Magnification Endoscopy: Does It Improve Mucosal Surface Analysis for the Diagnosis of Gastrointestinal Neoplasias?

The prognosis for patients with gastrointestinal malignancies is strictly dependent on early detection of premalignant and malignant lesions. In Western countries detection rates are low for early carcinomas; in Germany, in fact, only 5 % of all gastric cancers are diagnosed at an early stage [1]. Early cancers and adenomatous lesions can be removed endoscopically with various techniques (e. g. polypectomy, mucosal resection). The patient is cured by complete resection of such lesions. It is important to detect flat and depressed neoplasias, not only the more common polypoid type of malignant and adenomatous lesions. Hence, the goal of every routine endoscopy is detection, and if possible resection, of early cancers and premalignant lesions.

Radiological and endoscopic screening methods already have a long tradition in Japan, because of the higher incidence and prevalence of gastric cancer. What should an ideal screening endoscopy be able to do? The technique should allow detection of discreet mucosal alterations. Ideally, endoscopic discrimination between neoplastic and non-neoplastic lesions would be possible. The endoscopic characterization of mucosal surface detail might yield important clinical and economic advantages.

Newly developed high-resolution and magnification endoscopes offer image quality which is significantly better than that of first-generation video endoscopes or the older fiberoptic systems. The resolution of an endoscopic image is a different quality from the magnification, and is defined as the ability to distinguish between two points that are close together. High-resolution imaging improves the ability to discriminate detail while magnification enlarges the image. In digital video imaging, resolution is a function of pixel density. By incorporating high-pixel density charged-coupled devices (CCD), high-resolution endoscopes provide slightly magnified views of the gastrointestinal tract with greater mucosal detail. Magnification endoscopy utilizes a movable lens controlled by the endoscopist to vary the degree of magnification, which ranges from × 1.5 to × 150. Newly designed magnification endoscopes provide high-resolution and magnification features [2]. So far, the clinical value of such detailed mucosal surface evaluation has not been proven.

The role of the modern endoscopist moves towards that of the pathologist because of the optical power of the new endoscopic systems. The wealth of imaging detail now available accelerates the need for standardization of mucosal surface classification in various conditions. The question arises as to which surface characteristics allow the best differentiation between neoplastic and non-neoplastic tissue. Should we focus on glandular openings (pits), surface vascular architecture, the macroscopic type of a lesion, or on all of these features?

Surface Vascular Architecture

The surface capillary structure of neoplasias can be identified by using magnifying endoscopes. Kumagai et al. [3] examined the appearance of intrapapillary capillary loops (IPCL) in patients with superficial esophageal cancer. Depending on depth of invasion (m1 - m3, sm), characteristic changes in IPCL were seen, by means of stereoscopic microscopy as well as magnifying endoscopy. In m1 invasion, there was dilatation of IPCL; in m2 invasion, there was dilatation and elongation; in m3 invasion there was a mixed appearance, with ICPL and tumor vessels; and in sm invasion, there was complete replacement by tumor vessels.

In the current issue of Endoscopy, Tajiri et al. [4] present magnification endoscopy as a screening tool for the diagnosis of early gastric cancers. A newly developed magnifying endoscope was used which enables a magnification of up to × 80. A total of 318 patients were examined and 211 lesions were finally included in the study. Their findings for surface vascular architecture were as follows. In elevated-type early cancer with tubular adenocarcinoma, the capillary vessels observed by magnifying endoscopy were irregular, relatively thick and short compared with those observed in hyperplastic polyps or adenomas. In elevated-type early cancer with papillary adenocarcinoma, the authors found long, tortuous capillary vessels. The abnormal capillary vessels of depressed-type early cancer could be observed by magnifying endoscopy. Changes in vascular architecture are an early sign of malignant transformation and can be made visible with magnification endoscopes, even without the use of chromoendoscopy. Prediction of neoplasia based only on changes in vascular architecture is not sufficient, since inflammatory changes can mimic similar capillary alterations.

Tajiri et al. [4] described only slightly dilated and tortuous capillary vessels in regenerated epithelium. The density of the vessels increased but no irregularities could be observed, in contrast to neoplastic lesions. Yagi et al. [5], on the other hand compared the vascular architecture in patients with normal gastric mucosa and in those with Helicobacter-induced gastritis. The anterior wall or greater curvature of the middle body of the stomach was observed by magnifying endoscopy. Patients with normal gastric mucosa showed collecting venules with true capillaries forming a network, and gastric pits resembling pinholes. In contrast to the observation of Tajiri et al., patients with inflammation showed irregular capillaries.

Irregularities in vascular architecture can obviously be signs of early malignant change, but may also be identified in inflamed tissue. Anti-inflammatory therapy prior to magnification endoscopy is recommended, and may help to increase the detection rate for neoplastic alterations and decrease the rate of false-positive findings.

Pit Pattern Structure

In addition to vascular changes the arrangement of surface glandular openings (pit pattern structure) is an important parameter for an endoscopic prediction of neoplastic or non-neoplastic change. Chromoendoscopy, used in conjunction with powerful endoscopes, is an ideal tool for analyzing pit pattern. The macroscopic type and exact margins of a suspicious lesion can be easily recognized. The pit pattern structure becomes immediately visible after dye spraying and the surface mucosal structure can be examined meticulously. Indigo carmine dye (0.2 - 0.4 %) is commonly used for this. It is a contrast stain and is not absorbed. The blue color fills the bottom of shallow depressed areas and facilitates recognition of depressed early cancers [6].

High-magnification and high-resolution endoscopes were designed to be used in conjunction with chromoendoscopy. Intravital staining helps to unmask and to identify the real extent of small lesions. After dye spraying, the surface structure is analyzed in the magnification mode. The combination of both techniques has led to a standard classification of the surface staining pattern of colorectal lesions. The so-called pit pattern classification [7] categorizes five subtypes, with types I and II being predictors of non-neoplastic mucosal changes, and types III - V predicting neoplastic mucosal changes. With the use of the pit pattern classification, it is possible to distinguish between neoplastic and non-neoplastic tissue with a sensitivity and specificity of 92 - 98 % and 61 - 95 %, respectively [8] [9] [10]. The pit pattern classification was primarily developed for magnification endoscopy, but sufficient surface detail is visible with high-resolution imaging alone [10].

We have recently, investigated the value of conventional colonoscopy, chromoendoscopy, and magnification endoscopy combined with chromoendoscopy. Each one of 1000 patients was randomly allocated to examination by these methods [11]. With the use of chromoendoscopy and magnification endoscopy, significantly more lesions were detected compared with conventional endoscopy and there was a higher rate of detection of adenomas. In addition more dysplastic adenomas (independent of type) were observed. On the basis of pit pattern, the sensitivity and specificity for discrimination between nonadenomatous and adenomatous tissue were 84 % and 68 % for chromoendoscopy, and 98 % and 74 % for magnification endoscopy. Chromoendoscopy therefore facilitates early detection of colorectal adenomas. The surface staining pattern allows targeted biopsies, and a greater number of dysplastic adenomas of smaller size can be detected in comparison with conventional endoscopy. In addition, magnification endoscopy improves accuracy in distinguishing between adenomatous and nonadenomatous changes.

With regard to changes in esophageal mucosa, the newly developed endoscopes also provide significant improvements regarding early diagnosis. Barrett’s esophagus is a precancerous condition, for which specialized columnar epithelium (SCE) is pathognomonic and often distributed focally within the columnar-lined lower esophagus. So far, endoscopic discrimination between different types of epithelium within the columnar-lined lower esophagus (e. g. gastral or junctional epithelium) has not been possible. Chromoendoscopy and magnification endoscopy help to overcome this problem. Methylene blue selectively stains specialized columnar epithelium, as shown by Canto et al. [12] in 1996. In comparison with random biopsies, methylene blue-directed biopsies more often provide the diagnosis of Barrett’s esophagus, even when fewer samples are taken [13]. These early studies did not take the mucosal surface structure into account. Guelrud et al. [14] analyzed Barrett’s epithelium after instillation of acetic acid (1.5 %), using magnification endoscopes with a zoom range of up to 35. The surface mucosal pattern was categorized into four types (i. e., round, reticular, villous, and ridged pits). The yields for detecting specialized columnar epithelium were significantly improved for mucosal patterns with villous and ridged pits (87 % and 100 %). Endo et al. [15] combined intravital staining with methylene blue and magnification endoscopy (zoom × 80). They classified columnar epithelium into five types: small round, straight, long oval, tubular, and villous pits. Tubular and villous pit patterns were predictive of specialized columnar epithelium in 100 % of cases. These pit patterns stained with methylene blue significantly more often than others. Interestingly the staining rates were 50 % and 60 %, which was considerably lower than expected. Tubular and villous pits were associated with the histological presence of intestinal-mucin type, which corresponds to the histological nature of Barrett’s epithelium (incomplete intestinal metaplasia). Small round and straight pits were predictive of gastric metaplasia and did not conform to a formal diagnosis of Barrett’s esophagus. These patterns showed no staining at all, and no specialized columnar epithelium was observed in targeted biopsies.

So far there is no standardized classification of mucosal surface patterns for identifying early gastric cancers at magnifying endoscopy. Tajiri et al. [4] analyzed mucosal surface details in 211 gastric lesions including 89 early gastric cancers (31 elevated and 58 depressed type). The elevated-type early gastric cancer showed an irregular pit pattern on magnifying endoscopy and on dissecting microscopy. All depressed-type early cancers had irregular tubular structures. The pit pattern architecture was finer than in the surrounding mucosa. The sensitivity and specificity of magnification endoscopy for diagnosis of gastric neoplasia were 96 % and 95.5 %, respectively.

These criteria were evaluated retrospectively after endoscopic and histological examination. Therefore, they are only of empirical value so far. Thus, there is a urgent need for studies with prospective defined endoscopic criteria for gastric malignancies.

Macroscopic Type

The accurate determination of the macroscopic type of a lesion influences endoscopic management. For an ideal endoscopic screening procedure, it is important to look for flat and depressed neoplastic changes, and not just for the more easily identifiable polypoid lesions.

Depressed early colorectal carcinomas have high malignant potential and lead to early malignant invasion even when they are still small in size [16]. For a long time these lesions were not recognized by western endoscopists. However, Fujii et al. [17] demonstrated impressively, in England also, that detection of depressed early colorectal malignancies was possible with the help of chromoendoscopy and magnification endoscopy, after a thorough bowel lavage. Knowledge of the appearance of these lesions is a prerequisite for diagnosis. First, the endoscopist must recognize the often inconspicuous lesions. After cleaning with tap water, the lesion is stained, usually with indigo carmine, which facilitates the determination of the macroscopic type (flat vs. depressed). Magnification endoscopy follows and allows an endoscopic prediction with regard to neoplasia.

In early gastric cancer, discrimination between three categories and their different subtypes is important for deciding upon the appropriate surgical or endoscopic resection. Classification into subtypes is made easier by the use of chromoendoscopy. Intravital staining techniques are very commonly employed in Japan and belong to the endoscopist’s armamentarium. Depressed areas are generally signs of a greater depth of infiltration of the lesion. Therefore, only flat early cancers up to a size of 1 cm with a shallow depression (type IIc) are removed endoscopically. Ulcerated early carcinomas should primarily be resected surgically [18] [19]. Elevated-type early gastric cancers (types I and II), up to a size of 3 cm, can be removed endoscopically. For all these subtypes, it is the case that only well to moderately differentiated adenocarcinomas limited to the mucosa should be resected endoscopically [19] [20].

These strict criteria are beginning to change since the publication of a study by Gotoda et al. [20]. They investigated 5265 patients who had undergone gastrectomy with lymph node dissection for early gastric cancer. Several clinicopathological factors were assessed for their possible association with lymph node metastasis. None of the 1230 well-differentiated intramucosal cancers of less than 30 mm diameter were associated with metastases, regardless of the presence of ulceration, and none of the 929 lesions without ulceration were associated with nodal metastases regardless of tumor size. With submucosal cancers, similarly to the findings for intramucosal cancers, there was a significant correlation between tumor size larger than 30 mm and lymphatic or vascular involvement, with an increased risk of lymph node metastases. None of the 145 differentiated adenocarcinomas of less than 30 mm diameter and without lymphatic or venous invasion were associated with lymph node metastases, as long as the lesion had invaded less than 500 µm into the submucosa.

The visual evaluation of minute detail allowed by magnification endoscopy is promising, but some points of criticism must be discussed.

Inflammation can cause significant disturbance of the image seen when magnifying endoscopy is used to look for the minute changes indicative of neoplasia, and there is a danger of false-positive results. Inflamed epithelium should be treated prior to final endoscopic evaluation whenever possible.

It is not useful or practical to permanently use the zoom mode when screening the upper or lower gastrointestinal tract. The initial evaluation is performed in conventional mode and depends strictly on the knowledge and experience of the endoscopist. After the initial detection of discreet lesions, chromoendoscopy and magnification endoscopy are the tools used to enhance surface mucosal patterns. These techniques disclose a plethora of mucosal detail, the evaluation of which increases the procedure time, at least when the endoscopist is learning the technique.

The intra- and interobserver variability for different classifications for stained and magnified lesions have not yet been determined. In an unpublished study we have investigated the interobserver variability of pit pattern classification. The study was designed as an online questionnaire in which 10 different stained colorectal lesions, with known histology and surface appearance, could be viewed. The 166 participants (137 gastroenterologists, 10 medical students, 3 nurses, and 16 others) discriminated neoplastic lesions (pit patterns III - V) from non-neoplastic ones (pit patterns I and II) with the low sensitivity and specificity of 39 % and 26 %, respectively. Only gastroenterologists with experience in chromoendoscopy fared significantly better (sensitivity 76 %, specificity 61 %), independently of their general overall experience in endoscopy.

The proposed classifications for colorectal lesions, for Barrett’s esophagus, or for early gastric cancer types are too complex relative to their practical clinical value. A simplification is recommended. We are facing the same dilemma as our pathologist colleagues when evaluating dysplasias, with the difference that histological preparation is widely standardized whereas standardized procedure recommendations are lacking for chromoendoscopy and magnification endoscopy.

Another difficulty in the use of magnifying endoscopes is the high magnification levels. The newly developed systems allow enlargement of up to × 150. A sharp image is focused by manually adjusting the movable lens. Close examination can be difficult due to peristalsis and respiratory movements.

Chromoendoscopy has been known for a long time. The recent revival of dye spraying has been made possible by the use of newly developed magnification and high-resolution endoscopes. In combination, these techniques have led to remarkable progress in endoscopic diagnosis. Typical Barrett epithelium (specialized columnar epithelium) can be recognized endoscopically. This can be exploited for targeted biopsies and may lead to improvement in surveillance strategies. There is a noticeable trend away from random biopsies towards targeted ”smart“ biopsies, e. g. for patients with ulcerative colitis of long duration [21]. The diagnosis of depressed early colorectal cancer and flat adenoma [11] is facilitated, which provides a higher rate of endoscopic resection procedures (e. g. polypectomy, mucosal resection).

Questions arise concerning who should offer these new endoscopic techniques, and for which patients. Technical evolution continues at a very rapid rate. In a few years’ time video endoscopes with high optical performance will be widely used. It is therefore worthwhile to familiarize oneself with the newly visible details of the mucosal surface. The learning curve for achieving proficiency is considerable. Because of the high costs, general screening for gastrointestinal malignancies with high-resolution or magnification endoscopes cannot currently be recommended. So far, these endoscopes have been used mainly in gastroenterological centers. It is reasonable to use the new systems with selected patients. Screening for early malignant and premalignant changes is necessary in patient populations with a high prevalence of gastrointestinal neoplasia, for example in patients with long-standing ulcerative colitis, with Barrett’s esophagus and suspected intraepithelial neoplasia, and with familial colon cancer syndromes, or for gastric cancers in Japan. Also, prior to endoscopic mucosal resection, exact staging of the lesion in question is mandatory. Here, modern endoscopes have an important impact.

In summary, the newly developed high-resolution and magnification endoscopes offer features which allow more and new mucosal detail to be seen. They are used commonly in conjunction with chromoendoscopy. The analysis of mucosal surface detail is beginning to resemble histological examination. More accurate recognition of small flat and depressed neoplastic lesions is possible. Endoscopic prediction of neoplastic and non-neoplastic tissue is frequently possible. This can influence the endoscopic management. However, the differentiation of lesions and the diagnosis of early neoplastic changes remains difficult, despite the existing recommended classifications for various mucosal surface details. In general, endoscopists, especially western endoscopists, are at the very beginning of a long learning curve. Technical developments will continue to provide rapid improvements, and the availability of such endoscopes will increase. The endoscopist is therefore well advised to become familiar with the newly visible surface detail.

References

• 1 Sendler A, Bottcher K, Etter M. et al . Gastric carcinoma.  Internist. 2000;  41 817-830
• 2 Nelson D B, Block K P, Bosco J J. et al . High resolution and high-magnification endoscopy, guidelines: technology status evaluation report.  Gastrointest Endosc. 2000;  52 864-866
• 3 Kumagai Y, Inoue H, Nagai K. et al . Magnifying endoscopy, stereoscopic microscopy, and the microvascular architecture of superficial esophageal carcinoma.  Endoscopy. 2002;  34 369-375
• 4 Tajiri H, Doi T, Endo H. et al . Routine endoscopy using a magnifying endoscope for gastric cancer diagnosis.  Endoscopy. 2002;  34 772-776
• 5 Yagi K, Nakamura A, Sekine A. Comparison between magnifying endoscopy and histological, culture and urease test findings from the gastric mucosa of the corpus.  Endoscopy. 2002;  34 376-381
• 6 Jung M, Kiesslich R. Chromoendoscopy and intravital staining techniques.  Baillière’s Best Pract Res Clin Gastroenterol. 1999;  13 11-19
• 7 Kudo S, Tamura S, Nakajima T. et al . Diagnosis of colorectal tumorous lesions by magnifying endoscopy.  Gastrointest Endosc. 1996;  44 8-14
• 8 Kudo S, Rubio C A, Teixeira C R. et al . Pit pattern in colorectal neoplasia: endoscopic magnifying view.  Endoscopy. 2001;  33 367-373
• 9 Axelrad A M, Fleischer D E, Geller A J. et al . High resolution chromoendoscopy for the diagnosis of diminutive colon polyps: implications for colon cancer screening.  Gastroenterology. 1996;  110 1253-1258
• 10 Kiesslich R, von Bergh M, Hahn M. et al . Chromoendoscopy with indigocarmine improves the detection of adenomatous and nonadenomatous lesions in the colon.  Endoscopy. 2001;  33 1001-1006
• 11 Kiesslich R, Ohlerth C, Graebner H. et al . Early detection of colorectal adenoma and early cancer. Benefit of chromo-, magnification- or conventional colonoscopy? A multicenter prospective trial [abstract].  Gastrointest Endosc. 2002;  55 AB90
• 12 Canto M I, Setrakian S, Petras . et al . Methylene blue selectively stains intestinal metaplasia in Barrett’s esophagus.  Gastrointest Endosc. 1996;  44 1-7
• 13 Sharma P, Topalovski M, Mayo M S. et al . Methylene blue chromoendoscopy for detection of short-segment Barrett’s esophagus.  Gastrointest Endosc. 2001;  54 289-293
• 14 Guelrud M, Herrera I, Essenfeld H. et al . Enhanced magnification endoscopy: a new technique to identify specialized intestinal metaplasia in Barrett’s esophagus.  Gastrointest Endosc. 2001;  53 559-565
• 15 Endo T, Awakawa T, Takahashi H. et al . Classification of Barrett’s epithelium by magnifying endoscopy.  Gastrointest Endosc. 2002;  55 641-647
• 16 Ajioka Y, Watanabe H, Kazama S. et al . Early colorectal cancer with special reference to the superficial nonpolypoid type from a histopathologic point of view.  World J Surg. 2000;  24 1075-1080
• 17 Fujii T, Rembacken B J, Dixon M F. et al . Flat adenomas in the United Kingdom: are treatable cancers being missed?.  Endoscopy. 1998;  30 437-443
• 18 Yamao T, Shirao K, Ono H. et al . Risk factors for lymph node metastasis from intramucosal gastric carcinoma.  Cancer. 1996;  15 602-606
• 19 Ono H, Kondo H, Gotoda T. et al . Endoscopic mucosal resection for treatment of early gastric cancer.  Gut. 2001;  48 225-229
• 20 Gotoda T, Yanagisawa A, Sasako M. et al . Incidence of lymph node metastasis from early gastric cancer: estimation with a large number of cases at two large centers.  Gastric Cancer. 2000;  3 219-225
• 21 Kiesslich R, Fritsch J, Holtmann M. et al . Chromoendoscopy as a novel method for the detection of colitis-associated dysplasias and colon carcinoma in ulcerative colitis: results of a prospective randomised trial [abstract].  Gastrointest Endosc. 2002;  55 AB91