Real-time Histology in Colonoscopy

Gastrointestinal endoscopy had major technological improvements and novel technologies in recent years. High-definition endoscopy has permitted an increasingly detailed view of the mucosa during colonoscopy. Filter techniques that enhance analysis of vessel and surface structures. Autofluorescence imaging relies on functional imaging of tissue alterations. Endocytoscopy is an ultrahigh-contact microscopy procedure for cellular analysis of the epithelium. Endomicroscopy is an adaption of laser scanning microscopy for real-time intravital surface and subsurface microscopy during endoscopy. With these technologies, endoscopy has moved from prediction of histology based on morphologic patterns toward visualization of cellular and subcellular details, providing real-time histology.

Key points

•

High-definition white light endoscopy reveals an increasing amount of mucosal details.
•

Light filters (narrow-band imaging) or digital filters (i-scan, Fuji Intelligent Chromo Endoscopy) provide tools to support analysis of suspicious lesion by highlighting tissue and vessel patterns. Studies from expert centers suggest a good accuracy for differentiation of nonneoplastic from neoplastic lesions.
•

Endocytoscopy is an adaption of light microscopy that reveals cellular epithelial details after topical dye application, which may allow not only analysis of the surface of the lesion but also prediction of depth of invasion.
•

Endomicroscopy has been widely studied and permits a detailed analysis of the mucosal microarchitecture. Neoplastic lesions of the colorectum can be visualized with high accuracy.
•

For all these techniques, thorough training and ongoing evaluation in clinical trials should be sought to further corroborate their value in gastrointestinal endoscopy.

Introduction

Twenty years ago, studies showed for the first time that histology could be predicted during endoscopy by magnification chromoendoscopy with a high degree of confidence based on pit pattern analysis. Chromoendoscopy is covered elsewhere in this issue, but it has set the stage for the aim of the endoscopist “to establish an immediate endoscopic diagnosis that is virtually consistent with the histologic diagnosis.” Such analysis by magnifying chromoendoscopy is based on the morphology of crypts ( Fig. 1 ), but not on cellular or microarchitectural imaging. It therefore comes close to virtual histology, but still relies on prediction rather than visualization of ultrastructural changes. The recent introduction of high-definition (HD) endoscopy with virtual chromoendoscopy has similarly revealed many fine details of colonic lesions during ongoing endoscopy that can be used for immediate decision making on therapeutic strategies. Autofluorescence imaging (AFI) follows a different path and tries to predict the nature of a lesion on a molecular or functional basis more than on morphologic grounds. Endocytoscopy (EC) is an adaption of white light microscopy that enables intravital contact microscopy at the mucosal surface with a high degree of accuracy. Endomicroscopy uses miniaturized laser scanning devices for fluorescent cellular imaging on, but also below, the mucosal surface and can even provide visualization of molecular changes to characterize colonic lesions. The intention of all these techniques is to provide an on-site analysis of histology that should translate into an immediate, tissue-based therapeutic decision.

Introduction

Twenty years ago, studies showed for the first time that histology could be predicted during endoscopy by magnification chromoendoscopy with a high degree of confidence based on pit pattern analysis. Chromoendoscopy is covered elsewhere in this issue, but it has set the stage for the aim of the endoscopist “to establish an immediate endoscopic diagnosis that is virtually consistent with the histologic diagnosis.” Such analysis by magnifying chromoendoscopy is based on the morphology of crypts ( Fig. 1 ), but not on cellular or microarchitectural imaging. It therefore comes close to virtual histology, but still relies on prediction rather than visualization of ultrastructural changes. The recent introduction of high-definition (HD) endoscopy with virtual chromoendoscopy has similarly revealed many fine details of colonic lesions during ongoing endoscopy that can be used for immediate decision making on therapeutic strategies. Autofluorescence imaging (AFI) follows a different path and tries to predict the nature of a lesion on a molecular or functional basis more than on morphologic grounds. Endocytoscopy (EC) is an adaption of white light microscopy that enables intravital contact microscopy at the mucosal surface with a high degree of accuracy. Endomicroscopy uses miniaturized laser scanning devices for fluorescent cellular imaging on, but also below, the mucosal surface and can even provide visualization of molecular changes to characterize colonic lesions. The intention of all these techniques is to provide an on-site analysis of histology that should translate into an immediate, tissue-based therapeutic decision.

HD endoscopy and virtual chromoendoscopy

HD white light endoscopy (WLE) became possible after introduction of technical improvements from digital broadcasting to endoscopy. In digital imaging, resolution is a function of pixel density, and is improved by incorporating high–pixel density charge-coupled device (CCD) chips. Although standard-definition endoscopes incorporated approximately 200,000 to 400,000 pixels on 576 lines, HD endoscopes provide more than 2 × 10 ⁶pixels on 1080 lines, which results in visualization of subtle mucosal details ( Fig. 2 ).

Magnification or zoom endoscopy is obtained by a user-controlled movable lens in the endoscope tip and provides a continuously adaptable degree of magnification from 1.5-fold to 160-fold (see Fig. 1 ). This facility is different from electronic zooms, which enlarge the image without increasing pixel density at the trade-off of lower resolution despite software-based interpolation algorithms. HD endoscopes offer filter techniques (virtual chromoendoscopy) through the switch of a button on the endoscope, such as narrow-band imaging (NBI; Olympus, Japan), i-scan (Pentax, Japan), and Fuji Intelligent Chromo Endoscopy (FICE; Fujinon, Japan).

NBI

Although WLE uses the complete spectrum of visible light, this spectrum is narrowed around the blue and green band by a rotating filter in NBI. Because blue light is absorbed by hemoglobin, the false colored NBI image pronounces the mucosal vessel structure (in contrast with real chromoendoscopy, which highlights pit patterns). Initial hopes that NBI enhances adenoma detection rates in screening colonoscopy have not been fulfilled. Following these early trials, several classification systems have been developed for the use of NBI to predict histology (reviewed in Ref. ). These systems have recently been unified in the joint Japanese-European-American NBI International Colorectal Endoscopic (NICE) classification. This classification can be used with magnification or without, if close observation is accomplished with HD endoscopes. The NICE classification is based on color, vessel, and surface pattern. Type 1 shows the same or lighter color than the background, none or isolated lace vessels on the surface, and uniform surface pattern, and probably corresponds with a hyperplastic lesion. Type 2 is browner than the background (its color should arise from the vessels) and has thick brown vessels surrounding oval, tubular, or branched white structures (probable diagnosis is adenoma). Type 3 shows a (dark) brown color, areas with distorted or missing vessels, and areas of distorted or absent pattern, and indicates deep submucosal-invasive cancer.

Early trials have reported impressive accuracy rates of NBI for characterization of polyps with and without magnification, when high-confidence diagnoses were considered. Those trials were performed by expert endoscopists. However, accuracy rates have decreased in some trials, when endoscopists with less experience with in vivo diagnosis were asked to predict histology in vivo: in a Dutch multicenter trial at 6 nonacademic centers, accuracy rates were 75%. Thirteen US community gastroenterologists had 81% in vivo accuracy rates for diminutive adenomas. Considerable variation between different gastroenterologists and different batches of lesions within the study period was observed. Lower accuracy rates for in vivo (vs ex vivo) characterization were noted, potentially pointing to the importance of technical issues during colonoscopy, such as good visualization of the lesion for thorough on-site analysis.

FICE and i-scan

FICE and i-scan and use software-based post–image acquisition algorithms to display the endoscopic image in false colors. Different modalities either enhance the tissue surface or the vessel structure through the push of a button ( Fig. 3 ). Similar to NBI, virtual chromoendoscopy with digital filters has mainly contributed to better polyp characterization, whereas mucosal contrast enhancement in combination with HD imaging resulted in higher adenoma detection rates in some trials. In an expert center pilot study, i-scan with electronic filters, but without optical zoom, performed as well as chromoendoscopy for characterization of diminutive polyps of the rectum. In a follow-up trial with patients screened by colonoscopy, i-scan showed an accuracy rate of 98% to differentiate nonneoplastic from neoplastic lesions. Similar results were obtained for FICE: adenomas were correctly classified with 93% sensitivity and 61% specificity. This result was not significantly different from results obtained for chromoendoscopy in this study.