Confocal laser endomicroscopy is a new field of endoluminal

Transcription

1 Imaging and Advanced Technology Michael B. Wallace, Section Editor Probe-Based Confocal Laser Endomicroscopy MICHAEL B. WALLACE* and PAUL FOCKENS *Mayo Clinic, Jacksonville, Florida; and Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Confocal laser endomicroscopy is a new field of endoluminal imaging that offers extremely high magnification and resolution, approximating white light microscopy. This has the potential to fundamentally change the current algorithms of gastroenterologic diagnosis. A recent consensus conference (International Conference of Cellvizo Users, Miami, FL, Feb 22 23, 2009) on probebased confocal laser endomicroscopy (pcle) held in February 2009, established basic indications, techniques, research priorities and standards for image interpretation. This article summarizes the findings of that meeting. Since the inception of flexible endoscopy, the endoscope has been used as both a diagnostic and therapeutic instrument. The diagnostic component has relied heavily on endoscopically directed biopsy, with histology and all of its subforms serving as the gold standard. Although highly accurate, histology has major limitations including: incremental cost, risk, time delay, lack of in vivo information such as blood flow, and limited ability to predict disease course. In the case of bile duct cancers, biopsy is particularly prone to false-negative results. On the other hand, most endoscopic imaging tools, such as high-definition endoscopes, with or without optical enhancement, are useful for guiding biopsy, but are rarely able to make specific diagnoses of normal or abnormal tissue sufficient to replace biopsy. This paradigm seems likely to change. CLE can be performed currently with 1 of 2 FDA approved devices: 1 integrated into an endoscope (Pentax, Ft Wayne, NJ; herein termed ecle) and 1 as a stand-alone probe (herein termed pcle) capable of passage through the accessory channel of most endoscopes (Cellvizio, Mauna Kea Technologies, Paris, France). This review focuses on the pcle system (Figure 1). A previous column in this section of GASTROENTEROLOGY has discussed the ecle system. 1 pcle has several advantages and disadvantages compared with ecle. Advantages include the greater versatility of pcle probes, which can be used in conjunction with virtually any endoscope (or cholangioscope, bronchoscope, ureteroscope etc), ad hoc usage, such as when a lesion is detected with a normal endoscope, and acquisition at video frame rate of 12 frames/ sec allowing in vivo imaging of capillary flow (Video). Disadvantages include a slightly lower resolution (approximately 1 m compared with 0.7 m for ecle) and smaller field of view ( m). The fiber probes consist of a bundle of 30,000 optical fibers with a distal lens, and proximal precision connector. The fluorescent signal returning from the tissue is converted into an image using a detector (Avalanche Photo Diode), and software/hardware systems for image correction, stabilization, and display. Clinical image acquisition is optimized by use of a contrast agent. Although many previously published images with the ecle system have used topical acriflavine dye, concerns about DNA damage 2 by this and other nuclear stains have reduced its use. Most pcle imaging is performed with intravenous fluorescein, an agent FDA approved for diagnostic fluorescein angiography or angioscopy of the retina andiris vasculature. Fluorescein is a highly safe agent whose major side effects are short term (1 2 hours) and include yellowish skin discoloration and 1 2 days of bright yellow-colored urine. In a safety analysis of IV fluorescein for pcle imaging, no serious complications were observed in 410 consecutive cases. 3 The current potential indications for pcle imaging are broad and include almost all current applications of endoscopic biopsy. Early data suggest that the major capabilities of pcle will be to distinguish non-neoplastic tissue from neoplasia, such as surveillance of nondysplastic Barrett s esophagus (BE), chronic inflammatory bowel disease (IBD), small colorectal polyps, and indeterminate bile duct strictures. Other novel applications include detection of early rejection in small bowel transplantation, detection of residual neoplasia after endoscopic mucosal resection of large flat colorectal polyps, detection of microscopic colitis, and detection of celiac sprue. In most of these, the key role will be to detect nondiseased tissue, Abbreviations used in this paper: BE, Barrett s esophagus; ecle, endoscope-based confocal laser endomicroscopy; IBD, inflammatory bowel disease; IEN, intraepithelial neoplasia; NPV, negative predictive value; pcle, probe-based confocal laser endomicroscopy by the AGA Institute /09/$36.00 doi: /j.gastro GASTROENTEROLOGY 2009;136:

2 Figure 1. The probe-based confocal laser endomicroscopy (pcle) imaging system showing the fiber probe within a standard endoscopic accessory channel, and the laser scanning unit and software interface. and thus eliminate the large proportion of biopsies done, which yield no disease. In all of these applications, pcle will likely need to be used in conjunction with a red flag technology. pcle is a small-field imaging system, and thus is only appropriate for classification of tissue at a site already detected by standard or optically enhanced endoscopy. An example would be use of narrow-band imaging to detect regions of suspicion in BE, followed by pcle to confirm intraepithelial neoplasia (IEN), and guide immediate therapy. Barrett s Esophagus. Current surveillance guidelines for BE call for 4 quadrant random biopsies every 1 2 cm throughout the length of columnar epithelium in the esophagus. In patients without IEN, the annual incidence of high grade IEN or cancer is 1 in 200 per year. The pathology cost to Medicare alone for a single jar of 4 biopsies is substantial. Thus, a technology that could reliably exclude neoplasia has the potential to dramatically reduce the need for, and cost of, random biopsies. Early evidence using pcle for BE has identified key features of neoplasia, and was able to detect IEN with a per-biopsy sensitivity for 2 independent investigators of 75%, and specificity of 89% 91% with good interobserver agreement ( 0.6). 4 In the low-risk population studied, this led to a 98.8% negative predictive value (NPV), thus allowing nearly risk-free elimination of the random biopsy when pcle was negative. The features and example images indicative of neoplasia are shown in Table 1. A prospective, multicenter trial is now underway to evaluate the accuracy of pcle in comparison with high-definition white light and narrow-band imaging endoscopy. Potential future applications of pcle take advantage of the unique aspect of real-time in vivo imaging in living tissue. These include the potential for novel biomarkers of risk and prognosis such as angiogenesis, and the ability to image fluorescent-tagged molecular agents. Fundamentally, the field of Barrett s will need to move beyond reliance on histologic intestinal neoplasia alone as a biomarker of risk. Whether this will be accomplished by in vivo imaging markers, genomic biomarkers, proteomic biomarkers, or other means is not yet known. Colorectal Disease. Colorectal cancer screening with colonoscopy and polypectomy remains the gold standard for disease prevention. Despite advantages, there are major limitations to the current paradigm including the large number of benign (small distal hyperplastic) polyps, and increased risks and costs associated with polypectomy. Recent studies have shown that polypectomy is the single greatest risk factor for major complications of colonoscopy. 5 Table 1. Nondysplastic Barrett s (with permission from Pohl et al 4 ) Absence of criteria below. IEN Irregular epithelial lining; variable width of the epithelial lining; fusion of glands; presence of dark areas (decreased uptake of fluorescein); irregular vascular pattern. 1510

3 One third to one half of all polyps in the colon are hyperplastic; most of these are small ( 10 mm), distal lesions with extremely low malignant potential. With current methods of endoscopy, even with high-definition, optically enhanced colonoscopes, the accuracy of polyp classification is only 80% 90%. Current guidelines still call for removal and histologic classification of all colorectal polyps. 6 pcle has several potential roles in polyp management. The best-studied application is to distinguish between hyperplastic and adenomatous polyps, thus negating the need to remove hyperplastic polyps. This recognized that some large, proximal, hyperplastic lesions, especially those now being reclassified as serrated lesions, should be removed. In a large, prospective, blinded trial, our group evaluated pcle in 60 patients with 103 polyps and found a sensitivity of 80% and specificity of 94% for detection of adenomatous polyps. 7 A key issue for polyp application is the NPV. In a survey of 25 gastroenterologists attending the Miami conference, the majority would require a NPV 95% to leave a polyp in situ and 20% would require an NPV 99%. Such high levels of certainty are likely achievable with pcle, particularly with low-risk lesions. For example, a 6-mm, pale-appearing polyp in the rectosigmoid (estimated pretest probability of adenoma 10%) combined with negative pcle image (estimated sensitivity 80%) has a NPV of 98%. With a sensitivity of 90%, that NPV would exceed 99% and thus make leaving benign-appearing polyps in situ more acceptable. Societal guidelines and more validated studies are clearly needed, however. Surveillance in chronic IBD is recommended after longstanding disease. As is the case in BE, the yield for IEN on random biopsy is low, yet standard endoscopic imaging cannot identify most IEN. Recent studies have shown that targeting biopsies with methylene blue chromoendoscopy significantly increases the yield of IEN. More important, the yield of IEN in chromo-negative sites is 0.1%. Endoscope-based CLE has recently been shown to further increase the yield for IEN above and beyond methylene blue. 8 pcle also has the capacity to differentiate normal from inflamed tissue, and thus target biopsies for the purpose of grading and mapping the extent of colitis. 9,10 The role of pcle in chronic IBD surveillance is likely to be replacement of these random biopsies, particularly in chromo-negative regions, and to further target biopsies in circumscribed lesions detected with optical enhancement methods. Whether or not newer optical contrast methods (narrow-band imaging, AFI, FICE, iscan) will replace chromoendoscopy as the red flag methods, and how accurate these will be, remains to be determined, but early studies are promising. 11,12 It is likely that some highly sensitive methods such as pcle will still be needed to replace the random biopsy. Figure 2. Confocal image of pancreatic acini (B) obtain via endoscopic ultrasound-guided fine needle (A, arrow) pcle in a pig. pcle may also be valuable in detection of microscopic colitis in patients with chronic diarrhea. Although collagenous and lymphocytic colitis are rare, biopsy is recommended but of very low yield in patients with chronic diarrhea. Recent studies have suggested that confocal endomicroscopy has the capacity to detect both lymphocytic 13 and collagenous 14,15 colitis. Thus, pcle has the potential to replace or direct a large number of random biopsies in patients with chronic diarrhea, where the confocal image is normal. Pancreatobiliary Applications. One of the major advantages of probe-based CLE is the small diameter of the fiber, thus allowing imaging within the biliary or pancreatic duct, and even through fine-needle systems in solid organs and lymph nodes. Fibers as small as 300 m in diameter are currently available for animal imaging and have been used in endoscopic ultrasound guided fine needle aspiration of solid organs such as the pancreas (Figure 2). 1511

4 Figure 3. pcle images of normal bile duct (A) with fine-regular reticular pattern compared with the dilated irregular dark structures and bright, irregular vessels in cholangiocarcinoma (B). (Courtesy of Professor Alexander Meining.) The major role of pcle in the bile duct is likely to detect cancer in indeterminate bile duct and pancreatic strictures. Biopsy and other methods such as cytologic brushing and needle aspiration have very low levels of diagnostic accuracy in this setting. In the bile duct, confirmation and treatment of cancer is frequently delayed or requires operative intervention or long-term followup. In a pilot study of 14 patients with indeterminate biliary strictures, Meining et al were able to accurately (in fact even more accurately than histologic brushing and biopsy) distinguish malignant from benign strictures using pcle, based on the presence of large, irregular microvessels (Figure 3). Normal bile duct wall had a very regular, reticular pattern. A large, multicenter trial is now underway in the United States and Europe to confirm these findings and further improve the diagnostic criteria for neoplasia. Miscellaneous Applications. pcle imaging has the potential to replace or guide biopsy in almost any condition where biopsy is needed. Several novel applications include detection of rejection in small bowel transplantation where extensive biopsies are often needed, and biopsies carry an increased risk of complication in the transplanted bowel. pcle imaging is well suited to take advantage of recent advances in molecular imaging. Fluorescent tags, nanoparticles, and quantum dots can all be attached to specific monoclonal antibodies, or peptides that bind to specific molecular targets such as cathepsins or matrix metalloproteases. These molecular beacons can be tuned to either visible (488-nm) lasers or near-infrared lasers, which have the advantage of minimal background autofluorescence. Recent studies using a heptamer peptide-linked to fluorescein allowed sensitive and specific imaging of adenomatous colon polyps. 16 Although most research to date compares confocal imaging with histology, pcle also offers the potential to image structure and function that cannot be seen in excised tissue. By imaging in vivo, moving structures can easily be seen, particularly vascular flow (Video). Such information may allow new biomarkers of disease, prognosis, and prediction to be developed beyond histologic capabilities. Probe-based CLE is a rapidly emerging field of gastroenterology that bridges the interface between endoscopy and histology. It further expands our ability to image living tissue in real time and to provide therapy in the same setting. The immediate impact will be the ability to target biopsies much more precisely, and eliminate a large number of noninformative random biopsies. Longterm impacts will be a rethinking of the dependence on histology as our biomarker of choice for detection, prognostication, and prediction of gastrointestinal disease and therapy. Supplementary Data Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at and at doi: /j.gastro References 1. Wallace M. Leeuwenhoek meets Kussmaul: the evolution of endoscopist to endo-pathologist. Gastroenterology 2006;131: Iwamoto Y, Itoyama T, Yasuda K, et al. Photodynamic DNA strand breaking activities of acridine compounds. Biological & Pharmaceutical Bulletin. 1993;16: Wallace M, Meining A, Miehlke S, et al. Safety of intravenous fluorescein for probe-based Confocal Laser Endomicroscopy (pcle): a multicenter study. Gastroenterology In press. 4. Pohl H, Roesch T, Vieth M, et al. Miniprobe confocal laser microscopy for the detection of invisible neoplasia in patients with Barrett s esophagus. Gut 2008;57: Rabeneck L, Paszat LF, Hilsden RJ, et al. Bleeding and perforation after outpatient colonoscopy and their risk factors in usual clinical practice. Gastroenterology 2008;135: e

5 6. Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer and the American Cancer Society. CA Cancer J Clin ;56: Buchner AM GM, Murli K, Wolfsen HC, et al. High resolution confocal endomicroscopy probe system for in vivo diagnosis of colorectal neoplasia [abstract]. Gastroenterology 2008;135: Kiesslich R, Goetz M, Lammersdorf K, et al. Chromoscopy-guided endomicroscopy increases the diagnostic yield of intraepithelial neoplasia in ulcerative colitis. Gastroenterology 2007;132: Watanabe O, Ando T, Maeda O, et al. Confocal endomicroscopy in patients with ulcerative colitis. J Gastroenterol Hepatol 2008; 23(suppl 2):S Trovato C, Sonzogni A, Fiori G, et al. Confocal laser endomicroscopy for the detection of mucosal changes in ileal pouch after restorative proctocolectomy. Dig Liver Dis 2008 Nov 12 [Epub ahead of print]. 11. van den Broek FJC, Fockens P, van Eeden S, et al. Endoscopic tri-modal imaging for surveillance in ulcerative colitis: randomised comparison of high-resolution endoscopy and autofluorescence imaging for neoplasia detection; and evaluation of narrowband imaging for classification of lesions. Gut 2008;57: Dekker E, van den Broek FJ, Reitsma JB, et al. Narrow-band imaging compared with conventional colonoscopy for the detection of dysplasia in patients with longstanding ulcerative colitis. Endoscopy 2007;39: Meining A, Schwendy S, Becker V, et al. In vivo histopathology of lymphocytic colitis. Gastrointest Endosc 2007;66: Zambelli A, Villanacci V, Buscarini E, et al. Collagenous colitis: a case series with confocal laser microscopy and histology correlation. Endoscopy 2008;40: Kiesslich R, Hoffman A, Goetz M, et al. In vivo diagnosis of collagenous colitis by confocal endomicroscopy. Gut 2006;55: Hsiung PL, Hardy J, Friedland S, et al. Detection of colonic dysplasia in vivo using a targeted heptapeptide and confocal microendoscopy. Nat Med 2008;14: Reprint requests Address requests for reprints to: Michael B. Wallace, MD, MPH, Professor of Medicine, Mayo Clinic, 4500 San Pablo Rd, Jacksonville, Florida Wallace.michael@mayo.edu. Acknowledgments The authors thank the faculty of the International Conference of Cellvizio Users (ICCU) whose lectures provided a basis for this review article: Alexander Meining, MD, PhD; Prateek Sharma, MD; Yang Chen, MD; Marc Giovannini, MD; David Carr-Locke, MD; Julian Abrams, MD; Tom Wang, MD, PhD; Greg Lauwers, MD; Anna Buchner, MD, PhD; Christopher Thompson, MD; Adam Slivka, MD; and Simon Lo, MD. Conflicts of interest Dr Wallace receives research funding from Mauna Kea Technologies but no consulting or speaker s bureau fees. 1513