| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Hôpital Beaujon and Laboratoire d’Imagerie Médicale Paris-Nord, Faculté de Médecine Paris VII, Clichy, France (B.T., V.V., E.D., J.L.D., Y.M.); and Department of Radiology, University of California San Francisco, 505 Parnassus Ave, Rm M-372, Box 0628, San Francisco, CA 94143 (B.T., B.F.). From the 2001 RSNA scientific assembly. Received November 27, 2001; revision requested February 7, 2002; revision received March 15; accepted May 7. Supported by a grant from the French Radiology Society. Address correspondence to B.T. (e-mail: bachir.taouli@radiology.ucsf.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
MATERIALS AND METHODS: Sixty-six patients were examined with two single-shot echo-planar diffusion-weighted MR sequences. In the first sequence, liver diffusion isotropy was evaluated by using diffusion gradients in three directions with two b values. In the second sequence, a unidirectional diffusion gradient was used with four b values. ADCs were measured in 43 patients with 52 focal hepatic lesions more than 1 cm in diameter and in 23 patients with 14 normal and nine cirrhotic livers and were compared by using nonparametric tests.
RESULTS: Diffusion in the liver parenchyma was isotropic. ADCs of focal hepatic lesions were significantly different between sequences (P < .01). The mean (± SD) ADCs in the first sequence were 0.94 x 10-3 mm2/sec ± 0.60 for metastases, 1.33 x 10-3 mm2/sec ± 0.13 for HCCs, 1.75 x 10-3 mm2/sec ± 0.46 for benign hepatocellular lesions, 2.95 x 10-3 mm2/sec ± 0.67 for hemangiomas, and 3.63 x 10-3 mm2/sec ± 0.56 for cysts. There was a significant difference between benign (2.45 x 10-3 mm2/sec ± 0.96, isotropic value) and malignant (1.08 x 10-3 mm2/sec ± 0.50) lesions (P < .01 for both sequences).
CONCLUSION: Diffusion-weighted MR imaging can help differentiate benign from malignant hepatic lesions. The use of two b values in one direction could be sufficient for the design of MR sequences in the liver.
© RSNA, 2002
Index terms: Liver, diseases, 761.30, 761.79 • Liver, focal nodular hyperplasia, 761.3198 • Liver neoplasms, 761.30 • Liver neoplasms, MR, 761.121416, 761.12144 • Magnetic resonance (MR), diffusion study, 761.12144 • Magnetic resonance (MR), echo planar, 761.121416
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Investigators in a few preliminary studies measured the ADCs of abdominal organs and focal hepatic lesions by using a single-shot echo-planar MR imaging sequence (13–16). Results of these studies showed principally that diffusion-weighted MR imaging, by means of ADC measurement, can be used to characterize focal hepatic lesions. Benign lesions, such as hepatic cysts and hemangiomas, showed higher ADCs than those of malignant lesions (hepatocellular carcinomas [HCCs] and metastases).
Investigators in most previous studies measured the ADC in only one direction and implicitly assumed that, unlike in the brain (4,8) and kidney (17), the diffusion characteristics of the liver are isotropic. In addition, the ADCs of benign hepatocellular lesions, including cases of focal nodular hyperplasia and adenoma, have not yet been reported, to our knowledge.
The purpose of this study was to (a) evaluate liver diffusion isotropy, (b) compare two single-shot echo-planar diffusion-weighted MR imaging sequences for the characterization of focal hepatic lesions by using two or four b values, and (c) determine a threshold ADC value to differentiate benign lesions (eg, cysts, hemangiomas, cases of focal nodular hyperplasia, and adenomas) from malignant lesions (eg, HCCs and metastases).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Metastases.—Fifteen metastatic lesions (mean diameter, 3.4 cm; range, 1.0–8.7 cm) were evaluated in nine patients. The primary tumors were colorectal carcinoma (n = 6), carcinoid tumor (n = 8) of the gastrointestinal tract or pancreas, and renal carcinoma (n = 1). The diagnosis of metastasis was proven by means of surgery, needle biopsy, and follow-up imaging examinations, including ultrasonography (US), computed tomography (CT), and MR imaging, which showed progression of the lesions in two, three, and four patients, respectively. All the metastases were solid, without any cystic component.
HCCs.—Nine HCCs (mean diameter, 4.3 cm; range, 1.5–8.0 cm) were evaluated in nine patients. The diagnosis of HCC was assigned on the basis of MR imaging, CT, or US findings and confirmed by means of histologic findings or the elevation of serum 
-fetoprotein levels in four and five patients, respectively. All patients with HCC had concomitant cirrhosis related to chronic viral hepatitis B or C or to hemochromatosis in five, three, and one patient, respectively. The diagnosis of cirrhosis was assigned on the basis of clinical findings in five patients and on the basis of histologic findings in four patients.
Benign hepatocellular lesions.—Fifteen benign hepatocellular lesions (mean diameter, 6.8 cm; range, 2.0–15.5 cm), including 12 cases of focal nodular hyperplasia and three hepatic adenomas, were evaluated in 15 patients. The diagnosis of focal nodular hyperplasia (mean diameter, 6.0 cm; range, 2.0–11.4 cm) was established in the presence of typical MR imaging findings (18) in 12 patients without history of cancer and with normal hepatic function test results. On T2-weighted images, the cases of focal nodular hyperplasia were isointense (n = 8) or slightly hyperintense (n = 4) when compared with the surrounding normal hepatic tissue. The diagnosis of hepatocellular adenoma (mean diameter, 9.4 cm; range, 4.1–15.5 cm) was assigned at MR imaging (19) in three patients with heterogeneous lesions with some areas of hyperintensity on T2-weighted images; all diagnoses were confirmed by means of surgical resection.
Hemangiomas.—Seven hemangiomas (mean diameter, 3.8 cm; range, 2.1–6.4 cm) were evaluated in seven patients. The diagnosis of hemangioma was established by means of hyperintensity on T2-weighted images and the typical enhancement pattern seen at CT or MR imaging (slightly irregular or globular peripheral enhancement after injection of a bolus of contrast medium, with gradual filling of the center of the lesion on delayed images) (20,21).
Cysts.—Six cysts (mean diameter, 5.5 cm; range, 2.4–8.0 cm) were evaluated in three patients. One patient had adult polycystic renal disease with multiple hepatic cysts. The diagnosis of hepatic cysts was established by means of typical MR imaging findings (hypointense on T1-weighted images and hyperintense on T2-weighted images, with no enhancement after contrast material injection) and US appearances.
MR Imaging
Patients were examined with a 1.5-T superconducting MR system (Gyroscan; Philips Medical Systems, Best, the Netherlands) with a 23 mT/m maximum gradient capability and a surface phased-array coil. All patients underwent diffusion-weighted MR imaging in addition to imaging with a routine hepatic MR protocol to identify and select hepatic lesions suitable for ADC measurement. The protocol included a T1-weighted dual fast gradient-recalled-echo sequence (in-phase and out-of-phase sequences) (repetition time msec/echo time msec, 126/4.6 [in-phase], 2.3 [out-of-phase]; flip angle, 80°; matrix, 179 x 256; section thickness, 8 mm; intersection gap, 2.5 mm; one signal acquired; field of view, 320 mm), a T2-weighted fast spin-echo sequence with spectral fat saturation (1,800/85; fast spin-echo factor, 16; matrix, 512 x 512; section thickness, 8 mm; intersection gap, 2.5 mm; two signals acquired; field of view, 320 mm), and a T1-weighted gradient-echo out-of-phase sequence after dynamic injection of 0.1 mmol per kilogram of body weight of gadoterate meglumine (Dotarem; Guerbet, Aulnay-sous-Bois, France) through a power injector at a rate of 2 mL/sec.
Diffusion-weighted MR Imaging
Before contrast material injection, two breath-hold diffusion-weighted MR sequences were performed with the single-shot echo-planar imaging technique. These sequences combined diffusion gradient pulses before and after the 180° pulse. Spectral fat saturation was used systematically to exclude chemical shift artifacts.
First sequence.—With this sequence, liver isotropy was evaluated by using diffusion gradients with two b values (0 and 500 sec/mm2) along three directions: the frequency-encoding (x), phase-encoding (y), and section-select (z) directions. Five images were obtained: one image for b = 0, one image for each direction for b = 500 sec/mm2, and one isotropic image. The following parameters were used to acquire 20 sections (two sets of 10 sections each) in a 24-second breath hold: 2,400/104; gradient strength, 20 mT/m; diffusion gradient duration, 26 msec; matrix size, 82 x 128; section thickness, 6 mm; intersection gap, 2 mm; one signal acquired; field of view, 320 mm.
Second sequence.—In this sequence, a unidirectional diffusion gradient was applied along the section-select direction (z axis) with four increasing b values: 0, 134, 267, and 400 sec/mm2. Four images were obtained (one for each b value). The following parameters were used to acquire 15 sections in a 12-second breath hold: 3,106/104; gradient strength, 20.4 mT/m; diffusion gradient duration, 23.5 msec; matrix size, 128 x 256; section thickness, 8 mm; intersection gap, 3 mm; one signal acquired; field of view, 350 mm.
Phantom studies.—To validate our system, a preliminary phantom study was performed. Plastic tubes filled with water and acetone were imaged with increasing b values: 200, 300, 400, 500, 600, and 1,000 sec/mm2. The ADCs of water and acetone were calculated at ambient room temperature (21°C).
Image Analysis
All MR images were analyzed retrospectively in consensus by two experienced radiologists (V.V., B.T.) who were aware of the results of CT and US. The focal hepatic lesions were identified on the T1- and T2-weighted images, and their signal intensities, sizes, and patterns of enhancement after contrast material injection were noted. Because of the limited resolution of the diffusion-weighted sequences, only lesions larger than 1 cm in diameter were evaluated. The ADCs were measured in 52 lesions (>1 cm) in 43 patients. Quantitative ADC maps were derived automatically on a voxel-by-voxel basis by using commercially available software (Intera Workstation, release 8.1.3; Philips Medical Systems). The ADC was calculated with a linear regression analysis of the function S = S0 · exp(-b x ADC), where b is the diffusion factor, S is the signal intensity after application of the diffusion gradient, and S0 is the signal intensity at b = 0 sec/mm2.
One of the two radiologists (B.T.) established regions of interest in each lesion on the mapping images, and ADC values were obtained by using a workstation (EZ Vision Workstation, release 4; Philips Medical Systems). All regions of interest (round shape, at least 10 mm in diameter) were placed within the confines of the lesions; for heterogeneous lesions, regions of interest included the entire lesion. To ensure that the same areas were measured, the regions of interest were copied and pasted onto the T1-, T2-, and diffusion-weighted MR images and ADC maps. In normal and cirrhotic livers, regions of interest were always placed in the posterior segment of the right hepatic lobe to avoid artifacts from the great vessels. ADCs were each measured twice, and the measurements were averaged.
Statistical Evaluation
To evaluate liver isotropy, ADCs obtained in three directions (frequency encoding, phase encoding, and section select) were compared. To determine the ADCs of focal hepatic lesions and to try to determine whether the ADCs of focal hepatic lesions help differentiate benign from malignant lesions, the ADCs measured in every lesion on MR diffusion-weighted images were compared between the different groups of lesions and between normal and cirrhotic livers. ADCs of normal and cirrhotic livers were also compared for both sequences. In addition, ADCs were compared between both imaging sequences. The Kruskall-Wallis test was used for overall comparison, and the Wilcoxon signed rank test was used for comparison between two groups. The differences were considered significant when P values were less than .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Patients
ADCs were measured in all focal hepatic lesions included in the study and in 14 normal and nine cirrhotic livers.
Liver isotropy.—The normal and cirrhotic liver parenchyma and all explored focal hepatic lesions were isotropic, with small nonsignificant differences between the ADCs measured in the three directions (Table 1).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Comparison of ADCs of normal and cirrhotic livers.—The ADCs of normal liver ranged from 1.40 to 2.55 x 10-3 mm2/sec (isotropic values) with the first sequence and from 1.12 to 2.71 x 10-3 mm2/sec with the second sequence, whereas the ADCs of cirrhotic liver ranged from 0.21 to 1.81 x 10-3 mm2/sec with the first sequence and from 0.17 to 1.48 x 10-3 mm2/sec with the second sequence. The mean ADCs of cirrhotic liver (Table 2) were significantly lower than those of normal liver (P < .05 with both sequences). Despite the statistical difference, however, there was a considerable overlap.
Comparison of the two imaging sequences.—The ADCs of all hepatic lesions (except for HCCs) and normal and cirrhotic livers obtained with the second sequence were slightly lower than those obtained with the first sequence (Tables 2, 3), without reaching statistical significance.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
Characterization of Focal Hepatic Lesions and Differences between the Two Imaging Sequences
Differentiation of benign from malignant hepatic lesions is a frequent diagnostic problem at MR imaging. In our study, we obtained similar results by using two or four b values in terms of the characterization of focal hepatic lesions and the differentiation of benign from malignant lesions. The malignant lesions, including metastases and HCCs, had the lowest ADCs, whereas the benign lesions, including hemangiomas and cysts, had the highest ADCs. Benign hepatocellular lesions had intermediate ADCs. To our knowledge, we were the first investigators to explore the ADCs of benign hepatocellular lesions. Previous studies were focused on the differentiation between benign and malignant focal hepatic lesions by measuring the ADCs (11,13–15,24,25), and all of them showed lower ADCs in malignant lesions than those in benign lesions, with variable overlap. Except for one study (24) in which turboFLASH (Siemens, Erlangen, Germany) MR imaging was used, all previous studies (11,13–15) were designed with an echo-planar MR sequence, mostly with single-shot echo-planar MR imaging (13–15). However, the results of these studies showed considerable discrepancies in ADCs for abdominal organs and focal hepatic lesions (Table 4), probably related to the different b values, which ranged from very low to very high values (30–1,200 sec/mm2).
|
We found equivalent results for the characterization of focal hepatic lesions and for the differentiation of benign and malignant lesions by comparing the results with the two diffusion-weighted sequences implemented in our study. However, the ADCs were slightly higher when using b values of 0–500 sec/mm2 than those obtained with b values of 0–400 sec/mm2, without reaching statistical significance. The ADCs obtained with the second sequence were probably more precise than those obtained with the first sequence, because the number of data used for the ADC calculation was increased (four b values instead of two). In the study of Kim et al (15), the ADCs obtained with a higher b value (<850 sec/mm2) were lower than those obtained with lower b values (<100 and <410 sec/mm2). In our study, the small difference in the magnitude of the b values (100 sec/mm2) probably did not influence ADC calculation.
In our study, we decided to explore the ADCs of benign hepatocellular lesions, including cases of focal nodular hyperplasia and adenomas, despite their relatively short T2 values, because the lesions included were large (mean size of 6.8 mm) and easily depicted on the diffusion-weighted images. Unsurprisingly, the benign hepatocellular lesions had intermediate ADCs, ranging mostly between 1 and 2 x 10-3 mm2/sec (except for two cases of focal nodular hyperplasia with ADCs greater than 2 x 10-3 mm2/sec), which were not different from those of normal liver parenchyma. Despite some overlap, however, benign hepatocellular lesions had significantly different ADCs when compared with those of other hepatic lesions, particularly HCCs.
As in previous studies, we found that hepatic cysts had the highest ADCs because of their fluid content, with nonrestricted motion of water molecules. The ADCs of hemangiomas were also high (all but one had ADCs greater than 2 x 10-3 mm2/sec) but lower than those of cysts, without reaching statistical significance. This is probably related to the vascular content of hemangiomas, which is more viscous than cystic fluid (13). Moreover, we found that metastatic lesions and HCCs had the lowest ADCs, probably due to their high tumoral content, which restricts the water diffusion. All the metastatic lesions included in our study were solid and did not manifest cystic components that can increase the ADC (13). Hepatic metastases of carcinoid tumors can mimic hemangiomas on T1- and T2-weighted images, and the differentiation between these two lesions usually depends on dynamic gadolinium-enhanced imaging findings (21,26,27). Results of studies (28–30) have shown that T2 measurements can be useful to differentiate hepatic malignancies from hemangiomas and cysts and have questioned the routine use of gadolinium-based contrast material. Diffusion-weighted MR imaging, by means of ADC measurement, is an additional tool that can be used accurately for that purpose.
Finally, as in most previous studies (11,13–15,31), the results of our study showed lower ADCs of cirrhotic liver than those of normal liver, despite a considerable overlap. A possible explanation is the restricted water diffusion in fibrotic liver (11,31).
The present study had several potential limitations, mostly related to the imaging sequences. First, single-shot echo-planar imaging has a low spatial resolution and low signal-to-noise ratio, and therefore, infracentimetric lesions were not evaluated. Findings in recent studies have shown that new fast imaging sequences (32) can improve image quality and decrease echo-planar imaging–related artifacts or that the use of high-field-strength 3.0-T imagers (33) can potentially improve the signal in diffusion-weighted MR imaging. Second, the patient population was relatively small, but an approximately equivalent number of benign (n = 28) and malignant lesions (n = 24) were evaluated. However, further studies with a larger number of patients are needed in the future. Third, we did not use high b values because of the diminished image quality, and overestimation of the ADCs by including the perfusion fraction was possible; however, we did not intend—as in the study of Yamada et al (25)—to measure the true diffusion coefficient. Fourth, despite the significant differences between benign and malignant lesions, there was an overlap of their ADCs between 1 and 2 x 10-3 mm2/sec, particularly between benign hepatocellular lesions and HCCs.
In summary, we provide evidence that suggests that liver diffusion is isotropic and that diffusion-weighted MR imaging can give additional information for the characterization of hepatic lesions and can potentially be useful for the differentiation between benign and malignant hepatic lesions. A combined use of ADC and T2 measurements may reduce the need for the systematic use of gadolinium-based contrast material at MR imaging. In light of these results, for future clinical applications of diffusion-weighted MR imaging of the liver, we recommend the use of a monodirectional diffusion gradient with a minimum of two b values.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, V.V., B.T., E.D., Y.M.; study concepts and design, B.T., V.V., E.D., J.L.D.; literature research, B.T., J.L.D.; clinical studies, B.T.; experimental studies, B.T., E.D.; data acquisition, B.T.; data analysis/interpretation, B.T., V.V., B.F.; statistical analysis, B.T., E.D., B.F.; manuscript preparation, B.T.; manuscript definition of intellectual content, V.V., E.D., B.T., Y.M.; manuscript editing, B.T.; manuscript revision/review, B.T., V.V., E.D.; manuscript final version approval, B.T.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSION REFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Cui, X.-P. Zhang, Y.-S. Sun, L. Tang, and L. Shen Apparent Diffusion Coefficient: Potential Imaging Biomarker for Prediction and Early Detection of Response to Chemotherapy in Hepatic Metastases Radiology, September 1, 2008; 248(3): 894 - 900. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. I. Sainani, O. A. Catalano, N.-S. Holalkere, A. X. Zhu, P. F. Hahn, and D. V. Sahani Cholangiocarcinoma: Current and Novel Imaging Techniques RadioGraphics, September 1, 2008; 28(5): 1263 - 1287. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. C. van den Bos, S. M. Hussain, G. P. Krestin, and P. A. Wielopolski Liver Imaging at 3.0 T: Diffusion-induced Black-Blood Echo-planar Imaging with Large Anatomic Volumetric Coverage as an Alternative for Specific Absorption Rate-intensive Echo-Train Spin-Echo Sequences: Feasibility Study Radiology, July 1, 2008; 248(1): 264 - 271. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Zhang, Y. Mazaheri Tehrani, L. Wang, N. M. Ishill, L. H. Schwartz, and H. Hricak Renal Masses: Characterization with Diffusion-weighted MR Imaging--A Preliminary Experience Radiology, May 1, 2008; 247(2): 458 - 464. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. F. Hanna, D. A. Aguirre, N. Kased, S. C. Emery, M. R. Peterson, and C. B. Sirlin Cirrhosis-associated Hepatocellular Nodules: Correlation of Histopathologic and MR Imaging Features RadioGraphics, May 1, 2008; 28(3): 747 - 769. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. G. Lall, A. M. Aisen, N. Bansal, and K. Sandrasegaran Nonalcoholic Fatty Liver Disease Am. J. Roentgenol., April 1, 2008; 190(4): 993 - 1002. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Parikh, S. J. Drew, V. S. Lee, S. Wong, E. M. Hecht, J. S. Babb, and B. Taouli Focal Liver Lesion Detection and Characterization with Diffusion-weighted MR Imaging: Comparison with Standard Breath-hold T2-weighted Imaging Radiology, March 1, 2008; 246(3): 812 - 822. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-H. Shen, Y.-Y. Chiou, J.-H. Wang, M.-S. Yen, R.-C. Lee, C.-R. Lai, and C.-Y. Chang Diffusion-Weighted Single-Shot Echo-Planar Imaging with Parallel Technique in Assessment of Endometrial Cancer Am. J. Roentgenol., February 1, 2008; 190(2): 481 - 488. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Inan, A. Arslan, G. Akansel, Y. Anik, H. T. Sarisoy, E. Ciftci, and A. Demirci Diffusion-Weighted Imaging in the Differential Diagnosis of Simple and Hydatid Cysts of the Liver Am. J. Roentgenol., November 1, 2007; 189(5): 1031 - 1036. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Taouli, A. J. Tolia, M. Losada, J. S. Babb, E. S. Chan, M. A. Bannan, and H. Tobias Diffusion-Weighted MRI for Quantification of Liver Fibrosis: Preliminary Experience Am. J. Roentgenol., October 1, 2007; 189(4): 799 - 806. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.-M. Koh and D. J. Collins Diffusion-Weighted MRI in the Body: Applications and Challenges in Oncology Am. J. Roentgenol., June 1, 2007; 188(6): 1622 - 1635. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Matoba, H. Tonami, T. Kondou, H. Yokota, K. Higashi, H. Toga, and T. Sakuma Lung Carcinoma: Diffusion-weighted MR Imaging--Preliminary Evaluation with Apparent Diffusion Coefficient Radiology, May 1, 2007; 243(2): 570 - 577. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D.-M. Koh, E. Scurr, D. Collins, B. Kanber, A. Norman, M. O. Leach, and J. E. Husband Predicting Response of Colorectal Hepatic Metastasis: Value of Pretreatment Apparent Diffusion Coefficients Am. J. Roentgenol., April 1, 2007; 188(4): 1001 - 1008. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Yoshikawa, H. Kawamitsu, D. G |
