- Email: [email protected]
Evaluation of gastric cancer by high-resolution three-dimensional CISS MR imaging in vitro
Clinical Imaging 33 (2009) 354 – 360
Evaluation of gastric cancer by high-resolution three-dimensional CISS MR imaging in vitro Ichiro Yamada a,⁎, Kimiya Takeshita b , Naoya Saito b , Norio Yoshino c , Akemi Tetsumura c , Jiro Kumagai d , Hitoshi Shibuya a a
Department of Diagnostic Radiology and Oncology, Graduate School, Tokyo Medical and Dental University, Tokyo, Japan b Department of Surgery, Tokyo Medical and Dental University, Tokyo, Japan c Department of Oral and Maxillofacial Radiology, Tokyo Medical and Dental University, Tokyo, Japan d Department of Pathology, Tokyo Medical and Dental University, Tokyo, Japan Received 1 September 2008; accepted 8 January 2009
Abstract Objective: The objective of this study was to demonstrate the usefulness of high-resolution three-dimensional (3D) constructive interference in steady-state (CISS) MR imaging for evaluating mural invasion and the morphologic features of gastric cancers in vitro. Materials and methods: High-resolution 3D-CISS MR images were obtained in three surgical specimens containing three different gastric cancers. Results: In early carcinoma, advanced carcinoma, and leiomyosarcoma, the depth of mural invasion at 3D-CISS MR imaging correlated well with the histopathologic stage. Conclusion: High-resolution 3D-CISS MR imaging is a useful method for evaluating mural invasion and the macroscopic features of gastric cancers in vitro. © 2009 Elsevier Inc. All rights reserved. Keywords: Gastric cancer; High resolution; CISS; MR imaging; Staging
1. Introduction The prognosis of patients with gastric cancer is closely correlated with the pathological stage, and accurate preoperative staging of gastric cancer definitely improves the selection of the most appropriate therapy [1,2]. Since assessment of the depth of cancer invasion, which is one of the most important factors in staging, requires that the layers of the gastric wall be depicted, CT and endoscopic US have been widely used to assess depth of invasion, but the accuracy of tumor staging by these methods is still a matter of controversy [2,3]. Thus, the diagnostic ⁎ Corresponding author. Department of Diagnostic Radiology and Oncology, Graduate School, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. Tel.: +81 3 5803 5310; fax: +81 3 5803 0147. E-mail address: [email protected] (I. Yamada). 0899-7071/09/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.clinimag.2009.01.011
methods currently available to evaluate depth of invasion are very limited. Previous studies have reported the usefulness and limitations of T2-weighted MR imaging for evaluating mural invasion of gastric cancer [4–8], but the spatial resolution of conventional spin-echo MR images for depicting detailed structures is limited. Three-dimensional (3D) constructive interference in steady-state (CISS) MR imaging has recently been employed to obtain highspatial-resolution and heavily T2-weighted images in intracranial diseases, esophageal cancer, and colorectal cancer [9–12]. To our knowledge, however, there have been no reports on the use of 3D-CISS MR imaging to evaluate gastric cancer in surgical specimens. Thus, the purpose of our study was to assess the usefulness of highresolution 3D-CISS MR imaging as a method of evaluating mural invasion and the morphologic features of gastric cancer in vitro.
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
355
2. Materials and methods 2.1. Materials The materials consisted of three surgical specimens containing three gastric cancers that had been obtained from three consecutive gastric cancer patients at our institution and that had been histopathologically confirmed to be gastric cancer. A 62-year-old man had early gastric carcinoma, a 57-year-old man had advanced gastric carcinoma, and a 62-year-old man had leiomyosarcoma of the stomach. All specimens were imaged after fixation in formalin. 2.2. Imaging technique High-resolution MR imaging was performed by using a 1.5-T system with a 25-mT/m maximum gradient capability (Magnetom Vision; Siemens, Erlangen, Germany) and a 4-cm-diameter loop coil. High-resolution 3D MR imaging was performed by using a 3D-CISS sequence having the following parameters: 12.25 ms/5.9 ms (repetition time/echo time), flip angle of 70°, and two signals acquired. The acquisition matrix was 512×512 for a field of view of 80 mm×80 mm, and the other imaging parameters included a slab thickness of 47.5 mm, 68 slices, and a bandwidth of 195 Hz per pixel, which yielded a slice thickness of 0.7 mm and a voxel size of 0.017 mm3. The acquisition time was 14 min 14 s. The orientation of the acquisition slab was along the longitudinal axis of the resected stomach, so that the entire tumor lesion was imaged. The 3D data sets were postprocessed with image-processing software dedicated to DICOM images (OsiriX version 2.4; Digital Imaging Unit, University Hospital of Geneva, Switzerland) and commercially available software (VoxBlast; VayTek Inc., Fairfield, USA), and surface rendering was performed to generate virtual MR endoscopic images of the stomach.
Fig. 1. Images of the normal gastric wall. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) clearly depicts the normal gastric wall as having four to six layers, which correlate well with the histopathologic layers: m, mucus; M, mucosa; SM, submucosa; ICM, inner circular muscle; ICT, intermuscular connective tissue; OLM, outer longitudinal muscle; SS, subserosa. (B) Histopathologic section of the normal gastric wall shows the mucosa, submucosa, muscularis propria (inner circular muscle, intermuscular connective tissue, and outer longitudinal muscle), and subserosa (hematoxylin–eosin stain; original magnification, ×3.3).
2.3. Image analysis The MR images of each lesion were interpreted by two independent radiologists (I.Y. and N.Y.) who were blinded Table 1 Signal intensity of the layers of the gastric wall on high-resolution 3D-CISS MR images Structure
3D-CISS MR images
Mucus Mucosa Submucosa Submucosal fat Muscularis propria Inner circular muscle Intermuscular connective tissue Outer longitudinal muscle Subserosa or serosa Subserosal fat
High Low High Low Low Low High Low High Low
to the histopathologic findings. The histopathologic findings were used as the reference standard for analysis of the MR imaging findings. The MR images were reviewed for signal intensity and continuity of each layer of the normal gastric wall in the three surgical specimens, and thus, we analyzed the signal intensity characteristics of the layers of the normal gastric wall on high-resolution 3D-CISS MR images. Based on high-resolution 3D-CISS MR images, the depth of tumor invasion of the gastric wall was recorded as the deepest layer invaded: mucosa, submucosa, muscularis propria, or subserosa or serosa. The macroscopic type of the tumor was assessed by examining virtual MR endoscopic images, based on the Japanese Classification of Gastric Carcinoma by the Japanese Research Society for Gastric Cancer [13]. Highresolution 3D-CISS MR images were compared with
356
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
histopathologic findings, and virtual MR endoscopic images were compared with macroscopic findings at surgery.
3. Results 3.1. Signal intensity of the layers of the normal gastric wall The signal intensity of the layers of the normal gastric wall on high-resolution 3D-CISS MR images is shown in Table 1. The high-resolution 3D-CISS MR images depicted the surface material, including adherent mucus, as high signal intensity, the mucosa, including the epithelial glands and lamina propria mucosae, as low signal intensity, and the muscularis mucosae as low signal intensity (Fig. 1). The submucosa was depicted as high signal intensity, and histopathologic examination showed that it contained loose connective tissue, capillary vessels, and lymph channels.
The submucosal fat tissue appeared as low signal intensity. The muscularis propria was depicted as low signal intensity, but the high-resolution 3D-CISS MR images often resolved the muscularis propria into three layers: an inner circular muscle layer and an outer longitudinal muscle layer visualized as low-signal-intensity structures separated by a thin, high-signal-intensity band. The subserosa was depicted as high signal intensity, and the subserosal fat tissue appeared as low signal intensity. 3.2. Case 1: early gastric carcinoma A 62-year-old man underwent partial gastrectomy for a gastric lesion detected during a routine medical examination. High-resolution 3D-CISS MR images of the surgical specimen showed irregular thickenings in the mucosa, and histopathologic examination showed carcinoma confined within the mucosa (Fig. 2). Virtual MR
Fig. 2. Early gastric carcinoma confined within the mucosa. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) shows irregular thickenings (arrows) in the mucosa, and the submucosa of high signal intensity appears to be intact. (B) Corresponding histopathologic section shows carcinoma confined within the mucosa (arrows), as well as intact submucosa (hematoxylin–eosin stain; original magnification, ×1.7). (C) Virtual MR endoscopic image shows an irregularly shaped, flat tumor (arrows), indicative of a superficial tumor (type 0). (D) Macroscopic image confirms a superficial-type tumor (type 0-IIa +IIc) (arrows).
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
endoscopic images showed an irregularly shaped, flat tumor, indicating a superficial tumor (type 0), and macroscopic inspection confirmed a superficial-type tumor (type 0-IIa+IIc). 3.3. Case 2: advanced gastric carcinoma A 57-year-old man underwent partial gastrectomy for a gastric tumor detected during a routine medical examination. High-resolution 3D-CISS MR images showed an irregularly shaped tumor partially replacing the muscularis propria layer, and histopathologic examination showed carcinoma involving the muscularis propria (Fig. 3). Virtual MR endoscopic images showed a well-demarcated tumor with an irregular surface, indicating an irregularly surfaced, localized tumor (type 2), and macroscopic inspection confirmed an ulcerative and localized-type tumor (type 2).
357
3.4. Case 3: leiomyosarcoma of the stomach A 62-year-old man underwent partial gastrectomy for a gastric tumor detected during a routine medical examination. High-resolution 3D-CISS MR images showed a welldefined, lobulated tumor in the submucosa, and histopathologic examination showed a well-defined, lobulated submucosal tumor in the enlarged submucosal layer (Fig. 4). Virtual MR endoscopic images showed a welldefined, protruding tumor with a smooth surface, indicating a submucosal tumor, and macroscopic inspection confirmed a well-defined, protruding submucosal tumor.
4. Discussion Our initial findings have demonstrated that highresolution 3D-CISS MR images depict the normal gastric
Fig. 3. Advanced gastric carcinoma involving the muscularis propria. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) shows that an irregularly shaped tumor (arrows) partially replaces the muscularis propria layer (arrowheads) but that it does not penetrate through the muscularis propria layer. (B) Corresponding histopathologic section shows carcinoma involving the muscularis propria (arrowheads) (hematoxylin–eosin stain; original magnification, ×1.25). (C) Virtual MR endoscopic image shows a well-demarcated tumor with an irregular surface (arrow), indicative of an irregularly surfaced and localized tumor (type 2). (D) Macroscopic image confirms an ulcerative and localized-type tumor (type 2) (arrow).
358
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
Fig. 4. Leiomyosarcoma of the stomach. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) shows a well-defined, lobulated tumor (arrowheads) in the submucosa (SM). (B) Corresponding histopathologic section shows a well-defined, lobulated submucosal tumor (arrowheads) in the enlarged submucosa layer (SM) (hematoxylin–eosin stain; original magnification, ×1.04). (C) Virtual MR endoscopic image shows a well-defined, protruding tumor with a smooth surface (arrow), indicative of a submucosal tumor. (D) Macroscopic image confirms a well-defined, protruding submucosal tumor (arrow).
wall as consisting of four to six layers that correspond well with the layers of the gastric wall observed histopathologically. 3D-CISS MR imaging provides higher soft-tissue contrast than CT or US, and there are none of the artifactual interface echoes in the gastric wall that occur with US. Thus, high-resolution 3D-CISS MR imaging enables more accurate differentiation of the normal layers of the gastric wall than either CT or US [2,3]. Previous studies have described the gastric wall as consisting of three to six layers on conventional spin-echo T2-weighted MR images [4–8,14], whereas our results demonstrated that highresolution 3D-CISS MR imaging depicted the layers of the gastric wall observed histopathologically. Recent papers have shown that T2 relaxation times account for different signal intensities in the rectal wall layers [15]. Probably in the stomach, as well, different T2 times, which are due to histopathologic components, content of water, and the presence of fat, account for different signal intensities in the gastric wall layers.
Our findings showed that high-resolution 3D-CISS MR imaging was able to correctly depict the depth of invasion of the gastric wall in the surgical specimens containing early gastric carcinoma, advanced gastric carcinoma, and leiomyosarcoma. Thus, high-resolution 3D-CISS MR imaging was found to be a useful method for evaluating depth of invasion by gastric cancer. The reason appears to be attributable to the combination of high soft-tissue contrast and high spatial resolution that 3D-CISS MR imaging provides. Clinical MR imaging has already been employed to stage gastric cancer [4–6]. On the basis of a systematic review using articles published between 2000 and 2005, Kwee and Kwee [2] reported that conventional MR imaging for T-staging of gastric cancer had the diagnostic accuracy ranging from 71.4% to 82.6%. The previous reports suggest that conventional MR imaging may have substantial limitations in regard to the T-staging of gastric cancer [2,4–8]. Therefore, more accurate methods than conventional MR imaging may be required for gastric cancer.
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
The 3D-CISS sequence is basically a gradient-echo-based sequence that consists of two 3D Fourier transformation sequences: (a) fast imaging with steady-state precession (FISP) and (b) PSIF (reverse FISP). By summing the signals from the FISP and PSIF sequences, it is possible to obtain heavily T2-weighted images with high spatial resolution in the 3D-CISS sequence [9–12]. With its high spatial resolution and good soft-tissue contrast, 3D-CISS MR imaging has been reported to be highly accurate for evaluating intracranial diseases, esophageal cancer, and colorectal cancer [9–12]. Because of the high spatial resolution and soft-tissue contrast, our data demonstrate that high-resolution 3D-CISS MR imaging is a useful method for evaluating the level of mural invasion in gastric cancer in surgical specimens. Our findings have also demonstrated that the same 3DCISS imaging data set can be used to generate virtual MR endoscopic images. Furthermore, the virtual MR endoscopic images clearly depicted the macroscopic type of the cancers found at surgery. Knowledge of the gross configuration of the cancer and correct depth of invasion is essential to select the most appropriate therapy for gastric cancer [1–6]. Virtual endoscopic MR images have recently been employed to detect gastric cancer [5,16], and Zhong et al. [5] have found that MR gastrography with virtual endoscopic viewing is an effective method for detection of gastric cancer. High-resolution 3D-CISS MR imaging, which has higher spatial resolution, is expected to be a useful method for the detection and macroscopic evaluation of gastric cancers. The first limitation of our study is that the specimens were imaged after fixation in formalin. However, previous reports have shown no substantial effect of formalin fixation, on the signal intensity and soft-tissue contrast of spin-echo T2-weighted images in the gastric wall, colorectal wall, and esophageal wall [14,17,18]. The 3D-CISS images are essentially T2-weighted images, though they employ a gradient-echo-based sequence. Thus, the findings in our study may be applicable to high-resolution 3D-CISS MR imaging of in vivo specimens as well as formalinfixed specimens. The second limitation is that the observers were not blinded to the presence of gastric cancer but blinded only to the depth of tumor invasion. In this preliminary study, we attempted to assess 3D-CISS MR imaging only as a confirmatory test, not as a screening examination. Thus, it was not demonstrated whether pathological lesions could be reliably distinguished from normal sites. The third limitation is that we performed only 3D-CISS imaging and compared its findings with histopathologic findings as the reference standard. Thus, a comparison of this sequence with conventional T1-weighted, T2-weighted, or fast spin-echo sequences was not made in our study. Finally, since our results were obtained by imaging specimens along the axis of the stomach using long acquisition times, the results cannot be directly extrapolated
359
to an in vivo setting. There are many technical issues associated with performing preoperative gastric imaging in vivo, including motion (peristalsis, respiratory motion, and gross patient movement), necessity for gastric distention, and limitation on lumen contents. Clinical application of highresolution MR imaging of the stomach is still in the preliminary stage. However, high-resolution MR imaging of gastric cancer in vivo may become possible with the development of faster MR imaging techniques and by using an endoluminal coil technique or phased-array coil technique [19,20]. Higher field strength (3.0 T or more) also may enable high-resolution MR imaging of the stomach in reduced acquisition times. In conclusion, our study has demonstrated that highresolution 3D-CISS MR imaging in vitro depicts the internal architecture of the gastric wall and is a useful method for evaluating mural invasion and the macroscopic features of gastric cancer. Thus, high-resolution 3D-CISS MR imaging may enable accurate preoperative tumor staging and morphologic evaluation of gastric cancer after evaluation of this technique in vivo. References [1] American Joint Committee on Cancer. Stomach. In: Greene FL, Page DL, Fleming ID, et al, editors. AJCC cancer staging manual. 6th ed. New York: Springer-Verlag, 2002. p. 99–106. [2] Kwee RM, Kwee TC. Imaging in local staging of gastric cancer: a systematic review. J Clin Oncol 2007;25:2107–16. [3] Habermann CR, Weiss F, Riecken R, et al. Preoperative staging of gastric adenocarcinoma: comparison of helical CT and endoscopic US. Radiology 2004;230:465–71. [4] Sohn K-M, Lee JM, Lee S-Y, Ahn B-Y, Park S-M, Kim K-M. Comparing MR imaging and CT in the staging of gastric carcinoma. AJR Am J Roentgenol 2000;174:1551–7. [5] Zhong K, Li L, Sun JH, Xu JR. Preoperative diagnosis of gastric cancer using 2-D magnetic resonance imaging with 3-D reconstruction techniques. Chin J Dig Dis 2005;6:159–64. [6] Arocena MG, Barturen A, Bujanda L, et al. MRI and endoscopic ultrasonography in the staging of gastric cancer. Rev Esp Enferm Dig 2006;98:582–90. [7] Yamada I, Saito N, Takeshita K, et al. Early gastric carcinoma: evaluation with high-spatial-resolution MR imaging in vitro. Radiology 2001;220:115–21. [8] Palmowski M, Grenacher L, Kuntz C, Heye T, Dux M. Magnetic resonance imaging for local staging of gastric carcinoma: results of an in vitro study. J Comput Assist Tomogr 2006;30:896–902. [9] Yamada I, Tsunoda A, Noguchi Y, Komatsuzaki A, Shibuya H. Tumor volume measurements of acoustic neuromas with three-dimensional constructive interference in steady state and conventional spin-echo MR imaging. J Magn Reson Imaging 2000;12:826–32. [10] Yoshino N, Akimoto H, Yamada I, et al. Trigeminal neuralgia: evaluation of neuralgic manifestation and site of neurovascular compression with 3D CISS MR imaging and MR angiography. Radiology 2003;228:539–45. [11] Yamada I, Izumi Y, Kawano T, et al. Esophageal carcinoma: evaluation with high-resolution three-dimensional constructive interference in steady state MR imaging in vitro. J Magn Reson Imaging 2006;24: 1326–32. [12] Yamada I, Okabe S, Enomoto M, et al. Colorectal carcinoma: in vitro evaluation with high-spatial-resolution 3D constructive interference in steady-state MR imaging. Radiology 2008;246:444–53.
360
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
[13] Japanese Research Society for Gastric Cancer, editor. Japanese classification of gastric carcinoma. 1st English ed. Tokyo: Kanehara and Co., Ltd, 1995. [14] Auh YH, Lim T-H, Lee DH, et al. In vitro MR imaging of the resected stomach with a 4.7-T superconducting magnet. Radiology 1994;191: 129–34. [15] Stollfuss JC, Becker K, Sendler A, et al. Rectal carcinoma: highspatial-resolution MR imaging and T2 quantification in rectal cancer specimens. Radiology 2006;241:132–41. [16] Schmid MR, Hany TF, Knesplova L, Schlumpf R, Debatin JF. 3D MR gastrography: exoscopic and endoscopic analysis of the stomach. Eur Radiol 1999;9:73–7.
[17] Imai Y, Kressel HY, Saul SH, et al. Colorectal tumors: an in vitro study of high-resolution MR imaging. Radiology 1990;177:695–701. [18] Yamada I, Izumi Y, Kawano T, et al. Superficial esophageal carcinoma: an in vitro study of high-resolution MR imaging at 1.5T. J Magn Reson Imaging 2001;13:225–31. [19] Marcos HB, Semelka RC. Stomach diseases: MR evaluation using combined T2-weighted single-shot echo train spin-echo and gadolinium-enhanced spoiled gradient-echo sequences. J Magn Reson Imaging 1999;10:950–60. [20] Heye T, Kuntz C, Dux M, et al. CT and endoscopic ultrasound in comparison to endoluminal MRI—preliminary results in staging gastric carcinoma. Eur J Radiol 2008, doi:10.1016/j.ejrad.2008.01.037.
Evaluation of gastric cancer by high-resolution three-dimensional CISS MR imaging in vitro Ichiro Yamada a,⁎, Kimiya Takeshita b , Naoya Saito b , Norio Yoshino c , Akemi Tetsumura c , Jiro Kumagai d , Hitoshi Shibuya a a
Department of Diagnostic Radiology and Oncology, Graduate School, Tokyo Medical and Dental University, Tokyo, Japan b Department of Surgery, Tokyo Medical and Dental University, Tokyo, Japan c Department of Oral and Maxillofacial Radiology, Tokyo Medical and Dental University, Tokyo, Japan d Department of Pathology, Tokyo Medical and Dental University, Tokyo, Japan Received 1 September 2008; accepted 8 January 2009
Abstract Objective: The objective of this study was to demonstrate the usefulness of high-resolution three-dimensional (3D) constructive interference in steady-state (CISS) MR imaging for evaluating mural invasion and the morphologic features of gastric cancers in vitro. Materials and methods: High-resolution 3D-CISS MR images were obtained in three surgical specimens containing three different gastric cancers. Results: In early carcinoma, advanced carcinoma, and leiomyosarcoma, the depth of mural invasion at 3D-CISS MR imaging correlated well with the histopathologic stage. Conclusion: High-resolution 3D-CISS MR imaging is a useful method for evaluating mural invasion and the macroscopic features of gastric cancers in vitro. © 2009 Elsevier Inc. All rights reserved. Keywords: Gastric cancer; High resolution; CISS; MR imaging; Staging
1. Introduction The prognosis of patients with gastric cancer is closely correlated with the pathological stage, and accurate preoperative staging of gastric cancer definitely improves the selection of the most appropriate therapy [1,2]. Since assessment of the depth of cancer invasion, which is one of the most important factors in staging, requires that the layers of the gastric wall be depicted, CT and endoscopic US have been widely used to assess depth of invasion, but the accuracy of tumor staging by these methods is still a matter of controversy [2,3]. Thus, the diagnostic ⁎ Corresponding author. Department of Diagnostic Radiology and Oncology, Graduate School, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. Tel.: +81 3 5803 5310; fax: +81 3 5803 0147. E-mail address: [email protected] (I. Yamada). 0899-7071/09/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.clinimag.2009.01.011
methods currently available to evaluate depth of invasion are very limited. Previous studies have reported the usefulness and limitations of T2-weighted MR imaging for evaluating mural invasion of gastric cancer [4–8], but the spatial resolution of conventional spin-echo MR images for depicting detailed structures is limited. Three-dimensional (3D) constructive interference in steady-state (CISS) MR imaging has recently been employed to obtain highspatial-resolution and heavily T2-weighted images in intracranial diseases, esophageal cancer, and colorectal cancer [9–12]. To our knowledge, however, there have been no reports on the use of 3D-CISS MR imaging to evaluate gastric cancer in surgical specimens. Thus, the purpose of our study was to assess the usefulness of highresolution 3D-CISS MR imaging as a method of evaluating mural invasion and the morphologic features of gastric cancer in vitro.
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
355
2. Materials and methods 2.1. Materials The materials consisted of three surgical specimens containing three gastric cancers that had been obtained from three consecutive gastric cancer patients at our institution and that had been histopathologically confirmed to be gastric cancer. A 62-year-old man had early gastric carcinoma, a 57-year-old man had advanced gastric carcinoma, and a 62-year-old man had leiomyosarcoma of the stomach. All specimens were imaged after fixation in formalin. 2.2. Imaging technique High-resolution MR imaging was performed by using a 1.5-T system with a 25-mT/m maximum gradient capability (Magnetom Vision; Siemens, Erlangen, Germany) and a 4-cm-diameter loop coil. High-resolution 3D MR imaging was performed by using a 3D-CISS sequence having the following parameters: 12.25 ms/5.9 ms (repetition time/echo time), flip angle of 70°, and two signals acquired. The acquisition matrix was 512×512 for a field of view of 80 mm×80 mm, and the other imaging parameters included a slab thickness of 47.5 mm, 68 slices, and a bandwidth of 195 Hz per pixel, which yielded a slice thickness of 0.7 mm and a voxel size of 0.017 mm3. The acquisition time was 14 min 14 s. The orientation of the acquisition slab was along the longitudinal axis of the resected stomach, so that the entire tumor lesion was imaged. The 3D data sets were postprocessed with image-processing software dedicated to DICOM images (OsiriX version 2.4; Digital Imaging Unit, University Hospital of Geneva, Switzerland) and commercially available software (VoxBlast; VayTek Inc., Fairfield, USA), and surface rendering was performed to generate virtual MR endoscopic images of the stomach.
Fig. 1. Images of the normal gastric wall. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) clearly depicts the normal gastric wall as having four to six layers, which correlate well with the histopathologic layers: m, mucus; M, mucosa; SM, submucosa; ICM, inner circular muscle; ICT, intermuscular connective tissue; OLM, outer longitudinal muscle; SS, subserosa. (B) Histopathologic section of the normal gastric wall shows the mucosa, submucosa, muscularis propria (inner circular muscle, intermuscular connective tissue, and outer longitudinal muscle), and subserosa (hematoxylin–eosin stain; original magnification, ×3.3).
2.3. Image analysis The MR images of each lesion were interpreted by two independent radiologists (I.Y. and N.Y.) who were blinded Table 1 Signal intensity of the layers of the gastric wall on high-resolution 3D-CISS MR images Structure
3D-CISS MR images
Mucus Mucosa Submucosa Submucosal fat Muscularis propria Inner circular muscle Intermuscular connective tissue Outer longitudinal muscle Subserosa or serosa Subserosal fat
High Low High Low Low Low High Low High Low
to the histopathologic findings. The histopathologic findings were used as the reference standard for analysis of the MR imaging findings. The MR images were reviewed for signal intensity and continuity of each layer of the normal gastric wall in the three surgical specimens, and thus, we analyzed the signal intensity characteristics of the layers of the normal gastric wall on high-resolution 3D-CISS MR images. Based on high-resolution 3D-CISS MR images, the depth of tumor invasion of the gastric wall was recorded as the deepest layer invaded: mucosa, submucosa, muscularis propria, or subserosa or serosa. The macroscopic type of the tumor was assessed by examining virtual MR endoscopic images, based on the Japanese Classification of Gastric Carcinoma by the Japanese Research Society for Gastric Cancer [13]. Highresolution 3D-CISS MR images were compared with
356
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
histopathologic findings, and virtual MR endoscopic images were compared with macroscopic findings at surgery.
3. Results 3.1. Signal intensity of the layers of the normal gastric wall The signal intensity of the layers of the normal gastric wall on high-resolution 3D-CISS MR images is shown in Table 1. The high-resolution 3D-CISS MR images depicted the surface material, including adherent mucus, as high signal intensity, the mucosa, including the epithelial glands and lamina propria mucosae, as low signal intensity, and the muscularis mucosae as low signal intensity (Fig. 1). The submucosa was depicted as high signal intensity, and histopathologic examination showed that it contained loose connective tissue, capillary vessels, and lymph channels.
The submucosal fat tissue appeared as low signal intensity. The muscularis propria was depicted as low signal intensity, but the high-resolution 3D-CISS MR images often resolved the muscularis propria into three layers: an inner circular muscle layer and an outer longitudinal muscle layer visualized as low-signal-intensity structures separated by a thin, high-signal-intensity band. The subserosa was depicted as high signal intensity, and the subserosal fat tissue appeared as low signal intensity. 3.2. Case 1: early gastric carcinoma A 62-year-old man underwent partial gastrectomy for a gastric lesion detected during a routine medical examination. High-resolution 3D-CISS MR images of the surgical specimen showed irregular thickenings in the mucosa, and histopathologic examination showed carcinoma confined within the mucosa (Fig. 2). Virtual MR
Fig. 2. Early gastric carcinoma confined within the mucosa. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) shows irregular thickenings (arrows) in the mucosa, and the submucosa of high signal intensity appears to be intact. (B) Corresponding histopathologic section shows carcinoma confined within the mucosa (arrows), as well as intact submucosa (hematoxylin–eosin stain; original magnification, ×1.7). (C) Virtual MR endoscopic image shows an irregularly shaped, flat tumor (arrows), indicative of a superficial tumor (type 0). (D) Macroscopic image confirms a superficial-type tumor (type 0-IIa +IIc) (arrows).
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
endoscopic images showed an irregularly shaped, flat tumor, indicating a superficial tumor (type 0), and macroscopic inspection confirmed a superficial-type tumor (type 0-IIa+IIc). 3.3. Case 2: advanced gastric carcinoma A 57-year-old man underwent partial gastrectomy for a gastric tumor detected during a routine medical examination. High-resolution 3D-CISS MR images showed an irregularly shaped tumor partially replacing the muscularis propria layer, and histopathologic examination showed carcinoma involving the muscularis propria (Fig. 3). Virtual MR endoscopic images showed a well-demarcated tumor with an irregular surface, indicating an irregularly surfaced, localized tumor (type 2), and macroscopic inspection confirmed an ulcerative and localized-type tumor (type 2).
357
3.4. Case 3: leiomyosarcoma of the stomach A 62-year-old man underwent partial gastrectomy for a gastric tumor detected during a routine medical examination. High-resolution 3D-CISS MR images showed a welldefined, lobulated tumor in the submucosa, and histopathologic examination showed a well-defined, lobulated submucosal tumor in the enlarged submucosal layer (Fig. 4). Virtual MR endoscopic images showed a welldefined, protruding tumor with a smooth surface, indicating a submucosal tumor, and macroscopic inspection confirmed a well-defined, protruding submucosal tumor.
4. Discussion Our initial findings have demonstrated that highresolution 3D-CISS MR images depict the normal gastric
Fig. 3. Advanced gastric carcinoma involving the muscularis propria. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) shows that an irregularly shaped tumor (arrows) partially replaces the muscularis propria layer (arrowheads) but that it does not penetrate through the muscularis propria layer. (B) Corresponding histopathologic section shows carcinoma involving the muscularis propria (arrowheads) (hematoxylin–eosin stain; original magnification, ×1.25). (C) Virtual MR endoscopic image shows a well-demarcated tumor with an irregular surface (arrow), indicative of an irregularly surfaced and localized tumor (type 2). (D) Macroscopic image confirms an ulcerative and localized-type tumor (type 2) (arrow).
358
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
Fig. 4. Leiomyosarcoma of the stomach. (A) High-resolution 3D-CISS MR image (12.25 ms/5.9 ms; flip angle, 70°) shows a well-defined, lobulated tumor (arrowheads) in the submucosa (SM). (B) Corresponding histopathologic section shows a well-defined, lobulated submucosal tumor (arrowheads) in the enlarged submucosa layer (SM) (hematoxylin–eosin stain; original magnification, ×1.04). (C) Virtual MR endoscopic image shows a well-defined, protruding tumor with a smooth surface (arrow), indicative of a submucosal tumor. (D) Macroscopic image confirms a well-defined, protruding submucosal tumor (arrow).
wall as consisting of four to six layers that correspond well with the layers of the gastric wall observed histopathologically. 3D-CISS MR imaging provides higher soft-tissue contrast than CT or US, and there are none of the artifactual interface echoes in the gastric wall that occur with US. Thus, high-resolution 3D-CISS MR imaging enables more accurate differentiation of the normal layers of the gastric wall than either CT or US [2,3]. Previous studies have described the gastric wall as consisting of three to six layers on conventional spin-echo T2-weighted MR images [4–8,14], whereas our results demonstrated that highresolution 3D-CISS MR imaging depicted the layers of the gastric wall observed histopathologically. Recent papers have shown that T2 relaxation times account for different signal intensities in the rectal wall layers [15]. Probably in the stomach, as well, different T2 times, which are due to histopathologic components, content of water, and the presence of fat, account for different signal intensities in the gastric wall layers.
Our findings showed that high-resolution 3D-CISS MR imaging was able to correctly depict the depth of invasion of the gastric wall in the surgical specimens containing early gastric carcinoma, advanced gastric carcinoma, and leiomyosarcoma. Thus, high-resolution 3D-CISS MR imaging was found to be a useful method for evaluating depth of invasion by gastric cancer. The reason appears to be attributable to the combination of high soft-tissue contrast and high spatial resolution that 3D-CISS MR imaging provides. Clinical MR imaging has already been employed to stage gastric cancer [4–6]. On the basis of a systematic review using articles published between 2000 and 2005, Kwee and Kwee [2] reported that conventional MR imaging for T-staging of gastric cancer had the diagnostic accuracy ranging from 71.4% to 82.6%. The previous reports suggest that conventional MR imaging may have substantial limitations in regard to the T-staging of gastric cancer [2,4–8]. Therefore, more accurate methods than conventional MR imaging may be required for gastric cancer.
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
The 3D-CISS sequence is basically a gradient-echo-based sequence that consists of two 3D Fourier transformation sequences: (a) fast imaging with steady-state precession (FISP) and (b) PSIF (reverse FISP). By summing the signals from the FISP and PSIF sequences, it is possible to obtain heavily T2-weighted images with high spatial resolution in the 3D-CISS sequence [9–12]. With its high spatial resolution and good soft-tissue contrast, 3D-CISS MR imaging has been reported to be highly accurate for evaluating intracranial diseases, esophageal cancer, and colorectal cancer [9–12]. Because of the high spatial resolution and soft-tissue contrast, our data demonstrate that high-resolution 3D-CISS MR imaging is a useful method for evaluating the level of mural invasion in gastric cancer in surgical specimens. Our findings have also demonstrated that the same 3DCISS imaging data set can be used to generate virtual MR endoscopic images. Furthermore, the virtual MR endoscopic images clearly depicted the macroscopic type of the cancers found at surgery. Knowledge of the gross configuration of the cancer and correct depth of invasion is essential to select the most appropriate therapy for gastric cancer [1–6]. Virtual endoscopic MR images have recently been employed to detect gastric cancer [5,16], and Zhong et al. [5] have found that MR gastrography with virtual endoscopic viewing is an effective method for detection of gastric cancer. High-resolution 3D-CISS MR imaging, which has higher spatial resolution, is expected to be a useful method for the detection and macroscopic evaluation of gastric cancers. The first limitation of our study is that the specimens were imaged after fixation in formalin. However, previous reports have shown no substantial effect of formalin fixation, on the signal intensity and soft-tissue contrast of spin-echo T2-weighted images in the gastric wall, colorectal wall, and esophageal wall [14,17,18]. The 3D-CISS images are essentially T2-weighted images, though they employ a gradient-echo-based sequence. Thus, the findings in our study may be applicable to high-resolution 3D-CISS MR imaging of in vivo specimens as well as formalinfixed specimens. The second limitation is that the observers were not blinded to the presence of gastric cancer but blinded only to the depth of tumor invasion. In this preliminary study, we attempted to assess 3D-CISS MR imaging only as a confirmatory test, not as a screening examination. Thus, it was not demonstrated whether pathological lesions could be reliably distinguished from normal sites. The third limitation is that we performed only 3D-CISS imaging and compared its findings with histopathologic findings as the reference standard. Thus, a comparison of this sequence with conventional T1-weighted, T2-weighted, or fast spin-echo sequences was not made in our study. Finally, since our results were obtained by imaging specimens along the axis of the stomach using long acquisition times, the results cannot be directly extrapolated
359
to an in vivo setting. There are many technical issues associated with performing preoperative gastric imaging in vivo, including motion (peristalsis, respiratory motion, and gross patient movement), necessity for gastric distention, and limitation on lumen contents. Clinical application of highresolution MR imaging of the stomach is still in the preliminary stage. However, high-resolution MR imaging of gastric cancer in vivo may become possible with the development of faster MR imaging techniques and by using an endoluminal coil technique or phased-array coil technique [19,20]. Higher field strength (3.0 T or more) also may enable high-resolution MR imaging of the stomach in reduced acquisition times. In conclusion, our study has demonstrated that highresolution 3D-CISS MR imaging in vitro depicts the internal architecture of the gastric wall and is a useful method for evaluating mural invasion and the macroscopic features of gastric cancer. Thus, high-resolution 3D-CISS MR imaging may enable accurate preoperative tumor staging and morphologic evaluation of gastric cancer after evaluation of this technique in vivo. References [1] American Joint Committee on Cancer. Stomach. In: Greene FL, Page DL, Fleming ID, et al, editors. AJCC cancer staging manual. 6th ed. New York: Springer-Verlag, 2002. p. 99–106. [2] Kwee RM, Kwee TC. Imaging in local staging of gastric cancer: a systematic review. J Clin Oncol 2007;25:2107–16. [3] Habermann CR, Weiss F, Riecken R, et al. Preoperative staging of gastric adenocarcinoma: comparison of helical CT and endoscopic US. Radiology 2004;230:465–71. [4] Sohn K-M, Lee JM, Lee S-Y, Ahn B-Y, Park S-M, Kim K-M. Comparing MR imaging and CT in the staging of gastric carcinoma. AJR Am J Roentgenol 2000;174:1551–7. [5] Zhong K, Li L, Sun JH, Xu JR. Preoperative diagnosis of gastric cancer using 2-D magnetic resonance imaging with 3-D reconstruction techniques. Chin J Dig Dis 2005;6:159–64. [6] Arocena MG, Barturen A, Bujanda L, et al. MRI and endoscopic ultrasonography in the staging of gastric cancer. Rev Esp Enferm Dig 2006;98:582–90. [7] Yamada I, Saito N, Takeshita K, et al. Early gastric carcinoma: evaluation with high-spatial-resolution MR imaging in vitro. Radiology 2001;220:115–21. [8] Palmowski M, Grenacher L, Kuntz C, Heye T, Dux M. Magnetic resonance imaging for local staging of gastric carcinoma: results of an in vitro study. J Comput Assist Tomogr 2006;30:896–902. [9] Yamada I, Tsunoda A, Noguchi Y, Komatsuzaki A, Shibuya H. Tumor volume measurements of acoustic neuromas with three-dimensional constructive interference in steady state and conventional spin-echo MR imaging. J Magn Reson Imaging 2000;12:826–32. [10] Yoshino N, Akimoto H, Yamada I, et al. Trigeminal neuralgia: evaluation of neuralgic manifestation and site of neurovascular compression with 3D CISS MR imaging and MR angiography. Radiology 2003;228:539–45. [11] Yamada I, Izumi Y, Kawano T, et al. Esophageal carcinoma: evaluation with high-resolution three-dimensional constructive interference in steady state MR imaging in vitro. J Magn Reson Imaging 2006;24: 1326–32. [12] Yamada I, Okabe S, Enomoto M, et al. Colorectal carcinoma: in vitro evaluation with high-spatial-resolution 3D constructive interference in steady-state MR imaging. Radiology 2008;246:444–53.
360
I. Yamada et al. / Clinical Imaging 33 (2009) 354–360
[13] Japanese Research Society for Gastric Cancer, editor. Japanese classification of gastric carcinoma. 1st English ed. Tokyo: Kanehara and Co., Ltd, 1995. [14] Auh YH, Lim T-H, Lee DH, et al. In vitro MR imaging of the resected stomach with a 4.7-T superconducting magnet. Radiology 1994;191: 129–34. [15] Stollfuss JC, Becker K, Sendler A, et al. Rectal carcinoma: highspatial-resolution MR imaging and T2 quantification in rectal cancer specimens. Radiology 2006;241:132–41. [16] Schmid MR, Hany TF, Knesplova L, Schlumpf R, Debatin JF. 3D MR gastrography: exoscopic and endoscopic analysis of the stomach. Eur Radiol 1999;9:73–7.
[17] Imai Y, Kressel HY, Saul SH, et al. Colorectal tumors: an in vitro study of high-resolution MR imaging. Radiology 1990;177:695–701. [18] Yamada I, Izumi Y, Kawano T, et al. Superficial esophageal carcinoma: an in vitro study of high-resolution MR imaging at 1.5T. J Magn Reson Imaging 2001;13:225–31. [19] Marcos HB, Semelka RC. Stomach diseases: MR evaluation using combined T2-weighted single-shot echo train spin-echo and gadolinium-enhanced spoiled gradient-echo sequences. J Magn Reson Imaging 1999;10:950–60. [20] Heye T, Kuntz C, Dux M, et al. CT and endoscopic ultrasound in comparison to endoluminal MRI—preliminary results in staging gastric carcinoma. Eur J Radiol 2008, doi:10.1016/j.ejrad.2008.01.037.