Confocal laser endomicroscopy: technical status and current indications

Transcription

1 A. Hoffman 1 M. Goetz 1 M. Vieth 2 P. R. Galle 1 M. F. Neurath 1 R. Kiesslich 1 Confocal laser endomicroscopy: technical status and current indications Confocal laser endomicroscopy is a newly introduced endo scopic tool that makes it possible to carry out confocal micro scopic examination of the mucosal layer during ongoing endos copy. Different types of tissue and diseases can be diagnosed im mediately, facilitating early diagnosis of gastrointestinal cancer. Analysis of the in vivo microarchitecture is helpful in targeting biopsies to relevant areas. In addition, subsurface imaging can unmask microscopic diseases ± (microscopic colitis) or bacterial infection (Helicobacter pylori), for example. Molecular imaging is becoming feasible, and this will shortly open the door to new in dications in gastrointestinal endoscopy (e.g., in vivo receptor a nalysis). Introduction Principles of confocal microscopy Gastroenterologists still rely on the results of histological diag nosis. Suspicious areas identified during endoscopy are targeted and biopsied or removed endoscopically. However, there are sev eral disadvantages that may be associated with white light en doscopy with biopsies, including bleeding or infection. Non rep resentative biopsies may miss relevant portions of tissue, leading to underestimation of the diagnosis [1]. Random biopsy sam pling can also be time consuming. The method of confocal laser endomicroscopy has recently been developed, which allows im mediate in vivo microscopy of the mucosal layer. A miniaturized confocal microscope is integrated into the distal tip of a conven tional endoscope [2]. This allows high resolution in vivo histolo gical assessment, so that changes in vessels, connective tissue, and cellular or subcellular structures can be evaluated during on going endoscopy examinations [3]. Confocal microscopy provides better spatial resolution than con ventional fluorescence microscopy, as the images are not con taminated by light scattering from other focal planes [4]. A low powered laser is focused onto a single point in a defined micro scopic field of view, and the same lens is used as both the con denser and objective folding optical path [4]. The point of illumi nation thus coincides with the point of detection within the spe cimen. Light emanating from that point is focused through a pin hole to a detector, and light emanating from outside the illumi nated spot is rejected. The illumination and detection systems are in the same focal plane and are termed confocal. All detect ed signals from the illuminated spot are captured and measured. The gray scale image created is an optical section representing one focal plane within the examined specimen [5]. The image of a scanned region can be constructed and digitized by measuring the light returning to the detector from successive points. Single points are typically scanned in a raster pattern Institution 1 First Dept. of Medicine, Johannes Gutenberg University, Mainz, Germany 2 Institute of Pathology, Bayreuth Hospital, Bayreuth, Germany Corresponding author R. Kiesslich, M.D., PhD I. Med. Klinik und Poliklinik Johannes Gutenberg Universität Mainz Langenbeckstraße Mainz Germany Fax: E mail: info@ralf kiesslich.de Submitted 29 March 2006 Accepted after revision 25 April 2006 Bibliography Endoscopy 2006; 38 (12): 1275±1283 Georg Thieme Verlag KG Stuttgart New York DOI /s ISSN X

2 1276 Figure 1 Scheme of confocal laser endomicroscopy. (A) A confocal la ser microscope is incorporated into the distal tip of a conventional en doscope. (B, C) A blue laser light is emitted (B) onto the mucosal sur face and into deeper parts of the mucosal layer (C). The returning light is measured, and its intensity is displayed as a gray scale image. (D) A single image represents an optical horizontal section of the mucosal layer. Series of confocal images within successive planes can be used to observe fine cellular or subcellular structures, and three dimen sional structures in the specimen can be imaged. Confocal microscopy has become a standard method for molecular ima ging in basic research, in conjunction with fluorescence labeling techniques, selectively imaging the position of specific proteins at distinct cellular locations [6]. However, confocal microscopy has so far mainly been carried out on a microscope stage on the laboratory bench. Recently, a unique miniaturized design was developed, using a single optical mode fiber acting as both the illumination point source and the detection pinhole (Optiscan Pty. Ltd., Notting Hill, Victoria, Australia). In addition to standard video imaging, the confocal laser microscope, integrated into the distal tip of a conventional video endoscope (Pentax EC 3870CIFK; Pentax, To kyo, Japan), also allows confocal microscopy of the mucosal layer at high resolution (lateral resolution 0.7 mm) [7, 8] (Figure 1). Figure 2 a The confocal laser endoscope has a short protrusion at the distal tip, which contains the confocal microscope. Gentle contact with the mucosal surface is necessary to obtain microscopic images. b Actuation of the imaging plane depth is controlled by two additional buttons on the endoscope s control. Figure 3 The orientation of images and contrast agents used for en domicroscopy. Conventional histology allows differentiation of the mucosal and submucosal layers on a vertical axis, whereas endomicro scopy provides images in a horizontal axis. The depth of infiltration therefore has to be assessed using a series of images. Most commonly, endomicroscopy is carried out after intravenous administration of fluorescein. Colonic crypts, goblet cells, and connective tissue are readily visible at high resolution. However, due to the pharmacokinetic properties of fluorescein, nuclei are not visible in all horizontal sec tions. Acriflavine can be used as a topical dye, which highlights espe cially nuclei and cell membranes. Components of the confocal laser endoscope The distal tip of this newly developed colonoscope contains the components of the confocal laser microscope, an air and water jet nozzle, two light guides, an auxiliary water jet channel (used for topical application of the contrast agent) and a 2.8 mm work ing channel, allowing endomicroscopically guided biopsies (Fig ure 2). allowing targeted endomicroscopic imaging, and the position of the focal plane within the specimen is adjusted by using two ad ditional buttons on the endoscope control. Images from the sur face to the deeper parts of mucosal layer can be obtained, and images of interest can be stored digitally on the hard drive using a foot switch. Contrast agents The diameters of the distal tip and the insertion tube are 13.4 mm and 12.8 mm, respectively. During laser endoscopy, an argon ion laser delivers an excitation wavelength of 488 nm, and the maximum laser power output is 1 mw at the surface of the tissue. The confocal image data are collected at scanning rates of 0.8 frames/s ( pixels) or 1.6 frames/s ( pix els). The optical slice thickness is 7 mm, with a lateral resolution of 0.7 mm (field of view mm). The range of the z axis is 0±250 mm below the surface layer. The confocal endoscope can be handled in the same way as a standard endoscope. The distal tip of the endoscope is placed in gentle contact with the mucosa, Confocal imaging at high resolution is possible using an exoge nous fluorescence technique [9]. Potentially suitable agents are fluorescein, acriflavine, tetracycline, or cresyl violet. The most commonly used contrast agents are acriflavine hydrochloride (0.05 % in saline; topical use only) or fluorescein sodium (5 ± 10 ml of a 10% solution; intravenous application) [10]. Confocal imaging following staining with acriflavine hydrochloride and fluorescein sodium shows the characteristic morphology of mu cosal tissue [8]. Whereas topically used acriflavine hydrochloride strongly labels the superficial epithelial cells, including nuclei, intravenously administered fluorescein sodium distributes

3 throughout the entire mucosa, with strong contrast in the con nective tissue and the capillary network [11]. Fluorescein binds to serum albumin, and remaining unbound dye molecules pass across the systemic capillaries and enter the tissue, highlighting the extracellular matrix. Confocal images can be generated si multaneously with endoscopic images, making it possible to identify typical histological structures in the upper and lower gastrointestinal tract (Figure 3) [12, 13]. In the colon, mucin containing goblet cells and columnar epithe lial cells can be readily identified. The luminal openings of the crypts appear in the horizontal axis as black holes projecting onto the surface of the mucosa, and each crypt is covered with a layer of epithelial cells. The microvascularization is highlighted within the lamina propria in deeper parts of the mucosal layer. The vasculature within the mucosa of the colon shows a typical honeycomb appearance that represents a network of capillaries. Red blood cells are not labeled by fluorescein, and appear as moving black dots in the lumen of the vessels. Imaging of the duodenum or the terminal ileum allows visualiza tion of intestinal villi, including the brush border. The villi appear as small finger like extensions projecting from the surface. The epithelial cells covering the villi can be readily identified, and single goblet cells can be identified. In the stomach, the architecture of the gastric pits can be ob served as small invaginations on the luminal surface, consistent with the known histology of the stomach. The surface cells pres ent a typical cobblestone appearance in endomicroscopy, a find ing familiar from electron microscopy. resected colorectal lesions from 90 patients [14]. Nuclei could not be visualized in normal colonic mucosa or hyperplastic polyps, but were more often visible in neoplastic lesions. Conver sely, goblet cells were visible less often in malignant or premalig nant lesions. However, it should be noted that it was not individ ual nuclei, but rather dark areas in irregular cells that were visi ble. The study found a statistically significant difference between nonneoplastic and neoplastic lesions in relation to the detection rate of nuclei (dark areas) in laser scanning confocal microscopy images. On the basis of these results, the authors recommended prelim inary criteria for a confocal imaging classification of high grade intraepithelial neoplasia and cancer. Neoplasia was character ized by the presence of any structural abnormality and clear vi sualization of nuclei. However, the sensitivity of this method for predicting neoplasms in the colorectum was only 60%, reflecting the limited resolution of the system that was used. In a recently published study using the newly developed endo microscopic system, 42 patients with indications for screening or surveillance colonoscopy after previous polypectomy under went in vivo endomicroscopy with the confocal laser endoscope [15]. The aim of the study was to assess the histology in vivo dur ing ongoing colonoscopy in order to diagnose intraepithelial neoplasias and colon cancer. Fluorescein guided endomicrosco py of intraepithelial neoplasias and colon cancers showed a tub ular, villous, or irregular architecture, with a reduced number of goblet cells. In addition, neovascularization in neoplasms is char acterized by an irregular vessel architecture with fluorescein leakage. The luminal surface of the esophagus is composed of nonkerati nized squamous epithelium with polygonal epithelial cells and microvasculature loops within the papillae, which can be recog nized during confocal microscopy. Even the intercellular spaces in between single cells can be evaluated. The squamocolumnar junction at the Z line in the distal esophagus appears individ ually as a mosaic pattern, with columnar epithelial cells indicat ing cardiac mucosa. A simple classification of the confocal pattern (Table 1), based on initial experience with confocal endomicroscopy, was developed to allow differentiation between neoplastic and nonneoplastic tissue. Macroscopic and microscopic images were taken together to allow an immediate prediction of the histopathology. A total of confocal images from 390 locations were compared with the histological data from 1038 biopsies [15]. It was possible to predict the presence of neoplastic changes using the newly de 1277 Confocal imaging of colon pathology Colorectal cancer is still one of the leading causes of cancer relat ed death in the Western world. Screening colonoscopy is widely accepted as the gold standard for early diagnosis of cancer. The prognosis for patients with colonic neoplasms is strictly depen dent on the depth of infiltration, and therefore depends on early detection of preinvasive and neoplastic changes. Early detection makes it possible to cure the patient by immediate endoscopic resection. In 2003, Sakashita et al. reported initial experience with real time confocal endoscopy in ex vivo specimens [14]. The proto type endomicroscope that was used (Olympus Optical Ltd., To kyo, Japan) was passed through the working channel of an endo scope. The aim of the study was to establish new criteria for dis tinguishing between benign lesions and high grade dysplasia or cancer. The authors examined 100 endoscopically or surgically Table 1 Confocal laser endomicroscopy classification of patterns in colorectal lesions Grading Vessel architecture Crypt architecture Normal Regenera tion Neoplasia Hexagonal, honeycomb appear ance that presents a network of capillaries outlining the stro ma surrounding the luminal openings of the crypts Hexagonal, honeycomb appear ance with no increase or only a slight increase in the number of capillaries Regular luminal openings and distribution of crypts covered by a homogeneous layer of epithelial cells, including goblet cells Star shaped luminal crypt openings or focal aggregation of regular shaped crypts with a regular or reduced amount of goblet cells Dilated and distorted vessels Ridge lined irregular epithelial with increased leakage; irregu layer with loss of crypts and lar architecture, with little or no goblet cells; irregular cell orientation to the adjoining tis architecture, with little or no sue mucin

4 Figure 4 A flat hy perplastic lesion in the sigmoid. Chro moendoscopy with methylene blue can be used to reveal flat lesions. a The atypi cal star shaped opening of crypts can be seen on close inspection. b Endo microscopy can con firm the star shaped luminal opening of single crypts. c, d Fusion of crypts is visible. However, the arrangement of the columnar cells and goblet cells is regu lar, indicating non neoplastic tissue. Figure 5 Aberrant crypt foci can pro gress into colorectal adenoma and can cer. a Aberrant crypt foci can only be identified with the help of chromoen doscopy. b Magnifi cation makes it pos sible to analyze the crypt arrangement. Endomicroscopy re vealed similar chang es. c Three lumina within a single crypt can be seen, in addi tion to the cellular and vascular archi tecture. d Targeted biopsies confirmed the presence of aber rant crypt foci with three luminal open ings veloped confocal pattern classification with a sensitivity of 97.4%, a specificity of 99.4 %, and an accuracy of 99.2%, respec tively (Figures 4 ± 6). Ulcerative colitis It is not possible to examine the whole surface of the colon in the endomicroscopic mode. In patients with ulcerative colitis, it is therefore important to combine endomicroscopy with chromo endoscopy. Panchromoendoscopy with either methylene blue or indigo carmine is a valid diagnostic tool for improving the diag nostic yield of intraepithelial neoplasia using the SURFACE re commendations [16]. Chromoendoscopy can reveal circumscrib ed lesions [17], and chromoscopy guided confocal laser endomi croscopy can be used to predict intraepithelial neoplasias with a high degree of accuracy [15]. Targeted biopsies of relevant le sions can therefore be taken, and rapid confirmation of neoplas tic changes using confocal laser endoscopy during colonoscopy may lead to significant improvements in the clinical manage

5 Figure 6 Colorectal neoplasia. a Video endoscopy shows a polypoid lesion with a villous surface. b Endomicroscopy after acriflavine staining shows the typical villous archi tecture. Nuclei with different shapes and sizes are visible, in dicating neoplasia. The final histological analysis revealed a villous adenoma with high grade in traepithelial neopla sia. c A circular ste nosis is evident in the rectum. d Endo microscopy shows destruction of the basement mem brane and malignant infiltration of the la mina propria. The fi nal histological anal ysis identified a moderately differen tiated adenocarcino ma. pared with the histological results from 1392 biopsies. Sixty sev en of 95 circumscribed lesions were only visible after chromoen doscopy with methylene blue [17]. Different cellular structures (epithelial and blood cells), capillaries, and connective tissue limited to the mucosal layer were identified by confocal micros copy [18]. Due to the pharmacokinetic properties of fluorescein, nuclei could not be seen [11]. However, the presence of neoplas tic changes (sensitivity 94.4 %, specificity 95.6 %, accuracy 99.3 %) and inflammation were predictable with a high degree of accura cy [17]. In the first randomized trial of endomicroscopy, 153 patients with long term ulcerative colitis who were in clinical remission were randomly assigned at a ratio of 1 : 1 to undergo either con ventional colonoscopy or panchromoendoscopy using 0.1% met hylene blue in conjunction with endomicroscopy to detect in traepithelial neoplasia or colorectal cancer [18]. Circumscribed lesions in the colonic mucosa detected by chromoendoscopy were evaluated with endomicroscopy for cellular and vascular changes in accordance with the confocal pattern classification for predicting neoplasia. Targeted biopsies from the areas exam ined were taken and histologically graded according to the World Health Organization and new Vienna classification [18]. In the standard colonoscopy group, randomized biopsies every 10 cm between the anus and cecum were taken, as well as target ed biopsies of visible mucosal changes. The primary outcome a nalysis was a histological diagnosis of neoplasia. Using chromo endoscopy in conjunction with endomicroscopy (80 patients, average examination time 42 min), significantly more intraepi thelial neoplasia was detected (19 versus 4 cases; P = 0.007) than with standard colonoscopy (73 patients, average examina tion time 31 min). Endomicroscopy revealed different cellular structures (epithelial and blood cells), capillaries, and connective tissue limited to the mucosal layer. A total of 5580 confocal im ages from 134 circumscribed lesions were compared with the histological results from 311 biopsies. The presence of neoplastic changes was predicted with a high degree of accuracy (sensitiv ity 94.7 %, specificity 98.3 %, accuracy 97.8%) [18] In summary, chromoendoscopy is able to reveal circumscribed lesions, and confocal laser microscopy can be used to confirm in traepithelial neoplasias with a high degree of accuracy. Biopsies can therefore be limited to targeted sampling of relevant lesions. In vivo histology with endomicroscopy may lead to significant improvements in the clinical management of patients with ul cerative colitis (Figures 6 ± 8), with reduced numbers of biopsies being needed for confirmation of the condition and time being gained for immediate therapeutic intervention. ment of patients with ulcerative colitis. In 41 patients with long term ulcerative colitis who were in clinical remission, endomi croscopy in conjunction with methylene blue aided panchro moendoscopy was used for surveillance [17]. Chromoendoscopy with methylene blue did not lead to any interference with the la ser scanning system. A total of confocal images from 464 different locations (369 inconspicuous areas, 95 circumscribed lesions) were com Collagenous colitis Endomicroscopy makes it possible to locate and measure the dis tribution and thickness of collagenous bands underneath the epithelial layer, thus allowing targeted biopsies ± a new ap proach in collagenous colitis, particularly in cases with disrupted subepithelial collagen deposits. At present, randomized biopsies are recommended, preferably from the right colon. The distribu tion of the collagenous bands can be patchy and segmental in the colon. Confocal endomicroscopy helps differentiate between af fected and normal sites and can guide biopsies [19].

6 Figure 7 Barrett s esophagus. a A typi cal tongue like co lumnar epithelium is present. The blue la ser beam can be seen. b In a second case, a short seg ment Barrett s esophagus was also suspected. c, d En domicroscopy can be used to diagnose goblet cells within columnar lined epi thelium (arrows). Goblet cells can be identified due to black dots (mucin) incorporated into single columnar cells. Endomicrosco py guided biopsies can be taken, as the working channel is close to the confocal microscope at the distal tip of the en doscope. Figure 8 Barrett s associated neopla sia. a A type IIa ± IIc lesion is visible within a long seg ment Barrett s esophagus. b Chro moendoscopy helps reveal the borders of the lesion. c Targe ted endomicroscopy shows irregular ar chitecture in the tis sue and vessels. d Very dark cells with different shapes and sizes are characteris tic of the presence of Barrett s associated neoplasia Confocal imaging in the upper gastrointestinal tract Barrett s esophagus Barrett s esophagus is known to be a premalignant condition in patients with gastroesophageal reflux disease, and most adeno carcinomas of the distal esophagus have been shown to arise in Barrett s tissue. Barrett s esophagus is defined histologically by the presence of specialized columnar epithelium (SCE) with gob let cells. The columnar lined lower esophagus (CLE) can be iden tified during standard upper endoscopy. SCE is often present in a patchy mosaic contribution within CLE and can be overlooked by random biopsies, resulting in biopsies of the cardia type of mu cosa without goblet cells. However, it has been suggested that four quadrant step biopsies within CLE should serve as the gold standard for diagnosing Barrett s epithelium and Barrett s asso ciated neoplastic changes.

7 Table 2 Confocal diagnosis Gastric type epithelium Barrett s epithelium Neoplasia Confocal laser endomicroscopy classification of Barrett s esophagus Vessel architecture Capillaries with a regular shape only visible in the deeper parts of the mucosal layer Subepithelial capillaries with a regular shape underneath co lumnar lined epithelium visible in the upper and deeper parts of the mucosal layer Crypt architecture Regular columnar lined epi thelium with round glandular openings and typical cobble stone appearance Columnar lined epithelium with intermittent dark mucin in goblet cells in the upper parts of the mucosal layer. In the deeper parts, villous, dark, regular cylindrical Barrett s epithelial cells are present Irregular capillaries visible in Black cells with irregular apical the upper and deeper parts of and distal borders and shapes, the mucosal layer. Leakage of with strong dark contrast vessels leads to a heteroge against the surrounding tissue neous and brighter signal inten sity within the lamina propria Figure 9 Gastritis. a, b There is an ero sion in the antrum. c Endomicroscopy shows an increase in the vasculature (ar row). Prominent vessels filled with red blood cells can be readily visualized. d Intestinal metapla sia is visible in the gastric body (arrow). Goblet cells and mu cosal architecture similar to the colon can be observed. Targeted biopsies revealed chronic atrophic gastritis with intestinal meta plasia. Endomicroscopy makes it possible to identify CLE macroscopi cally and identify goblet cells microscopically in the distal esoph agus, allowing an immediate and reliable diagnosis of Barrett s esophagus. In an endomicroscopic study including 63 patients, different types of epithelial cell were distinguished and cellular and vas cular changes were detected using fluorescein guided endomi croscopy [20]. A classification of confocal images for the diagno sis of Barrett s epithelium and Barrett s associated neoplasias was developed on the basis of a comparison of the in vivo and conventional ex vivo histology (Table 2). The classification dis tinguishes between three types of epithelium (gastric epithe lium; Barrett s epithelium without neoplastic changes; and Bar rett s epithelium with neoplastic changes) Confocal imaging of the normal squamous epithelium of the esophagus demonstrated squamous cells at high resolution, showing capillaries (filled with red blood cells) within single pa pillae. It also became obvious that the number of papillae ap pears to increase after damage to the epithelium (e.g., in erosive esophagitis). It is possible to diagnose dilated intercellular spaces, which can be seen in patients with esophageal damage. Analysis of the Z line showed the clear border between squa mous and columnar lined epithelium. Goblet cells, which are pa thognomonic for Barrett s epithelium, are easily identified. The mucin (MUC2) in goblet cells appears as dark spots within single cells of columnar lined epithelium. The typical shape of Barrett s epithelium was villous, differing from the cardiac epithelium. High grade intraepithelial neoplasias or early cancers can be rec ognized by a distinct cell type in endomicroscopy. The highly ir regular and polygonal cells have a rather black appearance, with irregular borders. In addition, an irregular epithelial cell layer with typical black cells and loss of a regular basal border was found to indicate high grade intraepithelial neoplasia. The brightness of the lamina propria became heterogeneous due to the mixed vasculature of neoangiogenesis and leakage phenom ena. In this study [20], 156 areas and 3012 images were reassessed in accordance with the confocal Barrett classification and compar ed with the targeted biopsies (411 biopsies). The comparison showed that Barrett s esophagus can be predicted with the help of confocal endomicroscopy with a sensitivity of 98.1 % and a specificity of 94.1 %, respectively (accuracy 96.8%; positive pre dictive value 97.2 %; negative predictive value 96.0%). Moreover, Barrett s associated neoplastic changes can be predicted with a sensitivity of 92.9% and a specificity of 98.4 %, respectively (accu

8 1282 racy 97.4 %; positive predictive value 92.9 %; negative predictive value 98.4 %) (Figures 7, 8). Gastric cancer Confocal laser endomicroscopy has also been used to diagnose gastric cancer and precancerous conditions. Endomicroscopy was performed on five ex vivo gastrectomy specimens and in up per gastrointestinal endoscopies in vivo in eight patients [21]. Acriflavine hydrochloride dye was used for ex vivo examinations, and intravenous fluorescein sodium for in vivo examinations. A standard upper endoscopy examination was carried out, during which confocal images were obtained at standardized locations in the gastric antrum, body, and cardia, before biopsy specimens were taken from the same areas for histopathology. Confocal di agnostic criteria were established by comparison with histopa thology as the gold standard. Five endoscopists who were blind ed to the histological findings independently scored the confocal images on a template. The interobserver correlation was ana lyzed using kappa statistics. A total of 2766 confocal images from 132 different locations were obtained and compared with the histological findings in 44 biop sy specimens. Diagnostic confocal features were established for normal gastric mucosa, chronic gastritis (presence of chronic in flammatory cells), intestinal metaplasia (goblet cells), and cancer (architectural atypia, increased nuclear cytoplasmic ratio, chro matin condensation). Using these diagnostic features, a prospec tive and blinded evaluation showed that the presence of gastric cancer was capable of being predicted from confocal images with a high degree of accuracy (sensitivity 84 %; specificity 95 %; accu racy 80%). The kappa statistics showed that the interobserver agreement figures on the presence of intestinal metaplasia and cancer for the various sites (antrum, body, cardia) were 0.83, 0.89, and 0.63, respectively. and size of the bacteria, including the flagella, were identifiable. Helicobacter pylori infection was confirmed by histology and cul ture. Ex vivo examination of the cultures also showed active up take of acriflavine by Helicobacter pylori. Conclusion In vivo confocal laser endomicroscopy is a newly developed diag nostic tool that allows virtual histology of the mucosal layer dur ing ongoing endoscopy. The quality of the new, detailed images obtained with confocal laser endomicroscopy surely represents the start of a new era, in which this development in optical tech nology will allow unique visualization of living cells and cellular structures at and below the surface of the gut. Several prospec tive studies have already been published confirming the high level of diagnostic accuracy of confocal laser endomicroscopy. The diagnostic spectrum of confocal endomicroscopy is current ly expanding from screening and surveillance for colorectal can cer towards Barrett s esophagus, Helicobacter pylori associated gastritis, and gastric cancers. In addition, several other indica tions ± such as celiac disease, microscopic colitis, mucosa asso ciated lymphoid tissue (MALT) lesions, and squamous cell carci noma ± are also being investigated. Endomicroscopy is likely to play an increasingly important diagnostic role during gastroin testinal endoscopy in the future. Further technological develop ments and improvements are likely to enhance the imaging facil ities further. The door leading to in vivo functional and molecular diagnosis is about to open. Competing interests: The 1. Medical Clinical of the Johannes Gu tenberg University of Mainz was supported by an unrestricted grant by Pentax Europe. Confocal laser endoscopy allows immediate in vivo diagnosis of mucosal neoplasia and preneoplasia. Diagnoses of intestinal me taplasia and gastric carcinoma can be made using the criteria de scribed above with a reliable level of interobserver agreement. This allows cellular diagnoses during endoscopy, whereas intes tinal metaplasia was previously only capable of being diagnosed on the basis of the histopathology. In populations at risk of gas tric cancer, confocal endoscopy has potential clinical applica tions in connection with screening for gastric neoplasia and pre neoplasia [21] (Figure 9). Helicobacter pylori Usually, Helicobacter pylori infection can be diagnosed by nonin vasive and invasive methods such as the urea breath test, stool test, urease testing on endoscopic biopsies, and serological as says. The sensitivity of urease testing on antral biopsy specimens is 79 ± 100%, and the specificity ranges from 92 % to 100 % [22, 23]. However, an alternative diagnostic approach is in vivo identification of Helicobacter pylori using acriflavine guided endomicroscopy. Helicobacter pylori infection was first detected in vivo in 2005 in a 70 year old man [24]. Single as well as accumulated white dots were observed within the gastric mucosa after topical applica tion of acriflavine onto the gastric surface. The distinct shape References 1 McLaren W, Anikijenko P, Barkla D et al. In vivo detection of experi mental ulcerative colitis in rats using fiberoptic confocal imaging (FOCI). Dig Dis Sci 2001; 46: 2263 ± Da Costa RS, Wilson BC, Marcon NE. Optical techniques for the endo scopic detection of dysplastic colonic lesions. Curr Opin Gastroenterol 2005; 21: 70 ± 79 3 Macrae F, Polglase AL, McLaren WJ et al. Confocal endomicroscopy for detection of normal and disease human colonic mucosa in vivo. (Paper presented at Digestive Disease Week, Chicago: 14±19 May 2005) 4 Helmchen F. Miniaturization of fluorescence microscopes using fibre optics. Exp Physiol 2002; 87: 737 ± Robinson JP. Principles of confocal microscopy. Methods Cell Biol 2001; 63: 89 ± Polglase AL, McLaren WJ, Skinner SA et al. A fluorescence confocal en domicroscope for in vivo microscopy of the upper and the lower GI tract. Gastrointest Endosc 2005; 62: 686 ± Koenig F, Knittel J, Stepp H. Diagnosing cancer in vivo. Science 2001; 292: 1401 ± Kiesslich R, Neurath FM. Endoscopic confocal imaging. Clin Gastroen terol Hepatol 2005; 3 (7 Suppl 1): S58 ± S60 9 Kim J. The use of vital dyes in corneal disease. Curr Opin Ophthalmol 2000; 11: 241± Selkin B, Rajadhyaksha M, Gonzalez S, Langley RG. In vivo confocal mi croscopy in dermatology. Dermatol Clin 2001; 19: 369 ± 377, ix±x 11 Pawley JB. Limitations on optical sectioning in live cell confocal mi croscopy. Scanning 2002; 24: 241 ± Tadrous PJ. Methods for imaging the structure and function of living tissues and cells, 3: confocal microscopy and micro radiology. J Pathol 2000; 191: 345± 354

9 13 Delaney PM, Harris MR. Fiberoptics in confocal microscopy. In: Pawley JB (ed). Handbook of biological confocal microscopy, 2nd ed. New York: Plenum Press, 1995: 515± Sakashita M, Inoue H, Kashida J et al. Virtual histology of colorectal le sions using laser scanning confocal microscopy. Endoscopy 2003; 35: 1033± Kiesslich R, Burg J, Vieth M et al. Confocal laser endoscopy for diagnos ing intraepithelial neoplasias and colorectal cancer in vivo. Gastroen terology 2004; 127: 706 ± Kiesslich R, Neurath MF. Surveillance colonoscopy in ulcerative colitis: magnifying chromoendoscopy in the spotlight. Gut 2004; 53: 165± Kiesslich R, Fritsch J, Holtmann M et al. Methylene blue aided chro moendoscopy for the detection of intraepithelial neoplasia and colon cancer in ulcerative colitis. Gastroenterology 2003; 124: 880 ± Kiesslich R, Goetz M, Schneider C et al. Confocal endomicroscopy as a novel method to diagnose colitis associated neoplasias in ulcerative colitis: a prospective randomized trial. (Paper presented at Digestive Disease Week, Chicago: 14 ± 19 May 2005) 19 Kiesslich R, Hoffman A, Goetz M et al. In vivo diagnosis of collagenous colitis by confocal endomicroscopy. Gut 2006; 55: 591 ± Kiesslich R, Gossner L, Dahlmann A et al. In vivo histology of Barrett s esophagus and associated neoplasias by confocal laser endomicrosco py. Clin Gastroenterol Hepatol 2006; 8: 979± Yeoh KG, Salto Tellez M, Khor CJL et al. Confocal laser endoscopy is useful for in vivo rapid diagnosis of gastric neoplasia and pre neopla sia. (Paper presented at Digestive Disease Week, Chicago: 14 ± 19 May 2005) 22 Nakata H, Itoh H, Ishiguchi T et al. Immunological rapid urease test using monoclonal antibody for Helicobacter pylori. J Gastroenterol Hepatol 2004; 19: 970 ± Lim LL, Ho KY, Ho B, Salto Tellez M. Effect of biopsies on sensitivity and specificity of ultra rapid urease test for detection of Helicobacter py lori infection: a prospective evaluation. World J Gastroenterol 2004; 10: 1907± Kiesslich R, Goetz M, Burg J et al. Diagnosing Helicobacter pylori in vivo by confocal laser endoscopy. Gastroenterology 2005; 128: 2119±