Advances in the Endoscopic Assessment of Inflammatory Bowel Diseases: Cooperation between Endoscopic and Pathologic Evaluations

High definition endoscopy

High definition or high resolution endoscopy presents signal images with 850,000 to 1 million pixels, while standard definition signals offer 100,000 to 400,000 pixels on an SD format [14,15]. High resolution endoscopy results in visualization of subtle mucosal details and improves the sensitivity and specificity of dysplastic lesion detection. Furthermore, it facilitates endoscopic resection by delineating borders of neoplastic lesions in IBD patients.

The majority of published data comes from non-IBD patients and found high definition endoscopy to be superior compared with conventional endoscopy. A retrospective study with 160 colonoscopies including long-standing (>7 years) colonic IBD patients demonstrated 2.21 greater likelihood (95% confidence interval [CI], 1.09 to 4.45) adjusted prevalence ratio of detecting any dysplastic lesions and 2.99 (95% CI, 1.16 to 7.79) of detecting dysplastic lesions on targeted biopsy with high definition colonoscopy compared to conventional endoscopy [16]. There was also a 3-fold higher neoplasia detection rate with high definition endoscopy when compared with standard definition endoscopy in IBD patients.

Chromoendoscopy

Chromoendoscopy is considered a cost-effective technique intended to enhance visualization of mucosal detail, submucosal vascular patterns, and lesion characterization. In particular, chromoendoscopy can facilitate the identification of flat lesions harboring intraepithelial neoplasia. With this, it can guide biopsies and reduces the number of biopsies. It is divided into dye-based and dye-less imaging techniques.

Dye-based chromoendoscopy has been used for over a decade and increases the rate of dysplastic lesion detection, especially in patients with long-standing IBD (Fig. 1A). In addition, Dye-based chromoendoscopy allows for improved assessment of disease severity and extent. Absorptive agents (e.g., Lugol solution, methylene blue, toludine blue, and cresyl violet), contrast agents (e.g., indigo carmine and acetic acid), agents for tattooing (e.g., India ink, Indocyanine green, and methylene blue), and reactive staining agents (e.g., congo red and phenol red) can be used in dye-based chromoendoscopy [17,18]. Several studies have shown the superiority of chromoendoscopy compared to conventional white light endoscopy. Dye-based chromoendoscopy has a moderate to high sensitivity for diagnosis, improved dysplasia detection, and prediction of mucosal change using magnification techniques (Fig. 1B). Two meta-analyses also demonstrated the superiority of targeted biopsy with dye-based chromoendoscopy in diagnosing and assessing mucosal ulcerations and dysplasia [19,20] while reducing the number of biopsies. Most recently, Soetikno et al. [20] included 665 patients from 6 studies and confirmed that the rate of detection of any dysplasia was approximately 9 times higher with dye-based chromoendoscopy with targeted biopsy than using white light endoscopy, with an 8.9 pooled odds ratio (95% CI, 3.4 to 23.0). When comparing the difference in the mean procedure time, dye-based chromoendoscopy is 10.9 minutes shorter than white light endoscopy, including the time spent on random biopsies.

Fig. 1.

Chromoendoscopy using indigocarmine (A) and combined with magnification technique (B) for colonic dysplasia in ulcerative colitis (Courtesy of Dr. Jeong-Sik Byeon at Asan Medical Center).

Dye-less chromoendoscopy is a novel imaging technology that allows for a detailed examination of both the mucosal surface and the mucosal vascular pattern by pushing a button on the handle of the endoscope, thereby enabling high-contrast imaging of the mucosal surface in real time without the use of special equipment. These dye-less chromoendoscopy techniques are divided into two types. One is an optical filter system including narrow band imaging (NBI) from Olympus, Tokyo, Japan and Compound Band Imaging from Aohua, Shanghai, China, and the other is digital chromoendoscopy with a post-processing system including i-Scan from Pentax, Tokyo, Japan and FICE (Fuji intelligent color enhancement from Fujinon, Tokyo, Japan) [21,22].

Optical chromoendoscopy techniques are based on optical lenses integrated within the light source of the endoscope, usually in front of the excitation white light source, to narrow the bandwidth in the blue and green regions of the spectrum [23,24]. In contrast, digital chromoendoscopy uses digital postprocessing of endoscopic images made in real-time by the video processor [25]. Recent studies indicate that dye-less chromoendoscopy, including optical and digital ones, are useful and practical for the differentiation of adenoma versus hyperplastic colon polyps and have good histological correlations [26-28].

Narrow band imaging

NBI is the most recognized among the virtual chromoendoscopy. This in vivo method uses optical filters in front of the light source to narrow the wavelength of the projected light to a 30 nm wide blue (415 nm) and green (540 nm) spectra, which enables visualization of micro-vessel morphological changes in superficial neoplastic lesions. NBI enhances the visibility of the small irregularities that accompany non-neoplastic inflammatory changes using the same logic as dye-based chromoendoscopy (Fig. 2).

Fig. 2.

Observation findings of colonic dysplasia using white light endoscopy (A), narrow band imaging technique (B), and autofluorescence imaging technique (C) in ulcerative colitis (Courtesy of Dr. Jeong-Sik Byeon at Asan Medical Center).

However, the role of NBI in detecting dysplasia in IBD remains somewhat uncertain due to conflicting results in the literature. A paper by East et al. [29] was the first to describe the use of NBI to distinguish dysplastic from nondysplastic mucosa in patients with longstanding. Subsequent to this case report, several randomized controlled studies have been published. Dekker et al. [30] demonstrated that NBI does not improve the detection rate of neoplasia in UC compared with high-definition white light endoscopy with a randomized crossover study of 42 patients. Of 11 patients with neoplastic lesions, four were detected with both modalities, four with NBI alone, and three with standard white light colonoscopy alone.

Two additional randomized trials comparing NBI to white light endoscopy also found no significant difference in the detection of neoplastic lesions. Random background biopsies were also ineffective in detecting dysplasia. According to Ignjatovic et al. [31], dysplasia detection was 9% in each arm and the yield of dysplasia detection from random nontargeted biopsies was 0.04%. Van den Broek et al. [32] found 13 of 16 neoplastic lesions (81%) using high definition-NBI compared with 11 of 16 neoplastic lesions (69%) using high definition -white light endoscopy. A study using a new-generation NBI system compared with dye based chromoendoscopy for the early detection of colitis-associated dysplasia and cancer in patients with longstanding colonic IBD demonstrated that NBI is less time-consuming (26.87±9.89 minutes vs 15.74±5.62 minutes, p<.01), but has no advantages over conventional endoscopy for the detection of intraepithelial neoplasia [33]. However, NBI has some advantages over dye-based chromoendoscopy, as it does not require additional dye agents and is easier to use in practice. These findings have led to controversy regarding the real role of NBI in dysplasia detection in IBD patients.

i-Scan

Currently, two virtual chromoendoscopy techniques are available, including FICE and i-Scan is a new endoscopic system using post processing light filter technology based on software algorithms with real time image mapping. It enhances different elements of the mucosa by three different image processes such as surface enhancement, tone enhancement, and contrast enhancement. Activation between different modes is done by pushing a button on the handle of the endoscope [34,35]. To date, most randomized trials have not shown that NBI or FICE can improve the detection of colorectal neoplasia when comparing colonoscopy with and without filter enhancement.

A randomized controlled study was conducted on 78 IBD patients in Germany to identify whether i-Scan has the potential to enhance assessment of disease severity and extent in mild or inactive IBD patients. The average duration of the examination for high definition—white light endoscopy and i-Scan groups was 18 and 20.5 minutes, respectively, but these differences were not statistically significant. When comparing the endoscopic prediction of inflammatory extent and activity with the histological results, there was overall agreement of 48.71% and 53.85% in the high definition—white light endoscopy group and 92.31% and 89.74% in the i-Scan group (p<.001 and p=.066) [36].

Patients with intestinal food allergy present with lymphoid hyperplasia, slight mucosal edema, and blurred mucosal vascular pattern in the colon. Based on this, an observational study reported on the potential of i-Scan for prediction of mucosal changes with suspected food allergy. Positive and negative predictive values for i-Scan to predict food allergy were 92% and 80%, respectively. Moreover, i-Scan predicted food allergy with a sensitivity, specificity, and accuracy of 85%, 89%, and 86%, respectively [37].

Confocal laser endomicroscopy

Observation and characterization of the colonic mucosal surface and abnormalities of blood vessel architecture are crucial in predicting histology, and this can be performed more efficiently with chromoendoscopy. However, histologic confirmation is needed to determine whether the presence of mucosal abnormalities is a result of IBD or not. This can be accomplished by confocal endomicroscopy in vivo, which may provide images similar to histologic findings in real time (Fig. 3). Endomicroscopy is regarded as optical biopsy that can achieve an image of the cellular structure of the mucosa with 1,500 fold magnification [38]. Currently, two endomicroscopy systems are available including an integrated endoscopy system (iCLE, Pentax) and a probe-based system (pCLE, Cellvizio, Mauna Kea Technologies, Paris, France). In vivo CLE uses an excitation wavelength of 488 nm with a single line laser; the laser power output is up to 1 mW at the tissue surface. Images are collected at a scan rate of 0.8 frames per second at a resolution of 1,024×1,024 pixels or 1.6 frames per second with 1,024×512 pixels [39]. It can capture the z axis which enables interrogation of the epithelium and lamina propria 0–250 mm below the surface layer [40]. The pCLEsystem uses a fixed laser power and a fixed image plane depth. The purpose of the system is to observe mucosal microarchitecture with an increased field of view (4×2 mm) through postpocessing with Cellvio Viewer (Fig. 4A). It enables virtual staining of mucosal structures to further enhance tissue contrast. The probe requires an accessory channel of 2.8 mm and has a resolution of 1 μm with a field of view of 240 μm and a fixed image plane depth varying between 55–65 μm (Fig. 4B).

Fig. 3.

Confocal laser endomicroscopic findings for normal mucosa (A) and mucosa in active ulcerative colitis (B). In ulcerative colitis, lamina propria widening, inflammatory infiltrates, goblet cell depletion, and crypt distortion are observed.

Fig. 4.

Cellvizio system for probe based confocal laser endomicroscopy (A) and a probe (B).

Crypt architecture, microvascular alterations, fluorescein leakage, and cellular infiltrates within the lamina propria are important observational markers in CLE evaluation [12,41-44]. CLE can aid in demonstrating mucosal healing in terms of deep remission beyond the absence of mucosal ulceration.

Watanabe et al. [45] investigated the features of CLE in the inflamed and noninflamed rectal mucosa of 17 UC patients and compared these results to standard histology. In this study, the crypts of colonic mucosa in active UC were large, variously shaped and irregular in arrangement. Numerous inflammatory cells and capillaries were visible in the lamina propria with CLE. Li et al. [42] also assessed crypt architecture, fluorescein leakage, and microvascular alterations in 73 consecutive UC patients and showed a correlation with histological results (p<.001). On post-CLE objective assessment, subjective architectural classifications were supported by the number of crypts per image (p<.001), but not fluorescein leakage results by gray scale (p=.194). Most recently, CLE also proved a sensitive tool in predicting UC relapse. In this study, 17 of 20 patients (85%) with histologically confirmed normal or chronic inflammation were diagnosed as having nonactive inflammation by real-time CLE. Twenty two of 23 patients (96%) with histologically confirmed acute inflammation were diagnosed as having active inflammation by CLE. The results of CLE were highly consistent with those of conventional histology (kappa value=0.812). Eleven percent of patients in the nonactive inflammation group relapsed, while 64% of patients in the active inflammation group relapsed. The relapse rate of patients with active inflammation was significantly higher than of those with nonactive inflammation (p<.001).

Neumann et al. [46] proposed the Crohn’s Disease Endomicroscopic Activity Score for assessing CD activity in vivo from comparison data between CD patients and a normal control group with standard white-light endoscopy followed by CLE. Active CD patients showed a higher proportion of increased colonic crypt tortuosity, enlarged crypt lumens, microerosions, augmented vascularization, and increased cellular infiltrates within the lamina propria. In the case of quiescent CD patients, there was a significant increase in crypt and goblet cell number compared with controls.

Autofluorescence imaging

Autofluorescence imaging (AFI) is a technique using the natural principle that cells contain molecules that become fluorescent when excited by UV/Vis radiation of a certain wavelength. Among the endogenous fluorophores, collagen and elastin have a relatively high quantum yield, so the extracellular matrix usually contributes to the autofluorescence emission more than cellular components. Autofluorescence imaging videoendoscopy produces real-time pseudo-color images based on tissue autofluorescence emitted by excitation of endogenous tissue fluorophores.

It is well known that cell and tissue state change resulting from modifications of the amount and distribution of endogenous fluorophores and the chemical-physical properties of their microenvironment during physiological and/or pathological processes. Therefore, AFI can be utilized in order to obtain information about the morphological and physiological state of cells and tissues (Fig. 2).

AFI has been used to highlight various lesions, such as neoplastic tissue, minimal changes in reflux esophagitis, the extent of chronic atrophic fundal gastritis, and Barrett’s esophagus [47-50]. AFI improves detection rates of neoplasia in patients with IBD and decreases the number of random biopsies needed. In a randomized, comparative study with 50 UC patients, neoplasia miss-rates for AFI and white light endoscopy were 0% and 50%, respectively (p=.036) [51]. AFI had 100% of sensitivity since all neoplasia was colored purple on AFI, while NBI had a 75% of sensitivity according to the Kudo classification.

AFI also has the ability to detect inflammatory lesions, including microscopic activity, in the colonic mucosa. Osada et al. [52] evaluated 572 images from 42 UC patients including white light endoscopy and AFI to validate the clinical relevance of AFI endoscopy for the assessment of the severity of inflammation. The green color component of AFI corresponded more closely with mucosal inflammation sites (r=–0.62, p<.01) than the red (r=0.52, p<.01) or blue (r=0.56, p<.01) color components. There were significant differences in green color components between limited (0.399±0.042) and extensive (0.375±0.044) (p=.014) polymorphonuclear cell infiltration within MES-0. It was observed that the green color component of AFI decreased as the severity of the mucosal inflammation increased.

Endocytoscopy

Endocytoscopy (Olympus, Tokyo, Japan) is a new technique, enabling observation of the gastrointestinal mucosa at the cellular level. Microscopic imaging for the gut mucosal layer can be observed at a magnification up to 1,400-fold with a contact light microscope [53]. It requires preparation of the mucosal layer with absorptive contrast agents like methylene blue or toluidine blue. Thus, endoscopists can distinguish architectural details such as epithelial structure, cellular features, and vascular patterns in terms of size, leakage, and tortuosity [54-56].

Some studies suggest that endocytoscopy has a potential role in in vivo evaluation. A study in patients who have colorectal aberrant crypt foci demonstrated that endocystoscopy was able to detect tissue abnormalities in the normal mucosa surrounding CRC and to identify neoplasia in aberrant crypt foci with 91.4% sensitivity [56]. A pilot study with IBD patients showed that endocytoscopy could reliably distinguish single inflammatory cells with high sensitivities and specificities (neutrophilic [60% and 95%], basophilic [74.43% and 94.44%], eosinophilic granulocytes [75% and 90.48%], and lymphocytes [88.89% and 93.33%]). It also showed that the concordance between endocytoscopy and histopathology for grading intestinal disease activity in IBD was 100% [54]. This new imaging technique introduces possibilities for the development of in vivo research while allowing surface magnification at cellular and subcellular resolution, but little data is currently available on endocytoscopy.