correlation existed between histologic grade and either tumor size or ADC. To convert the three ordinal categories (benign, intermediate, and malignant) to discrete variables, we assigned a score of 1 to indicate benign, 2 to indicate intermediate, and 3 to indicate malignant when we performed Spearman correlation coefficient analysis. The difference in the proportion of positive imaging findings (other than size and ADC) among the three tumor categories (benign, intermediate, and malignant) was evaluated by using an extension of the Fisher exact test, and Bonferroni correction for multiple comparisons also was applied. We performed three comparisons: benign versus intermediate tumors, intermediate versus malignant tumors, and benign versus malignant tumors. P = .05 was considered to indicate a significant difference after Bonferroni correction. Computer software (SPSS 14.0 J for Windows; SPSS, Chicago, Ill) was used to analyze the raw data.

Histopathologic Analysis
Histopathologic diagnoses were made on the basis of findings in the specimens obtained at surgical resection of all lesions. All surgical specimens from the 30 patients were sliced perpendicular to an imaginary line from the nipple to the lesion in 5-mm intervals. Thereafter, each slice was cut parallel to an imaginary line from the nipple to the lesion in 5-mm intervals, and each histopathologic section was numbered sequentially. All specimens were stained with hematoxylin-eosin and examined by two experienced pathologists (T.E. and K.N., with 6 and 10 years experience, respectively) without knowledge of the MR findings. In addition, a radiologist (H.Y.) and a pathologist (T.E.) compared the significant MR findings with the histopathologic findings: A given MR image section was compared with the approximately corresponding region of the specimen seen on the photograph of the gross specimen. The histopathologic specimens were examined to confirm the diagnosis of phyllodes tumor of the breast.

The tumors were classified as benign (zero to four cell reproductions at 10 high-power fields, pushing margins, and minimal or moderate stromal overgrowth with minimal stromal cellularity and atypia), intermediate, or malignant (>10 cell reproductions at 10 high-power fields, infiltrative margins, moderate to marked stromal cellularity, atypia, and overgrowth) according to the criteria proposed by Azzopardi (13) and Salvadori et al (14). The MR features, including hyperintensity on T1-weighted images, low or equal signal intensity relative to surrounding normal tissue on T2-weighted images, cystic change with irregular wall, and low ADC area, were compared with the histopathologic findings.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Correlations
The mean maximal diameter (± standard deviation) was 5.3 cm ± 2.9 for the benign phyllodes tumors, 7.0 cm ± 3.0 for the intermediate tumors, and 7.6 cm ± 1.9 for the malignant tumors. There was no significant correlation between the size and the histologic grade of the phyllodes tumors (P = .174, = 0.287) (Table). There tended to be a high frequency of malignancy among both the patients who had cystic changes with an irregular wall (P = .003) and those who had tumor signal intensity lower than or equal to the normal breast tissue signal intensity on T2-weighted images (P = .005). Tumor signal intensity higher than normal breast tissue signal intensity on T1-weighted images was more frequent in the malignant (in three of five tumors) and intermediate (in three of six tumors) tumor groups than in the benign group (in two of 19 tumors); however, this finding did not correlate significantly with histologic grade (P = .024). Frequency of cystic change with smooth wall did not correlate significantly with degree of malignancy (P = .663). ADC correlated significantly with histologic grade (P = .001, = –0.888).

Histopathologic Comparisons
The 30 phyllodes tumors were histopathologically classified as benign in 19, intermediate in six, and malignant in five patients. Areas of cystic change with an irregular wall at MR imaging corresponded histopathologically to necrotic changes (Fig 1). Areas with signal intensity lower than or equal to the signal intensity of normal breast tissue on T2-weighted images and areas with a low ADC corresponded to areas of stromal hypercellularity (Figs 1, 2). Hyperintense areas on T1-weighted images corresponded to areas of hemorrhage (Fig 2).

View larger version (115K): [in this window] [in a new window] [Download PPT slide]   Figure 1a: Malignant phyllodes tumor of right breast in 17-year-old girl. (a) Coronal nonenhanced T1-weighted MR image (6.4/2.5) shows low-signal-intensity circumscribed mass (arrow). (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. Note the isointense area (arrow) in the lateral aspect of the tumor relative to the surrounding breast tissue (*). (c) Sagittal multiplanar reconstruction image constructed from contrast-enhanced subtraction image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall (arrow). An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve reveals a washout pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.41 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes (*) in lower aspect. Photomicrograph (not shown) showed frequent mitotic elements, and a diagnosis of malignant phyllodes tumor was confirmed. Comparison of MR and histopathologic findings revealed that cystic change with an irregular wall corresponded to necrosis, which was visible primarily in the lower aspect of the tumor. Isointense area in b and area with low ADC were found to correspond histopathologically to hypercellular stroma (not shown).

View larger version (86K): [in this window] [in a new window] [Download PPT slide]   Figure 1b: Malignant phyllodes tumor of right breast in 17-year-old girl. (a) Coronal nonenhanced T1-weighted MR image (6.4/2.5) shows low-signal-intensity circumscribed mass (arrow). (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. Note the isointense area (arrow) in the lateral aspect of the tumor relative to the surrounding breast tissue (*). (c) Sagittal multiplanar reconstruction image constructed from contrast-enhanced subtraction image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall (arrow). An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve reveals a washout pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.41 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes (*) in lower aspect. Photomicrograph (not shown) showed frequent mitotic elements, and a diagnosis of malignant phyllodes tumor was confirmed. Comparison of MR and histopathologic findings revealed that cystic change with an irregular wall corresponded to necrosis, which was visible primarily in the lower aspect of the tumor. Isointense area in b and area with low ADC were found to correspond histopathologically to hypercellular stroma (not shown).

View larger version (94K): [in this window] [in a new window] [Download PPT slide]   Figure 1c: Malignant phyllodes tumor of right breast in 17-year-old girl. (a) Coronal nonenhanced T1-weighted MR image (6.4/2.5) shows low-signal-intensity circumscribed mass (arrow). (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. Note the isointense area (arrow) in the lateral aspect of the tumor relative to the surrounding breast tissue (*). (c) Sagittal multiplanar reconstruction image constructed from contrast-enhanced subtraction image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall (arrow). An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve reveals a washout pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.41 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes (*) in lower aspect. Photomicrograph (not shown) showed frequent mitotic elements, and a diagnosis of malignant phyllodes tumor was confirmed. Comparison of MR and histopathologic findings revealed that cystic change with an irregular wall corresponded to necrosis, which was visible primarily in the lower aspect of the tumor. Isointense area in b and area with low ADC were found to correspond histopathologically to hypercellular stroma (not shown).

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 1d: Malignant phyllodes tumor of right breast in 17-year-old girl. (a) Coronal nonenhanced T1-weighted MR image (6.4/2.5) shows low-signal-intensity circumscribed mass (arrow). (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. Note the isointense area (arrow) in the lateral aspect of the tumor relative to the surrounding breast tissue (*). (c) Sagittal multiplanar reconstruction image constructed from contrast-enhanced subtraction image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall (arrow). An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve reveals a washout pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.41 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes (*) in lower aspect. Photomicrograph (not shown) showed frequent mitotic elements, and a diagnosis of malignant phyllodes tumor was confirmed. Comparison of MR and histopathologic findings revealed that cystic change with an irregular wall corresponded to necrosis, which was visible primarily in the lower aspect of the tumor. Isointense area in b and area with low ADC were found to correspond histopathologically to hypercellular stroma (not shown).

View larger version (51K): [in this window] [in a new window] [Download PPT slide]   Figure 1e: Malignant phyllodes tumor of right breast in 17-year-old girl. (a) Coronal nonenhanced T1-weighted MR image (6.4/2.5) shows low-signal-intensity circumscribed mass (arrow). (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. Note the isointense area (arrow) in the lateral aspect of the tumor relative to the surrounding breast tissue (*). (c) Sagittal multiplanar reconstruction image constructed from contrast-enhanced subtraction image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall (arrow). An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve reveals a washout pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.41 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes (*) in lower aspect. Photomicrograph (not shown) showed frequent mitotic elements, and a diagnosis of malignant phyllodes tumor was confirmed. Comparison of MR and histopathologic findings revealed that cystic change with an irregular wall corresponded to necrosis, which was visible primarily in the lower aspect of the tumor. Isointense area in b and area with low ADC were found to correspond histopathologically to hypercellular stroma (not shown).

View larger version (151K): [in this window] [in a new window] [Download PPT slide]   Figure 1f: Malignant phyllodes tumor of right breast in 17-year-old girl. (a) Coronal nonenhanced T1-weighted MR image (6.4/2.5) shows low-signal-intensity circumscribed mass (arrow). (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. Note the isointense area (arrow) in the lateral aspect of the tumor relative to the surrounding breast tissue (*). (c) Sagittal multiplanar reconstruction image constructed from contrast-enhanced subtraction image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall (arrow). An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve reveals a washout pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.41 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes (*) in lower aspect. Photomicrograph (not shown) showed frequent mitotic elements, and a diagnosis of malignant phyllodes tumor was confirmed. Comparison of MR and histopathologic findings revealed that cystic change with an irregular wall corresponded to necrosis, which was visible primarily in the lower aspect of the tumor. Isointense area in b and area with low ADC were found to correspond histopathologically to hypercellular stroma (not shown).

View larger version (114K): [in this window] [in a new window] [Download PPT slide]   Figure 2a: Intermediate phyllodes tumor of right breast in 39-year-old woman. (a) Coronal nonenhanced T1-weighted image (6.4/2.5) shows mass (arrows) with both high and low signal intensity relative to surrounding breast tissue. (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. (c) Sagittal multiplanar reconstruction image constructed from subtracted contrast-enhanced image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall. An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve shows plateau pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.45 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes in lateral area (*). MR imaging–histopathologic correlation analysis revealed that hyperintense area in a corresponds to hemorrhagic infarction. Photomicrograph (not shown) showed a few mitotic elements, and a diagnosis of intermediate phyllodes tumor was confirmed. Area with low ADC was found to correspond to hypercellular stroma (not shown).

View larger version (144K): [in this window] [in a new window] [Download PPT slide]   Figure 2b: Intermediate phyllodes tumor of right breast in 39-year-old woman. (a) Coronal nonenhanced T1-weighted image (6.4/2.5) shows mass (arrows) with both high and low signal intensity relative to surrounding breast tissue. (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. (c) Sagittal multiplanar reconstruction image constructed from subtracted contrast-enhanced image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall. An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve shows plateau pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.45 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes in lateral area (*). MR imaging–histopathologic correlation analysis revealed that hyperintense area in a corresponds to hemorrhagic infarction. Photomicrograph (not shown) showed a few mitotic elements, and a diagnosis of intermediate phyllodes tumor was confirmed. Area with low ADC was found to correspond to hypercellular stroma (not shown).

View larger version (122K): [in this window] [in a new window] [Download PPT slide]   Figure 2c: Intermediate phyllodes tumor of right breast in 39-year-old woman. (a) Coronal nonenhanced T1-weighted image (6.4/2.5) shows mass (arrows) with both high and low signal intensity relative to surrounding breast tissue. (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. (c) Sagittal multiplanar reconstruction image constructed from subtracted contrast-enhanced image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall. An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve shows plateau pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.45 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes in lateral area (*). MR imaging–histopathologic correlation analysis revealed that hyperintense area in a corresponds to hemorrhagic infarction. Photomicrograph (not shown) showed a few mitotic elements, and a diagnosis of intermediate phyllodes tumor was confirmed. Area with low ADC was found to correspond to hypercellular stroma (not shown).

View larger version (11K): [in this window] [in a new window] [Download PPT slide]   Figure 2d: Intermediate phyllodes tumor of right breast in 39-year-old woman. (a) Coronal nonenhanced T1-weighted image (6.4/2.5) shows mass (arrows) with both high and low signal intensity relative to surrounding breast tissue. (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. (c) Sagittal multiplanar reconstruction image constructed from subtracted contrast-enhanced image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall. An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve shows plateau pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.45 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes in lateral area (*). MR imaging–histopathologic correlation analysis revealed that hyperintense area in a corresponds to hemorrhagic infarction. Photomicrograph (not shown) showed a few mitotic elements, and a diagnosis of intermediate phyllodes tumor was confirmed. Area with low ADC was found to correspond to hypercellular stroma (not shown).

View larger version (46K): [in this window] [in a new window] [Download PPT slide]   Figure 2e: Intermediate phyllodes tumor of right breast in 39-year-old woman. (a) Coronal nonenhanced T1-weighted image (6.4/2.5) shows mass (arrows) with both high and low signal intensity relative to surrounding breast tissue. (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. (c) Sagittal multiplanar reconstruction image constructed from subtracted contrast-enhanced image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall. An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve shows plateau pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.45 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes in lateral area (*). MR imaging–histopathologic correlation analysis revealed that hyperintense area in a corresponds to hemorrhagic infarction. Photomicrograph (not shown) showed a few mitotic elements, and a diagnosis of intermediate phyllodes tumor was confirmed. Area with low ADC was found to correspond to hypercellular stroma (not shown).

View larger version (144K): [in this window] [in a new window] [Download PPT slide]   Figure 2f: Intermediate phyllodes tumor of right breast in 39-year-old woman. (a) Coronal nonenhanced T1-weighted image (6.4/2.5) shows mass (arrows) with both high and low signal intensity relative to surrounding breast tissue. (b) Transverse short inversion time inversion-recovery T2-weighted image (5970/103/150) shows circumscribed mass of heterogeneous signal intensity. (c) Sagittal multiplanar reconstruction image constructed from subtracted contrast-enhanced image shows heterogeneously enhancing mass. Note the cystic change with an irregular wall. An ROI (circle) was placed in superior part of the tumor. (d) Time–signal intensity curve shows plateau pattern at contrast-enhanced dynamic imaging. (e) ADC map constructed from diffusion-weighted images with b factors of 0 and 1000 sec/mm2 and obtained at same level as c shows low-signal-intensity mass. The ADC was 1.45 x 10–3 mm2/sec. An ROI (circle) was placed in superior part of the tumor. (f) Cut surface of gross tumor specimen reveals hemorrhagic changes in lateral area (*). MR imaging–histopathologic correlation analysis revealed that hyperintense area in a corresponds to hemorrhagic infarction. Photomicrograph (not shown) showed a few mitotic elements, and a diagnosis of intermediate phyllodes tumor was confirmed. Area with low ADC was found to correspond to hypercellular stroma (not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

Liberman and co-workers (7) compared the mammographic and US findings of phyllodes breast tumors with the histopathologic features and concluded that a tumor diameter of 3 cm or greater suggests a higher likelihood of malignancy; however, the imaging characteristics of benign and malignant tumors overlapped substantially. In that study, the US findings of phyllodes tumors, including internal echo texture, posterior acoustic properties, and cystic change, also were considered. However, their analysis did not include a description of the cyst walls as either smooth or irregular.

Histopathologically, phyllodes tumors contain leaflike, epithelium-lined papillary projections that penetrate into cystic spaces (2). Consequently, US frequently reveals cystic areas in these tumors (8). In our study, cystic areas with irregular walls were not observed in the benign phyllodes tumors. Therefore, the classification of a cyst wall as either smooth or irregular appears to be important for differentiating benign from intermediate or malignant phyllodes tumors. In addition, because phyllodes tumors tend to appear as large masses, MR imaging appears to be more suitable than US for observing the deep portion of these tumors.

Farria and co-workers (9) described the MR features of four benign phyllodes tumors of the breast in four patients. All four tumors had circumscribed margins, high signal intensity on T2-weighted images, and rapid enhancement at dynamic imaging. Kinoshita and colleagues (10) described the MR features of benign phyllodes tumors of the breast in eight patients. These tumors appeared as lobular, polygonal, or oval masses. The signal intensity of the tumors on T2-weighted images was high or both hyperintense and isointense, whereas low signal intensity was observed on the T1-weighted images. Five of the eight tumors showed rapid enhancement, and the remaining three tumors demonstrated gradual enhancement at dynamic contrast-enhanced imaging. However, to date, the MR features of malignant phyllodes tumors have not been well characterized. Moreover, to our knowledge, no information regarding comparisons of the MR characteristics with the histologic grades of phyllodes breast tumors had been available.

According to our comparison of MR and histopathologic findings, low ADC corresponded to stromal hypercellularity in the intermediate and malignant phyllodes tumors. High signal intensity on T1-weighted images corresponded histopathologically to hemorrhage, and cystic change with irregular wall corresponded to necrotic changes. Histopathologically, stromal cellularity, pleomorphism, nuclear atypia, and growth of stroma occasionally are seen in phyllodes tumors as the histologic grade of malignancy increases (13,14). We speculated that this gradation of stromal cellularity is reflected in the MR findings—for example, in the signal intensity on T2-weighted images and in the ADC on diffusion-weighted images. The correlation of cell density with ADC has been established in studies of brain and breast tumors (12,15). Although we did not compare the cell density of phyllodes tumors with the ADCs, a low ADC in a malignant phyllodes tumor was considered to be attributable to stromal hypercellularity.

Our study revealed no significant correlation between time–signal intensity curve pattern and histologic grade. Twelve of 19 benign phyllodes tumors exhibited an initially rapid enhancement pattern. Initially rapid enhancement is generally regarded as a malignant pattern at breast MR imaging; however, this does not apply to phyllodes tumors. Our results are consistent with those described in previous reports by Farria et al (9) and Kinoshita et al (10). It is possible that an initially rapid enhancement pattern is reflective of the epithelial component of benign phyllodes tumors.

Our study revealed also that mass shape, margin, and internal enhancement pattern did not correlate significantly with histologic grade of phyllodes breast tumor. Morphologic MR characteristics such as tumor margin, internal enhancement, and dynamic enhancement pattern have been identified as findings useful for differentiating benign from malignant breast tissue (16,17). Rim enhancement is an obvious characteristic pattern of malignancy (18). All of the phyllodes tumors in our study had a smooth margin, and internal enhancement pattern did not correlate significantly with histologic grade. Therefore, we believe that signal intensity on T1- and T2-weighted MR images, cyst wall appearance, and ADC—in addition to classic findings such as spiculated margin, rim enhancement, and dynamic enhancement pattern—can be useful in diagnosing the histologic grade of phyllodes tumors.

A major diagnostic problem associated with phyllodes tumors is that neither fine-needle aspiration cytology, nor mammography, nor US can be used to distinguish fibroadenomas from phyllodes tumors (7,8). Fine-needle aspiration cytology may lead to the misdiagnosis of a benign phyllodes tumor as a fibroadenoma. According to our present study results, either a hypointense area on T2-weighted images or a low ADC area on diffusion-weighted images is suggestive of an interstitial hypercellular lesion. Therefore, MR imaging can be used to guide the selection of an appropriate biopsy site.

Wide local resection with margins of at least 1–2 cm is generally recommended for the treatment of both benign and malignant phyllodes tumors (1,3). However, Buchanan (4) proposed performing simple mastectomy for treatment of tumors larger than 5 cm in diameter and those found to be malignant or borderline malignant. Therefore, any MR characteristics suggestive of malignancy may influence the surgical treatment for phyllodes tumor of the breast.

Our study had some limitations: First, diffusion-weighted MR imaging was performed in only 10 patients; therefore, the correlation between ADC and histologic grade may be weak. Second, the comparison of MR and histopathologic findings was not based on a detailed one-to-one comparison. All surgical specimens were sliced perpendicular to an imaginary line from the nipple to the lesion, whereas the MR imaging planes analyzed were sagittal or coronal. However, to our knowledge, this is the first report to demonstrate a correlation between histologic grade and MR findings of phyllodes breast tumor. In conclusion, several MR findings can be used to help determine the histologic grade of phyllodes tumor of the breast.

	ADVANCES IN KNOWLEDGE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 

• Cystic changes with irregular wall, tumor signal intensity lower than or equal to normal breast tissue signal intensity on T2-weighted images, and low ADC correlated significantly with histologic grade of phyllodes tumor.
  
• High tumor signal intensity compared with the normal breast tissue signal intensity on T1-weighted images was suggestive of malignancy but did not correlate significantly (P = .024) with histologic grade.
  
• Tumor size, cystic changes with smooth wall, and time–signal intensity curve pattern did not correlate significantly with histologic grade of phyllodes tumor.
  

	ACKNOWLEDGMENTS

 
We thank Tsukasa Sakemoto, RT, for his technical support and are grateful to Satoshi Toyoshima, MD, PhD, for advice regarding the histopathologic analyses.

	FOOTNOTES

 

Abbreviations: ADC = apparent diffusion coefficient • ROI = region of interest

Author contributions: Guarantor of integrity of entire study, H.Y.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, H.Y.; clinical studies, all authors; statistical analysis, H.Y.; and manuscript editing, H.Y., H.H.

Authors stated no financial relationship to disclose.

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 

1. Rowell MD, Perry RR, Hsju JG, et al. Phyllodes tumors. Am J Surg 1993;165:376–379.[CrossRef][Medline]
2. Rosen PP, Oberman HA. Fibroepithelial neoplasms. In: Rosai J, Sobin LH, eds. Atlas of tumor pathology: tumors of the mammary gland. Washington, DC: Armed Forces Institute of Pathology, 1993; 107–114.
3. Mangi AA, Smith BL, Gadd MA, et al. Surgical management of phyllodes tumors. Arch Surg 1999;134:487–492.[Abstract/Free Full Text]
4. Buchanan EB. Cystosarcoma phyllodes and its surgical management. Am Surg 1995;61:350–355.[Medline]
5. Petrek JA. Phyllodes tumors. In: Harris JR, Lippman S, Morrow M, et al, eds. Disease of the breast. 2nd ed. Philadelphia, Pa: Lippincott Williams & Wilkins, 2000; 669–674.
6. Krishnamurthy S, Ashfaq R, Shin HJ, et al. Distinction of phyllodes tumor from fibroadenoma: a reappraisal of an old problem. Cancer 2000;90(6):342–349.[CrossRef][Medline]
7. Liberman L, Bonaccio E, Hamele-Bena D, et al. Benign and malignant phyllodes tumors: mammographic and sonographic findings. Radiology 1996;198:121–124.[Abstract/Free Full Text]
8. Chao TC, Lo YF, Chen SC, et al. Sonographic features of phyllodes tumors of the breast. Ultrasound Obstet Gynecol 2002;20:64–71.[CrossRef][Medline]
9. Farria DM, Gorczyca DP, Barsky SH, et al. Benign phyllodes tumor of the breast: MR imaging features. AJR Am J Roentgenol 1996;167:187–189.[Free Full Text]
10. Kinoshita T, Fukutomi T, Kubochi K. Magnetic resonance imaging of benign phyllodes tumors of the breast. Breast J 2004;10:232–236.[CrossRef][Medline]
11. American College of Radiology. ACR BI-RADS magnetic resonance imaging. Reston, Va: American College of Radiology, 2003.
12. Guo Y, Cai YQ, Cai ZL, et al. Differentiation of clinically benign and malignant breast lesions using diffusion-weighted imaging. J Magn Reson Imaging 2002;16:172–178.[CrossRef][Medline]
13. Azzopardi JG. Sarcoma in the breast. In: Benningron J, ed. Problems in breast pathology. Vol 2, Major problems in pathology. Philadelphia, Pa: Saunders, 1979; 355–359.
14. Salvadori B, Cusumano F, Del Bo R, et al. Surgical treatment of phyllodes tumors of the breast. Cancer 1989;63:2532–2536.[CrossRef][Medline]
15. Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cerebral gliomas. J Magn Reson Imaging 1999;9:53–60.[CrossRef][Medline]
16. Wedegartner U, Bick U, Wortler K, et al. Differentiation between benign and malignant findings on MR-mammography: usefulness of morphological criteria. Eur Radiol 2001;11:1645–1650.[CrossRef][Medline]
17. Kuhl CK, Mielcareck P, Klaschik S, et al. Dynamic breast MR imaging: are signal intensity time course data useful for differential diagnosis of enhancing lesions? Radiology 1999;211:101–110.
18. Matsubayashi R, Matsuo Y, Edakuni G, et al. 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;217:841–848.[Abstract/Free Full Text]

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2413051470v1 241/3/702    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Yabuuchi, H. Articles by Honda, H. Search for Related Content PubMed PubMed Citation Articles by Yabuuchi, H. Articles by Honda, H.