- Email: [email protected]
Novel magnified endoscopic findings of microvascular architecture in intramucosal gastric cancer
Novel magnified endoscopic findings of microvascular architecture in intramucosal gastric cancer Kenshi Yao, MD, PhD, Tatsuhiro Oishi, MD, PhD, Toshiyuki Matsui, MD, PhD, Tsuneyoshi Yao, MD, PhD, Akinori Iwashita, MD, PhD Background: The color change observed endoscopically in early gastric cancer is thought to correlate with vascular density and architecture. This study investigated the endoscopic microvascular architecture in intramucosal gastric carcinoma in vivo. Methods: Intramucosal gastric carcinomas without ulceration in 27 patients were studied by using a new magnifying upper endoscope with attention to microvascular findings. The carcinomas were divided into two major types histologically: differentiated (18) and undifferentiated (9). Results: A regular subepithelial capillary network was demonstrated in noncancerous mucosa. The appearance of the carcinomas differed depending on histologic differentiation. With all of the differentiated carcinomas, there was a well-demarcated area where the regular capillary pattern of noncancerous mucosa had disappeared and irregular microvessels were proliferating. In contrast, with undifferentiated carcinomas there was only an ill-defined area with disappearance or a reduction in the density of capillaries in the noncancerous mucosa. Conclusions: Magnified endoscopic observation of microvessels may be of assistance in the identification of intramucosal gastric carcinomas that exhibit only subtle changes in color and shape at standard endoscopy.
Color change is a useful marker for the endoscopic identification of early stage gastric cancer and for determination of the spread of carcinoma.1,2 Our clinical study demonstrated that this color change in early stage gastric cancer depended mainly on mucosal hemoglobin content and that this was increased in differentiated intramucosal gastric carcinoma and decreased in undifferentiated intramucosal gastric cancer.3 In addition, the pattern of distribution in early stage gastric carcinoma was found to be more irregular in differentiated carcinomatous Received October 1, 2001. For revision January 4, 2002. Accepted March 7, 2002. Current affiliations: Departments of Gastroenterology and Pathology, Fukuoka University, Chikushi Hospital, Fukuoka, Japan. Presented in part at the annual Digestive Disease Week, May 1923, 2002, San Francisco, California (Gastrointest Endosc 2002;55:AB128). Reprint requests: Kenshi Yao, MD, PhD, Fukuoka University, Chikushi Hospital, 377-1 Ooaza Zokumyoin, Chikushino-city, Fukuoka Prefecture, Japan 818-8502. Copyright © 2002 by the American Society for Gastrointestinal Endoscopy 0016-5107/2002/$35.00 + 0 37/69/126061 doi:10.1067/mge.2002.126061 VOLUME 56, NO. 2, 2002
mucosa than in noncancerous background mucosa.4 Furthermore, it has been well demonstrated in pathologic studies that the change in color evident at endoscopy (i.e., mucosal hemoglobin content) probably originates from differences in the density and architecture of the microvasculature structure between cancerous and noncancerous mucosa.5,6 However, the significance of these changes in microvascular architecture has not been clarified in clinical investigations. In preliminary studies, a new magnifying upper endoscopy system clearly visualized a subepithelial microvascular structure or loop in the gastric mucosa.7,8 In the present study, the magnifying endoscopic appearance of the microvascular architecture was determined for both cancerous and noncancerous mucosa in vivo. PATIENTS AND METHODS Instrument specifications and endoscopic procedure A new magnifying endoscope (GIF-Q240Z, Olympus Optical Co., Ltd., Tokyo, Japan) for use in upper GI tract with an electronic endoscope system (EVIS 240, Olympus) was used in this study. The diameter and length of the insertion tube of this instrument are the same as those of a standard upper endoscope (e.g., GIF-Q240, Olympus). The new instrument can magnify the endoscopic image optically up to 80-fold through use of a zoom attachment. Before the examination, a food particle was mounted on the tip of the endoscope to enable the endoscopist to fix the focal distance at 3 mm between the tip of the instrument and the mucosal surface, the point at which maximal magnification of the endoscopic image was to be obtained. Preparation of patients for the magnifying examination was the same as that for standard upper endoscopy. For magnifying observation, the tip of the endoscope was allowed to touch the mucosa immediately after zooming up to the maximal magnifying level, and then a magnified endoscopic image in focus could be easily obtained. Maximal resolution of the magnifying endoscope in vitro The maximal resolution of the new magnifying system was tested under laboratory conditions9,10 to determine whether it provided adequate resolution to display the microvascular structure of the gastric mucosa, as described in another in vitro study.11 For this purpose, a resolution chart (Olympus) was used on which black lines of various known widths and lengths are displayed. The chart was placed on a stage arranged perpendicular to the axis of the insertion tube of the endoscope. At maximal magnification, the minimal width ( m) of the black lines that could be separated visually on the endoscopic monitor was determined and taken as the maximal resolution of the magnifying system.9 Patients and histopathologic investigations Twenty-seven consecutive patients with intramucosal gastric carcinomas who underwent magnifying endoscopy between May 2000 and August 2001 were included in this GASTROINTESTINAL ENDOSCOPY
279
K Yao, T Oishi, T Matsui, et al.
Figure 1. Nonmagnified endoscopic view of differentiated superficial depressed type carcinoma. An ill-defined area with faint, unusual redness is shown on the greater curvature of the antrum. uncontrolled, prospective cohort study. The protocol was approved by the Human Subject Committee of our university, and informed consent was obtained from all patients. In all patients, the gastric carcinoma had been diagnosed previously at conventional endoscopy including biopsies with histopathologic confirmation. Thus, the general endoscopic findings including macroscopic type and location of the lesion were known before magnifying endoscopy. Macroscopic type of the carcinoma was classified according to the classification for early stage gastric cancer of the Japanese Research Society of Gastric Cancer (Type I, protruded; Type IIa, superficial elevated; Type IIb, flat; Type IIc, superficial depressed; and Type III, excavated).12 Standard (nonmagnifying) and subsequent magnifying endoscopies were performed preoperatively to define the margin between carcinoma and noncarcinoma, and to determine depth of invasion. Standard (nonmagnifying) endoscopy was performed first to identify abnormal mucosal areas including reddened, slightly depressed or flat mucosa with irregular margin, flat focal lesions with ill-defined margins accompanied by abnormal redness, or focally pale mucosa.2 Immediately after taking endoscopic photographs of a non-magnified, abnormal mucosal lesion, the appearance of which raised a strong suspicion for carcinoma, the mucosal lesion and the surrounding mucosa were observed at maximal magnification with the zoom attachment of the magnifying endoscope. Once the magnified image was in focus, magnified endoscopic photographs were taken. If the observed area did not exhibit findings on nonmagnified endoscopy typical of carcinoma, a biopsy specimen was taken from the area at which the magnified endoscopic image was obtained. The following findings were recorded at standard endoscopy: macroscopic classification, endoscopic color (i.e., reddened, almost the same as the surrounding mucosa, or pale) and location of the carcinoma. The following magnifying endoscopic findings were recorded: microvascular architecture in both the cancerous and surrounding mucosa. One investigator (K.Y.) performed all of the magnifying endoscopic examinations and recorded all of the endoscopic findings. 280
GASTROINTESTINAL ENDOSCOPY
Intramucosal gastric cancer: microvascular architecture
After the carcinomas had been resected either endoscopically or surgically, the resected specimens were extended on boards with pins and fixed in 20% formalin. The cancerous lesions together with the surrounding noncancerous mucosa were cut into 2- to 5-mm wide serial step sections, at least one of which corresponded to the portion of the carcinomatous and noncarcinomatous mucosa in the magnified endoscopic images, as determined in discussions with a pathologist (T.O.). The histopathologic sections were interpreted by an independent pathologist (A.K.) blinded to the endoscopic findings. The histopathologic diagnostic criteria were based on the Vienna classification of GI epithelial neoplasia (Category 1, negative for neoplasia/dysplasia; Category 2, indefinite for neoplasia/dysplasia; Category 3, noninvasive low-grade neoplasia; Category 4, noninvasive highgrade dysplasia; Category 5, invasive neoplasia [5.1, intramucosal carcinoma; 5.2. submucosal carcinoma]).13 The carcinomas were further classified into 2 major differentiation groups, i.e., differentiated carcinoma which had tubular or papillary structures resembling metaplastic intestinal epithelium, and undifferentiated carcinoma, including poorly differentiated carcinoma, which was characterized by discretely proliferating tumor cells with sparse or no tubular formation.14 In addition, the sections were evaluated histologically for ulceration (i.e., regenerative epithelium or submucosal fibrosis). Histopathologic findings refer to the section that corresponded to the magnified endoscopic view. Carcinomas of the superficial depressed or flat types as determined at standard endoscopy were included in this study. Protruded and superficial elevated types were excluded as these are easily identified by standard endoscopy alone. Patients were excluded if the endoscopic or histopathologic findings included ulceration within the lesion, or submucosal invasion, because both of these findings may influence the microvascular architecture of the lesion.4
RESULTS Resolution of magnifying endoscope At maximal magnification, the smallest line width separable on the resolution chart was 7.874 m; lines 7.016 m in width could not be separated. Accordingly, the resolution power of this magnifying endoscope was determined to be 7.9 m. Nonmagnified endoscopic and histopathologic findings With nonmagnified endoscopy, all lesions had a slightly depressed or flat mucosal appearance with poor demarcation. Twenty-three carcinomas were of the superficial depressed type and 4 were flat type. On nonmagnified observation the color of the lesions was as follows: reddened, 18; pale, 8; and same color as surrounding mucosa, 1. Eighteen carcinomas were located in the gastric body; 9 were in the gastric antrum. Eleven carcinomas were resected by using VOLUME 56, NO. 2, 2002
Intramucosal gastric cancer: microvascular architecture
K Yao, T Oishi, T Matsui, et al.
A Figure 3. Photomicrograph of section corresponding to magnified view of marginal area between carcinoma and noncarcinomatous mucosa as shown in Figure 2C (H&E, orig. mag. 13). The margin (black arrow) is distinct. The cancerous tissue (area to left of arrow) consists of atypical glands accompanied by irregular interstitial tissue. The cancerous tissue has invaded expansively within the mucosa.
B
the endoscopic mucosal resection (EMR) technique15 and 16 were resected surgically. The median diameter of the differentiated and undifferentiated carcinomas was, respectively, 14 mm (range, 5-45 mm) and 25 mm (range, 12-35 mm). Histopathologic evaluation disclosed that all 27 carcinomas had invaded the lamina propria but not the submucosa (i.e., Category 5.1, intramucosal carcinoma, Vienna classification).13 There were 19 differentiated and 8 undifferentiated carcinomas. In comparing the endoscopic color to histologic type, 18 of the differentiated carcinomas were reddened and 1 was almost the same color as the surrounding mucosa, whereas all 8 undifferentiated carcinomas were pale. Mucosal microvascular findings by magnifying endoscopy
C Figure 2. Magnified endoscopic views of area indicated by white arrow in Figure 1. A, Magnified view of noncancerous mucosa showing subepithelial mesh-like capillary network whose shape and distribution exhibits a regular pattern. B, Magnified view of cancerous mucosa showing tortuous, irregularly dilated microvessels of various shapes such as branched (yellow arrow), loop-like (blue arrow), or ring-like (green arrow). The pattern is irregular compared with that for the capillaries in the noncancerous mucosa. C, Magnified view of border between cancerous and noncancerous mucosa showing well-demarcated line (white arrow).
VOLUME 56, NO. 2, 2002
Magnification findings were successfully obtained of both the noncancerous and cancerous mucosa and compared with histopathologic findings in all cases (Figs. 1-6). Noncancerous mucosa manifested a subepithelial capillary network or capillary loops in regular arrangement (Fig. 2A). Magnified views of the surface microvascular structure of the carcinomatous mucosa differed depending on histopathologic differentiation. All 19 differentiated carcinomas had a well-demarcated area within which the regular subepithelial capillary network of the surrounding noncancerous mucosa was replaced by microvessels of irregular size, shape and distribution (Fig. 3). The histologic findings corresponding to this magnified view of differentiated carcinoma included carcinomatous tissue composed of atypical glands in a tubular or papillary pattern with irregGASTROINTESTINAL ENDOSCOPY
281
K Yao, T Oishi, T Matsui, et al.
Intramucosal gastric cancer: microvascular architecture
A Figure 4. Nonmagnified endoscopic view of undifferentiated superficial depressed type carcinoma. An ill-defined area (arrow) of pale mucosa is present in the posterior wall of the gastric body.
ular interstitial tissue that had expanded to form an abrupt transition with noncancerous tissue (Fig. 3). In contrast, the magnified microvascular findings in the 9 undifferentiated carcinomas consisted of an ill-defined area within which there was either disappearance or a decrease in the density of the normal regular subepithelial capillary network of the surrounding noncancerous mucosa, but proliferation of irregularly shaped vessels was not observed (Fig. 5). Histopathologically, carcinomatous cells had infiltrated discretely and sparsely, without proliferation of interstitial tissue, within the lamina propria of the noncancerous mucosa (Fig. 6). DISCUSSION Previously developed magnifying endoscopy systems do not have adequate resolution for visualization of microvessels, including capillaries.16,17 The resolution of the endoscope used in the present study was 7.9 m. Because the minimum diameter of subepithelial capillaries in the gastric mucosa is found in anatomic studies to be 8 m,11 the magnifying endoscope used in the study can visualize microvessels and capillaries that are 8 m or greater in diameter. This improved resolution made it possible to observe the characteristics of microvessels in carcinomatous and noncarcinomatous mucosa for the first time in vivo. The preparation of patients for magnifying endoscopy and the procedure itself are the same as for standard upper endoscopy. For magnifying observation of mucosal microvessels, no special techniques, such as vital mucosal staining18 or acid loading,19 are needed. The method described is both simple and noninvasive. Magnifying endoscopy of early stage gastric carcinoma may be useful for identifying intramucosal carcinomas that can be cured by EMR or minimally 282
GASTROINTESTINAL ENDOSCOPY
B Figure 5. Magnified views of pale area indicated by white arrow in Figure 4. A, Magnified view of pale cancerous mucosa showing disappearance of mucosal capillary network without any proliferation of vessels (white arrows). B, Magnified endoscopic view of marginal area between carcinomatous tissue and noncarcinomatous mucosa showing gradual reduction in vascularity between subepithelial noncancerous capillary network and cancerous mucosa (white arrow).
invasive surgery, thereby offering the patient a prognosis comparable with that achievable by conventional surgery.20-22 The endoscopic diagnosis of intramucosal gastric carcinoma of the superficial depressed type or flat type with nonmagnified instruments is often difficult because such carcinomas, so-called “gastritis-like cancer,” are manifest as only subtle changes in color and shape.23 Differentiated carcinoma is often described at standard endoscopy as a poorly demarcated mucosal area with a slight, unusual redness.2,23 In contrast, magnifying endoscopy reveals differentiated carcinoma as a well-demarcated area within which the normal noncancerous regular capillary network is replaced by microvessels of irregular size, shape, and distribution. In clinical practice, a biopsy specimen should be taken for histopathologic diagnosis when a lesion with these findings is demonstrated by magnifying endoscopy. In addition, when the diagnosis of differentiated carciVOLUME 56, NO. 2, 2002
Intramucosal gastric cancer: microvascular architecture
noma has been established it may be helpful to determine the exact extension of the malignancy by using magnifying endoscopy to detect the demarcation line between the microvascular findings of normal and carcinomatous tissue. If absence of the magnifying endoscopic findings associated with carcinoma is demonstrated in subsequent studies to have a high negative predictive value for malignancy, it may be possible to avoid taking biopsy specimens unnecessarily from focally reddened flat mucosal lesions. Nevertheless, because the present study was uncontrolled and the number of cases included was small, a further prospective, blinded study with multiple endoscopists and a sufficient number of patients with and without carcinoma is needed to determine the sensitivity, specificity, and accuracy of magnifying endoscopic findings in clinical practice. For intramucosal undifferentiated carcinoma, magnifying endoscopy is not specific because it simply shows an illdefined area with a reduction in or disappearance of the normal mucosal capillary network of noncancerous mucosa. All undifferentiated carcinomas without ulceration in the present study exhibited a pale color on nonmagnifying endoscopy. Therefore, a forceps biopsy specimen should be obtained from all pale flat or depressed areas to differentiate malignant and nonmalignant lesions. It has been hypothesized that magnified endoscopy findings may reflect the histologic architecture of differentiated carcinoma, which is usually accompanied by interstitial tissue including microvessels, and which expands, displacing the surrounding noncancerous tissue.5,6 Conversely, however, the magnified endoscopic findings of intramucosal undifferentiated carcinoma (i.e., disappearance of or reduced density of subepithelial mucosal capillaries seen in noncancerous mucosa without proliferation of microvessels) may reflect the histologic findings of sparse infiltration without proliferation of interstitial tissue within the gastric mucosa.5,6 The reason(s) why the proliferation of interstitial tissue in intramucosal gastric carcinoma differs between differentiated and undifferentiated forms is unknown. In conclusion, magnified observation of the microvascular architecture of intramucosal gastric carcinoma may be useful for characterizing flat carcinomas that exhibit only subtle changes in color and shape at standard endoscopy. Magnifying endoscopy may also be useful for determining the extent of intramucosal spread of differentiated carcinomas that have an irregular margin. It may also provide a more precise endoscopic diagnosis because it offers more information regarding the quality and quantity of the color change that reflects the microvascular architecture. VOLUME 56, NO. 2, 2002
K Yao, T Oishi, T Matsui, et al.
A
B Figure 6. Photomicrographs of histologic sections corresponding to magnified endoscopic findings presented in Figure 5. A, Histologic section corresponding to Figure 5A. The signet-ring cell, poorly differentiated adenocarcinoma occupied almost the entire thickness of the lamina propria without any proliferation of stromal tissue (black arrows) (H&E, orig. mag. 50). B, Histologic section corresponding to Figure 5B. At the margin of the carcinomatous mucosa, individual malignant cells have infiltrated sparsely into the middle of the lamina propria of the noncancerous mucosa (black arrow) (H&E, orig. mag. 50). ACKNOWLEDGEMENTS We wish to thank H. Yabe (Olympus, Tokyo, Japan) for his critical discussion and specialized technical support of the present study, Dr. R. J. Schlemper (Department of Internal Medicine, Fukuoka University) for his critical discussion and Miss K. Miller (Royal English Language Centre, Fukuoka, Japan) for assistance in editing the manuscript. REFERENCES 1. Fujiwara T, Hiraoka H, Yao T, Koga Y, Okabe H. Observations on diagnostic problems in depressed type early gastric cancer [in Japanese with English abstract]. Stomach Intestine 1971; 6:157-74. 2. Yao T, Fujiwara T, Watanabe H, Koga Y, Okada Y, Okabe H. Endoscopic diagnosis of the extent of infiltration in gastric cancer [in Japanese with English abstract]. Stomach Intestine 1972;7:725-38. GASTROINTESTINAL ENDOSCOPY
283
E Miller, J Kim, J Gondehak, et al.
3. Yao K, Yao T, Matsui T, Iwashita A, Oishi T. Hemoglobin content in intramucosal gastric carcinoma as a marker of histologic differentiation: a clinical application of quantitative electronic endoscopy. Gastrointest Endosc 2000;52:241-5. 4. Yao K, Iwashita A, Yao T, Furukawa H, Furukawa K, Takemura S et al. Quantitative method to characterize electronic endoscopic images of early gastric carcinomas, using index of hemoglobin [in Japanese with English Abstract]. Gastroenterol Endosc 1997;39:2253-63. 5. Adachi Y, Mori M, Enjoji M, Sugimachi K. Microvascular architecture of early gastric carcinoma. Microvascularhistopathologic correlates. Cancer 1993;72:32-6. 6. Furukawa K, Yao K, Iwashita A, Matsui T, Yao T. Microvascular architecture of depressed type early gastric cancer. Quantified vascular density using computer analysis with special reference to the color of endoscopic findings [in Japanese with English abstract]. Gastroenterol Endosc 1997; 39:1358-69. 7. Yao K, Oishi T. Microgastroscopic findings of mucosal microvascular architecture as visualized by magnifying endoscopy. Dig Endosc 2001;13:527-33. 8. Yagi K. Endoscopic features and magnified views of the corpus in the Helicobacter pylori-negative stomach. Dig Endosc 2001;13:S34-5. 9. Udagawa T, Amano M, Okada F. Development of magnifying video endoscopes with high resolution. Dig Endosc 2001;13: 163-9. 10. Sakaki N. The resolving power of electronic endoscopy [Japanese]. Clin Gastroenterol 1989;4:677-84. 11. Gannon B. The vasculature and lymphatic drainage. Whitehead R ed. Gastrointestinal and oesophageal pathology. Edinburgh: Churchill Livingstone; 1995. p. 129-99. 12. Nishi M, Omoni Y, Miwa K, editors. Japanese classification of gastric carcinoma. First English ed. Tokyo: Kanehara Co; 1995. p. 4-5. 13. Schlemper RJ, Riddle RH, Kato Y, Borchard F, Cooper HS,
284
GASTROINTESTINAL ENDOSCOPY
A new device for measuring esophageal variceal pressure
14.
15. 16.
17.
18.
19.
20.
21.
22.
23.
Dawsey SM, et al. The Vienna classification of gastrointestinal epithelial neoplasia. Gut 2000:47;251-5. Lauren P. The two histologic main types of gastric carcinoma: diffuse and so-called intestinal-type carcinoma: an attempt at a histo-clinical classification. Acta Pathol Microbiol Scand 1965;64:31-49. Tada M, Murakami A, Karita M, Okita K. Endoscopic resection of early gastric cancer. Endoscopy 1993;25:445-50. Sakaki N, Iida Y, Saito M, Tada M, Odawara M, Okazaki Y, et al. New magnifying endoscopic classification of the fine gastric mucosa pattern [in Japanese with English abstract]. Gastroenterol Endosc 1980;22:377-83. Tobita K. Study on minute surface structures of the depressed type early gastric cancer with magnifying endoscopy. Dig Endosc 2001;13:121-6. Kudo S, Tamura S, Nakajima T, Yamano H, Kusaka H, Watanabe H. Diagnosis of colorectal tumorous lesions by magnifying endoscopy. Gastrointest Endosc 1996;44:8-14. Guelrud M, Herrera I, Essenfeld H, Castro J. Enhanced magnification endoscopy: a new technique to identify specialized intestinal metaplasia in Barrett’s esophagus. Gastrointest Endosc 2001;53:559-65. Adachi Y, Shiraishi N, Shiromizu A, Bandoh T, Aramaki M, Kitano S. Laparoscopy-assisted Biliroth I gastrectomy compared with conventional open gastrectomy. Arch Surg 2000; 135:806-10. Nakamura K, Ueyama T, Yao T, Xuan ZX, Ambe K, Adachi Y, et al. Pathology and prognosis of gastric carcinoma: findings in 10,000 patients who underwent primary gastrectomy. Cancer 1992;70:1030-7. Adachi Y, Yasuda K, Inomata M, Sato K, Shiraishi N, Kitano S. Pathology and prognosis of gastric carcinoma. Cancer 2000: 89;1418-24. Honmyo U, Mitsui A, Murakami A, Mizumoto S, Yoshinaka I, Maeda M, et al. Mechanisms producing color change in flat early gastric cancers. Endoscopy 1997;29:366-71.
VOLUME 56, NO. 2, 2002
Color change is a useful marker for the endoscopic identification of early stage gastric cancer and for determination of the spread of carcinoma.1,2 Our clinical study demonstrated that this color change in early stage gastric cancer depended mainly on mucosal hemoglobin content and that this was increased in differentiated intramucosal gastric carcinoma and decreased in undifferentiated intramucosal gastric cancer.3 In addition, the pattern of distribution in early stage gastric carcinoma was found to be more irregular in differentiated carcinomatous Received October 1, 2001. For revision January 4, 2002. Accepted March 7, 2002. Current affiliations: Departments of Gastroenterology and Pathology, Fukuoka University, Chikushi Hospital, Fukuoka, Japan. Presented in part at the annual Digestive Disease Week, May 1923, 2002, San Francisco, California (Gastrointest Endosc 2002;55:AB128). Reprint requests: Kenshi Yao, MD, PhD, Fukuoka University, Chikushi Hospital, 377-1 Ooaza Zokumyoin, Chikushino-city, Fukuoka Prefecture, Japan 818-8502. Copyright © 2002 by the American Society for Gastrointestinal Endoscopy 0016-5107/2002/$35.00 + 0 37/69/126061 doi:10.1067/mge.2002.126061 VOLUME 56, NO. 2, 2002
mucosa than in noncancerous background mucosa.4 Furthermore, it has been well demonstrated in pathologic studies that the change in color evident at endoscopy (i.e., mucosal hemoglobin content) probably originates from differences in the density and architecture of the microvasculature structure between cancerous and noncancerous mucosa.5,6 However, the significance of these changes in microvascular architecture has not been clarified in clinical investigations. In preliminary studies, a new magnifying upper endoscopy system clearly visualized a subepithelial microvascular structure or loop in the gastric mucosa.7,8 In the present study, the magnifying endoscopic appearance of the microvascular architecture was determined for both cancerous and noncancerous mucosa in vivo. PATIENTS AND METHODS Instrument specifications and endoscopic procedure A new magnifying endoscope (GIF-Q240Z, Olympus Optical Co., Ltd., Tokyo, Japan) for use in upper GI tract with an electronic endoscope system (EVIS 240, Olympus) was used in this study. The diameter and length of the insertion tube of this instrument are the same as those of a standard upper endoscope (e.g., GIF-Q240, Olympus). The new instrument can magnify the endoscopic image optically up to 80-fold through use of a zoom attachment. Before the examination, a food particle was mounted on the tip of the endoscope to enable the endoscopist to fix the focal distance at 3 mm between the tip of the instrument and the mucosal surface, the point at which maximal magnification of the endoscopic image was to be obtained. Preparation of patients for the magnifying examination was the same as that for standard upper endoscopy. For magnifying observation, the tip of the endoscope was allowed to touch the mucosa immediately after zooming up to the maximal magnifying level, and then a magnified endoscopic image in focus could be easily obtained. Maximal resolution of the magnifying endoscope in vitro The maximal resolution of the new magnifying system was tested under laboratory conditions9,10 to determine whether it provided adequate resolution to display the microvascular structure of the gastric mucosa, as described in another in vitro study.11 For this purpose, a resolution chart (Olympus) was used on which black lines of various known widths and lengths are displayed. The chart was placed on a stage arranged perpendicular to the axis of the insertion tube of the endoscope. At maximal magnification, the minimal width ( m) of the black lines that could be separated visually on the endoscopic monitor was determined and taken as the maximal resolution of the magnifying system.9 Patients and histopathologic investigations Twenty-seven consecutive patients with intramucosal gastric carcinomas who underwent magnifying endoscopy between May 2000 and August 2001 were included in this GASTROINTESTINAL ENDOSCOPY
279
K Yao, T Oishi, T Matsui, et al.
Figure 1. Nonmagnified endoscopic view of differentiated superficial depressed type carcinoma. An ill-defined area with faint, unusual redness is shown on the greater curvature of the antrum. uncontrolled, prospective cohort study. The protocol was approved by the Human Subject Committee of our university, and informed consent was obtained from all patients. In all patients, the gastric carcinoma had been diagnosed previously at conventional endoscopy including biopsies with histopathologic confirmation. Thus, the general endoscopic findings including macroscopic type and location of the lesion were known before magnifying endoscopy. Macroscopic type of the carcinoma was classified according to the classification for early stage gastric cancer of the Japanese Research Society of Gastric Cancer (Type I, protruded; Type IIa, superficial elevated; Type IIb, flat; Type IIc, superficial depressed; and Type III, excavated).12 Standard (nonmagnifying) and subsequent magnifying endoscopies were performed preoperatively to define the margin between carcinoma and noncarcinoma, and to determine depth of invasion. Standard (nonmagnifying) endoscopy was performed first to identify abnormal mucosal areas including reddened, slightly depressed or flat mucosa with irregular margin, flat focal lesions with ill-defined margins accompanied by abnormal redness, or focally pale mucosa.2 Immediately after taking endoscopic photographs of a non-magnified, abnormal mucosal lesion, the appearance of which raised a strong suspicion for carcinoma, the mucosal lesion and the surrounding mucosa were observed at maximal magnification with the zoom attachment of the magnifying endoscope. Once the magnified image was in focus, magnified endoscopic photographs were taken. If the observed area did not exhibit findings on nonmagnified endoscopy typical of carcinoma, a biopsy specimen was taken from the area at which the magnified endoscopic image was obtained. The following findings were recorded at standard endoscopy: macroscopic classification, endoscopic color (i.e., reddened, almost the same as the surrounding mucosa, or pale) and location of the carcinoma. The following magnifying endoscopic findings were recorded: microvascular architecture in both the cancerous and surrounding mucosa. One investigator (K.Y.) performed all of the magnifying endoscopic examinations and recorded all of the endoscopic findings. 280
GASTROINTESTINAL ENDOSCOPY
Intramucosal gastric cancer: microvascular architecture
After the carcinomas had been resected either endoscopically or surgically, the resected specimens were extended on boards with pins and fixed in 20% formalin. The cancerous lesions together with the surrounding noncancerous mucosa were cut into 2- to 5-mm wide serial step sections, at least one of which corresponded to the portion of the carcinomatous and noncarcinomatous mucosa in the magnified endoscopic images, as determined in discussions with a pathologist (T.O.). The histopathologic sections were interpreted by an independent pathologist (A.K.) blinded to the endoscopic findings. The histopathologic diagnostic criteria were based on the Vienna classification of GI epithelial neoplasia (Category 1, negative for neoplasia/dysplasia; Category 2, indefinite for neoplasia/dysplasia; Category 3, noninvasive low-grade neoplasia; Category 4, noninvasive highgrade dysplasia; Category 5, invasive neoplasia [5.1, intramucosal carcinoma; 5.2. submucosal carcinoma]).13 The carcinomas were further classified into 2 major differentiation groups, i.e., differentiated carcinoma which had tubular or papillary structures resembling metaplastic intestinal epithelium, and undifferentiated carcinoma, including poorly differentiated carcinoma, which was characterized by discretely proliferating tumor cells with sparse or no tubular formation.14 In addition, the sections were evaluated histologically for ulceration (i.e., regenerative epithelium or submucosal fibrosis). Histopathologic findings refer to the section that corresponded to the magnified endoscopic view. Carcinomas of the superficial depressed or flat types as determined at standard endoscopy were included in this study. Protruded and superficial elevated types were excluded as these are easily identified by standard endoscopy alone. Patients were excluded if the endoscopic or histopathologic findings included ulceration within the lesion, or submucosal invasion, because both of these findings may influence the microvascular architecture of the lesion.4
RESULTS Resolution of magnifying endoscope At maximal magnification, the smallest line width separable on the resolution chart was 7.874 m; lines 7.016 m in width could not be separated. Accordingly, the resolution power of this magnifying endoscope was determined to be 7.9 m. Nonmagnified endoscopic and histopathologic findings With nonmagnified endoscopy, all lesions had a slightly depressed or flat mucosal appearance with poor demarcation. Twenty-three carcinomas were of the superficial depressed type and 4 were flat type. On nonmagnified observation the color of the lesions was as follows: reddened, 18; pale, 8; and same color as surrounding mucosa, 1. Eighteen carcinomas were located in the gastric body; 9 were in the gastric antrum. Eleven carcinomas were resected by using VOLUME 56, NO. 2, 2002
Intramucosal gastric cancer: microvascular architecture
K Yao, T Oishi, T Matsui, et al.
A Figure 3. Photomicrograph of section corresponding to magnified view of marginal area between carcinoma and noncarcinomatous mucosa as shown in Figure 2C (H&E, orig. mag. 13). The margin (black arrow) is distinct. The cancerous tissue (area to left of arrow) consists of atypical glands accompanied by irregular interstitial tissue. The cancerous tissue has invaded expansively within the mucosa.
B
the endoscopic mucosal resection (EMR) technique15 and 16 were resected surgically. The median diameter of the differentiated and undifferentiated carcinomas was, respectively, 14 mm (range, 5-45 mm) and 25 mm (range, 12-35 mm). Histopathologic evaluation disclosed that all 27 carcinomas had invaded the lamina propria but not the submucosa (i.e., Category 5.1, intramucosal carcinoma, Vienna classification).13 There were 19 differentiated and 8 undifferentiated carcinomas. In comparing the endoscopic color to histologic type, 18 of the differentiated carcinomas were reddened and 1 was almost the same color as the surrounding mucosa, whereas all 8 undifferentiated carcinomas were pale. Mucosal microvascular findings by magnifying endoscopy
C Figure 2. Magnified endoscopic views of area indicated by white arrow in Figure 1. A, Magnified view of noncancerous mucosa showing subepithelial mesh-like capillary network whose shape and distribution exhibits a regular pattern. B, Magnified view of cancerous mucosa showing tortuous, irregularly dilated microvessels of various shapes such as branched (yellow arrow), loop-like (blue arrow), or ring-like (green arrow). The pattern is irregular compared with that for the capillaries in the noncancerous mucosa. C, Magnified view of border between cancerous and noncancerous mucosa showing well-demarcated line (white arrow).
VOLUME 56, NO. 2, 2002
Magnification findings were successfully obtained of both the noncancerous and cancerous mucosa and compared with histopathologic findings in all cases (Figs. 1-6). Noncancerous mucosa manifested a subepithelial capillary network or capillary loops in regular arrangement (Fig. 2A). Magnified views of the surface microvascular structure of the carcinomatous mucosa differed depending on histopathologic differentiation. All 19 differentiated carcinomas had a well-demarcated area within which the regular subepithelial capillary network of the surrounding noncancerous mucosa was replaced by microvessels of irregular size, shape and distribution (Fig. 3). The histologic findings corresponding to this magnified view of differentiated carcinoma included carcinomatous tissue composed of atypical glands in a tubular or papillary pattern with irregGASTROINTESTINAL ENDOSCOPY
281
K Yao, T Oishi, T Matsui, et al.
Intramucosal gastric cancer: microvascular architecture
A Figure 4. Nonmagnified endoscopic view of undifferentiated superficial depressed type carcinoma. An ill-defined area (arrow) of pale mucosa is present in the posterior wall of the gastric body.
ular interstitial tissue that had expanded to form an abrupt transition with noncancerous tissue (Fig. 3). In contrast, the magnified microvascular findings in the 9 undifferentiated carcinomas consisted of an ill-defined area within which there was either disappearance or a decrease in the density of the normal regular subepithelial capillary network of the surrounding noncancerous mucosa, but proliferation of irregularly shaped vessels was not observed (Fig. 5). Histopathologically, carcinomatous cells had infiltrated discretely and sparsely, without proliferation of interstitial tissue, within the lamina propria of the noncancerous mucosa (Fig. 6). DISCUSSION Previously developed magnifying endoscopy systems do not have adequate resolution for visualization of microvessels, including capillaries.16,17 The resolution of the endoscope used in the present study was 7.9 m. Because the minimum diameter of subepithelial capillaries in the gastric mucosa is found in anatomic studies to be 8 m,11 the magnifying endoscope used in the study can visualize microvessels and capillaries that are 8 m or greater in diameter. This improved resolution made it possible to observe the characteristics of microvessels in carcinomatous and noncarcinomatous mucosa for the first time in vivo. The preparation of patients for magnifying endoscopy and the procedure itself are the same as for standard upper endoscopy. For magnifying observation of mucosal microvessels, no special techniques, such as vital mucosal staining18 or acid loading,19 are needed. The method described is both simple and noninvasive. Magnifying endoscopy of early stage gastric carcinoma may be useful for identifying intramucosal carcinomas that can be cured by EMR or minimally 282
GASTROINTESTINAL ENDOSCOPY
B Figure 5. Magnified views of pale area indicated by white arrow in Figure 4. A, Magnified view of pale cancerous mucosa showing disappearance of mucosal capillary network without any proliferation of vessels (white arrows). B, Magnified endoscopic view of marginal area between carcinomatous tissue and noncarcinomatous mucosa showing gradual reduction in vascularity between subepithelial noncancerous capillary network and cancerous mucosa (white arrow).
invasive surgery, thereby offering the patient a prognosis comparable with that achievable by conventional surgery.20-22 The endoscopic diagnosis of intramucosal gastric carcinoma of the superficial depressed type or flat type with nonmagnified instruments is often difficult because such carcinomas, so-called “gastritis-like cancer,” are manifest as only subtle changes in color and shape.23 Differentiated carcinoma is often described at standard endoscopy as a poorly demarcated mucosal area with a slight, unusual redness.2,23 In contrast, magnifying endoscopy reveals differentiated carcinoma as a well-demarcated area within which the normal noncancerous regular capillary network is replaced by microvessels of irregular size, shape, and distribution. In clinical practice, a biopsy specimen should be taken for histopathologic diagnosis when a lesion with these findings is demonstrated by magnifying endoscopy. In addition, when the diagnosis of differentiated carciVOLUME 56, NO. 2, 2002
Intramucosal gastric cancer: microvascular architecture
noma has been established it may be helpful to determine the exact extension of the malignancy by using magnifying endoscopy to detect the demarcation line between the microvascular findings of normal and carcinomatous tissue. If absence of the magnifying endoscopic findings associated with carcinoma is demonstrated in subsequent studies to have a high negative predictive value for malignancy, it may be possible to avoid taking biopsy specimens unnecessarily from focally reddened flat mucosal lesions. Nevertheless, because the present study was uncontrolled and the number of cases included was small, a further prospective, blinded study with multiple endoscopists and a sufficient number of patients with and without carcinoma is needed to determine the sensitivity, specificity, and accuracy of magnifying endoscopic findings in clinical practice. For intramucosal undifferentiated carcinoma, magnifying endoscopy is not specific because it simply shows an illdefined area with a reduction in or disappearance of the normal mucosal capillary network of noncancerous mucosa. All undifferentiated carcinomas without ulceration in the present study exhibited a pale color on nonmagnifying endoscopy. Therefore, a forceps biopsy specimen should be obtained from all pale flat or depressed areas to differentiate malignant and nonmalignant lesions. It has been hypothesized that magnified endoscopy findings may reflect the histologic architecture of differentiated carcinoma, which is usually accompanied by interstitial tissue including microvessels, and which expands, displacing the surrounding noncancerous tissue.5,6 Conversely, however, the magnified endoscopic findings of intramucosal undifferentiated carcinoma (i.e., disappearance of or reduced density of subepithelial mucosal capillaries seen in noncancerous mucosa without proliferation of microvessels) may reflect the histologic findings of sparse infiltration without proliferation of interstitial tissue within the gastric mucosa.5,6 The reason(s) why the proliferation of interstitial tissue in intramucosal gastric carcinoma differs between differentiated and undifferentiated forms is unknown. In conclusion, magnified observation of the microvascular architecture of intramucosal gastric carcinoma may be useful for characterizing flat carcinomas that exhibit only subtle changes in color and shape at standard endoscopy. Magnifying endoscopy may also be useful for determining the extent of intramucosal spread of differentiated carcinomas that have an irregular margin. It may also provide a more precise endoscopic diagnosis because it offers more information regarding the quality and quantity of the color change that reflects the microvascular architecture. VOLUME 56, NO. 2, 2002
K Yao, T Oishi, T Matsui, et al.
A
B Figure 6. Photomicrographs of histologic sections corresponding to magnified endoscopic findings presented in Figure 5. A, Histologic section corresponding to Figure 5A. The signet-ring cell, poorly differentiated adenocarcinoma occupied almost the entire thickness of the lamina propria without any proliferation of stromal tissue (black arrows) (H&E, orig. mag. 50). B, Histologic section corresponding to Figure 5B. At the margin of the carcinomatous mucosa, individual malignant cells have infiltrated sparsely into the middle of the lamina propria of the noncancerous mucosa (black arrow) (H&E, orig. mag. 50). ACKNOWLEDGEMENTS We wish to thank H. Yabe (Olympus, Tokyo, Japan) for his critical discussion and specialized technical support of the present study, Dr. R. J. Schlemper (Department of Internal Medicine, Fukuoka University) for his critical discussion and Miss K. Miller (Royal English Language Centre, Fukuoka, Japan) for assistance in editing the manuscript. REFERENCES 1. Fujiwara T, Hiraoka H, Yao T, Koga Y, Okabe H. Observations on diagnostic problems in depressed type early gastric cancer [in Japanese with English abstract]. Stomach Intestine 1971; 6:157-74. 2. Yao T, Fujiwara T, Watanabe H, Koga Y, Okada Y, Okabe H. Endoscopic diagnosis of the extent of infiltration in gastric cancer [in Japanese with English abstract]. Stomach Intestine 1972;7:725-38. GASTROINTESTINAL ENDOSCOPY
283
E Miller, J Kim, J Gondehak, et al.
3. Yao K, Yao T, Matsui T, Iwashita A, Oishi T. Hemoglobin content in intramucosal gastric carcinoma as a marker of histologic differentiation: a clinical application of quantitative electronic endoscopy. Gastrointest Endosc 2000;52:241-5. 4. Yao K, Iwashita A, Yao T, Furukawa H, Furukawa K, Takemura S et al. Quantitative method to characterize electronic endoscopic images of early gastric carcinomas, using index of hemoglobin [in Japanese with English Abstract]. Gastroenterol Endosc 1997;39:2253-63. 5. Adachi Y, Mori M, Enjoji M, Sugimachi K. Microvascular architecture of early gastric carcinoma. Microvascularhistopathologic correlates. Cancer 1993;72:32-6. 6. Furukawa K, Yao K, Iwashita A, Matsui T, Yao T. Microvascular architecture of depressed type early gastric cancer. Quantified vascular density using computer analysis with special reference to the color of endoscopic findings [in Japanese with English abstract]. Gastroenterol Endosc 1997; 39:1358-69. 7. Yao K, Oishi T. Microgastroscopic findings of mucosal microvascular architecture as visualized by magnifying endoscopy. Dig Endosc 2001;13:527-33. 8. Yagi K. Endoscopic features and magnified views of the corpus in the Helicobacter pylori-negative stomach. Dig Endosc 2001;13:S34-5. 9. Udagawa T, Amano M, Okada F. Development of magnifying video endoscopes with high resolution. Dig Endosc 2001;13: 163-9. 10. Sakaki N. The resolving power of electronic endoscopy [Japanese]. Clin Gastroenterol 1989;4:677-84. 11. Gannon B. The vasculature and lymphatic drainage. Whitehead R ed. Gastrointestinal and oesophageal pathology. Edinburgh: Churchill Livingstone; 1995. p. 129-99. 12. Nishi M, Omoni Y, Miwa K, editors. Japanese classification of gastric carcinoma. First English ed. Tokyo: Kanehara Co; 1995. p. 4-5. 13. Schlemper RJ, Riddle RH, Kato Y, Borchard F, Cooper HS,
284
GASTROINTESTINAL ENDOSCOPY
A new device for measuring esophageal variceal pressure
14.
15. 16.
17.
18.
19.
20.
21.
22.
23.
Dawsey SM, et al. The Vienna classification of gastrointestinal epithelial neoplasia. Gut 2000:47;251-5. Lauren P. The two histologic main types of gastric carcinoma: diffuse and so-called intestinal-type carcinoma: an attempt at a histo-clinical classification. Acta Pathol Microbiol Scand 1965;64:31-49. Tada M, Murakami A, Karita M, Okita K. Endoscopic resection of early gastric cancer. Endoscopy 1993;25:445-50. Sakaki N, Iida Y, Saito M, Tada M, Odawara M, Okazaki Y, et al. New magnifying endoscopic classification of the fine gastric mucosa pattern [in Japanese with English abstract]. Gastroenterol Endosc 1980;22:377-83. Tobita K. Study on minute surface structures of the depressed type early gastric cancer with magnifying endoscopy. Dig Endosc 2001;13:121-6. Kudo S, Tamura S, Nakajima T, Yamano H, Kusaka H, Watanabe H. Diagnosis of colorectal tumorous lesions by magnifying endoscopy. Gastrointest Endosc 1996;44:8-14. Guelrud M, Herrera I, Essenfeld H, Castro J. Enhanced magnification endoscopy: a new technique to identify specialized intestinal metaplasia in Barrett’s esophagus. Gastrointest Endosc 2001;53:559-65. Adachi Y, Shiraishi N, Shiromizu A, Bandoh T, Aramaki M, Kitano S. Laparoscopy-assisted Biliroth I gastrectomy compared with conventional open gastrectomy. Arch Surg 2000; 135:806-10. Nakamura K, Ueyama T, Yao T, Xuan ZX, Ambe K, Adachi Y, et al. Pathology and prognosis of gastric carcinoma: findings in 10,000 patients who underwent primary gastrectomy. Cancer 1992;70:1030-7. Adachi Y, Yasuda K, Inomata M, Sato K, Shiraishi N, Kitano S. Pathology and prognosis of gastric carcinoma. Cancer 2000: 89;1418-24. Honmyo U, Mitsui A, Murakami A, Mizumoto S, Yoshinaka I, Maeda M, et al. Mechanisms producing color change in flat early gastric cancers. Endoscopy 1997;29:366-71.
VOLUME 56, NO. 2, 2002