Probe-based confocal laser endomicroscopy: scientific toy or clinical tool?

The incidence of esophageal adenocarcinoma and associated death has risen significantly over the past decades in Western countries [1] [2]. This coincides with the pandemic of obesity and associated increase in the incidence of reflux disease [3]. In tertiary referral patients, Barrett’s esophagus can be found in 4.9 % of patients with atypical symptoms and in 4.5 % of patients with typical reflux symptoms [4].

Screening for neoplasia in Barrett’s esophagus is a difficult but worthwhile effort. First, the annual combined incidence of adenocarcinoma or high-grade intraepithelial neoplasia in patients with a nondysplastic Barrett’s esophagus is estimated to be 1.3 % [5]. This is comparable to the annual risk of colorectal cancer for which cancer screening programs with colonoscopy have been implemented in many countries [6]. Secondly, retrospective surgical series [7] [8] [9] [10] [11] [12] have shown that patients who are diagnosed with adenocarcinoma while under surveillance have a higher rate of early stage lesions (76 %) than patients who had their cancer diagnosed outside a Barrett’s surveillance setting (15 %). This is reflected in a 3-year survival of 80 % and 31 %, respectively [7]. Endoscopic surveillance comprises both thorough inspection for possible visible lesions and biopsies taken according to the Seattle protocol (four-quadrant biopsies every 2 cm). However, adherence to this protocol by gastroenterologists is generally poor [13] and it is often perceived as time consuming and uncomfortable for the patient.

Confocal laser endomicroscopy (CLE) is a new endoscopic imaging technique that allows in vivo imaging with a (sub)cellular resolution, enabling the endoscopist to perform a histologic assessment during ongoing endoscopy [14] [15] [16] [17] [18] [19] [20] [21]. Currently, two systems are available. In the first system, a confocal probe is incorporated into the tip of a conventional colonoscope (eCLE). The second system is a probe-based system (pCLE), which can be passed through the biopsy channel of any conventional endoscope. The basics for obtaining images are the same for both systems. The confocal unit is connected to a monochrome blue light laser source, necessary to excite a fluorophore. The most commonly used fluorescent dye is intravenously administered fluorescein. Images can be obtained easily by applying the confocal window of the probe or endoscope onto the mucosa once the fluorescent dye is administered. Intravenously applied fluorescein causes a transient discoloration of the skin and concerns have been raised about its safety as a diagnostic aid. A recent multicenter study assessing fluorescein-assisted pCLE showed, however, that no major adverse events occurred in 2272 procedures and that mild adverse events were recorded in 1.4 % of the procedures, including nausea, vomiting, transient hypotension, injections site erythema, rash, and epigastric pain [22].

With a lateral resolution of 0.7 µm and a magnification of × 1000, the eCLE system currently has the best resolution compared with pCLE, which has a lateral resolution of 1µm [23]. Although the absolute difference of 0.3 µm appears rather small, this constitutes a 43 % decrease in resolution, which is comparable to the difference between standard video endoscopy and high-definition endoscopy. The eCLE system also allows the adjustment of the imaging plane in the Z-axis, which makes it possible to scan at different depths and visualize different histologic structures, such as blood vessels. The pCLE system has a fixed imaging plane but its use is more versatile as it can be used with any endoscope, at any moment. In addition, the temporal resolution of the pCLE system is higher (12 images/second) compared with the pCLE (± 1 image/second).

Studies suggest that confocal endomicroscopy enables the endoscopist to differentiate gastric mucosa, specialized intestinal metaplasia, and neoplastic changes in Barrett’s esophagus with an apparently high diagnostic accuracy [14] [15] [16] [19] [20] [21].

In this issue of Endoscopy, Bajbouj et al. [24] present a multicenter study on the diagnostic accuracy of pCLE in screening for Barrett’s neoplasia in patients under surveillance or referred for suspicion of neoplasia. In total, 670 pCLE-matched biopsies were obtained from 68 patients. The authors studied the diagnostic accuracy of pCLE for bedside diagnosis and in a blinded fashion by scoring only the images without having the endoscopic image available, based on a previously validated classification algorithm (Pohl et al.). During immediate bedside interpretation, 12 % of the areas with neoplasia were found to be positive on pCLE (i. e. sensitivity) and 18 % of pCLE positive images had neoplasia detected in the biopsies (positive predictive value). Of the areas without neoplasia, 95 % were negative on pCLE (specificity) and 92 % of pCLE-negative images were indeed free of neoplasia in the biopsies (negative predictive value). Sensitivity and positive predictive value of the technique were improved during blinded interpretation of CLE images to 28 % and 46 %, respectively. The authors correctly conclude that due to the low sensitivity and low positive predictive value, it is currently too premature to replace the classic Seattle biopsy protocol with pCLE.

Compared with other studies on CLE in Barrett’s esophagus, this study has two unique features. This is the first study to address the accuracy of a new advanced imaging technique in a population with a relatively low prevalence of neoplasia (8.3 % of the biopsies and 16 % of the patients), hereby more resembling a standard Barrett’s surveillance population. Secondly, this is a multicenter study illustrating the performance of the technique in the hands of different investigators. This study illustrates very well the difficulty of evaluating the performance of new imaging techniques in the detection of Barrett’s neoplasia. Most studies in this field do not reflect everyday gastroenterology practice as they evaluate techniques in populations with a high prevalence of Barrett’s neoplasia [25] [26]. This high pre-test probability results in much higher sensitivity rates compared with the 12 % – 28 % reported by Bajbouj et al. [24]. Pohl et al. [20] used the same pCLE system in a Barrett’s population with a 24 % prevalence of neoplasia and found an apparently higher sensitivity of 75 %. In two studies investigating the diagnostic accuracy of eCLE the prevalence of neoplasia ranged between 24 % and 41 % [14] [19]. These studies used a different eCLE classification system and reported impressive diagnostic accuracy with a sensitivity of 92.9 % and specificity of 98.4 % [14].

Although pCLE and eCLE basically evaluate more or less the same morphologic features, the criteria to define and diagnose neoplastic changes show considerable differences. The former uses criteria primarily based on architectural changes such as irregular epithelial lining, fusion of glands, dark areas, and irregular vascular pattern. These criteria were developed by consensus of experts based on the unblinded evaluation of 15 patients [20] and were used in the current study by Bajbouj et al. [24]. The eCLE-based criteria were proposed in a pilot study in 63 patients and are based on architectural but also cellular changes. The presence of irregular capillaries with leakage of fluorescein and the presence of irregular glands with so-called ”black“ cells are considered diagnostic for Barrett’s neoplasia [14]. Importantly, the eCLE criteria have been confirmed in an independent second study [19], assessing the increase in diagnostic yield of targeted eCLE biopsies. The superior diagnostic accuracy reported with eCLE may thus reflect differences in patient selection and study design or the use of a different classification scheme, which in itself may be a reflection of the higher resolution of the eCLE and its ability to scan through the mucosa at different levels. Nevertheless, data on the clinical use of CLE in Barrett’s esophagus are scarce and larger studies, preferably in nonselected patients in a multicenter setting need to confirm either classification.

The paper by Bajbouj et al. [24] also nicely demonstrates that interpretation of CLE actually depends on the integration of the white-light endoscopic and microscopic images, as on a per patient basis, a significant decrease in specificity from 95 % to 59 % was observed when investigators were blinded to the endoscopic findings. Indirectly, this may indicate that ”the eye sees what the mind expects“ and that the high negative predictive value of pCLE in this study is based on seeing endoscopically normal-appearing mucosa in a known setting with a low risk of neoplasia: only 8.3 % of biopsies showed neoplasia and the high negative predictive value of pCLE (92 %) therefore is virtually identical to the pre-test likelihood of absence of neoplasia. In addition, of the eleven patients with neoplasia, eight had visible abnormalities that were detected by white-light endoscopy. This is further supported by other studies on the use of CLE in Barrett’s esophagus. In the pioneer study by Kiesslich et al. [14], 12 of the 15 patients with neoplasia had mucosal irregularities or focal lesions that could be detected by standard white-light endoscopy. Dunbar et al. [19] found visible lesions in 56 % of patients referred for suspicion of neoplasia and no neoplasia was found in patients just undergoing standard surveillance.

So is there a future for CLE as a clinically relevant imaging tool for Barrett’s surveillance? In our opinion three conditions have to be met to meet this goal: 1) inspection with CLE should significantly decrease the already low pre-test likelihood of neoplasia below a clinically acceptable threshold (e. g. < 1 %); 2) in the absence of neoplasia, CLE should also enable risk stratification comparable to current histology-driven decision making (e. g. the management of nondysplastic Barrett’s esophagus and Barrett’s esophagus with low-grade intraepithelial neoplasia significantly differs); and 3) CLE should reliably detect early neoplasia – it is not enough to rule out neoplasia if only a minority of neoplastic lesions are actually recognized as such. These three prerequisites are surely not met by the currently available studies.

Theoretically, CLE is able to reduce the number of biopsies, increase the diagnostic yield, and reduce costs associated with biopsy processing. The question remains, however, whether this is feasible in a standard surveillance setting with a low prevalence of neoplasia, where the main goal is to exclude neoplastic changes. It is conceivable based on the currently available data, that in the absence of any visible abnormality on high-definition endoscopy, a negative CLE mapping could confirm the absence of neoplasia. However, in order to achieve this, the sensitivity needs to be improved compared with the study by Bajbouj et al. So far studies addressing this issue are lacking, but larger multicenter trials are on the way. In the setting of the presence of visible lesions limited data are available that indicate that eCLE can be used as a clinical tool to confirm the presence of neoplasia [14] [19]. The study by Dunbar et al. showed that the diagnostic yield of biopsies could be significantly increased from 17.2 % to 33.7 % by only targeting CLE-suspicious sites [19]. The ultimate goal would eventually be to combine CLE with immediate therapy as illustrated in some case reports, where even flat neoplasia was resected on the spot after CLE diagnosis [15] [19]. However, more data are needed to justify this approach and to find out what the additional value of CLE might be compared with, for instance, autofluorescence and narrow-band imaging [26] [27].

Currently, CLE still has to be regarded as an interesting scientific toy that is under clinical evaluation. However, it does have great potential to become an important clinical tool in the assessment of Barrett’s esophagus, both in surveillance and in patients with suspicion of neoplasia. Further improvement of the technique, the combination with molecular imaging, and standardization of the interpretation of images may help to achieve this goal.

From a medical-legal viewpoint the involvement of pathologists will probably also be indispensable, in order to make them acquainted with the interpretation of these gray-scale histologic images that ultimately may replace classic standard histology. For the endoscopist who is more reluctant to become an endomicroscopist, the idea of just electronically sending CLE slides to the pathologist may sound very appealing and may help to turn this scientific toy into an indispensable clinical tool.

Competing interests: None

References

• 1 Pohl H, Welch H G. The role of overdiagnosis and reclassification in the marked increase of esophageal adenocarcinoma incidence.  J Natl Cancer Inst. 2005;  97 142-146
• 2 Eloubeidi M A, Mason A C, Desmond R A. et al . Temporal trends (1973 – 1997) in survival of patients with esophageal adenocarcinoma in the United States: a glimmer of hope?.  Am J Gastroenterol. 2003;  98 1627-1633
• 3 Jacobson B C, Somers S C, Fuchs C S. et al . Body-mass index and symptoms of gastroesophageal reflux in women.  N Engl J Med. 2006;  354 2340-2348
• 4 Poelmans J, Feenstra L, Demedts I. et al . The yield of upper gastrointestinal endoscopy in patients with suspected reflux-related chronic ear, nose, and throat symptoms.  Am J Gastroenterol. 2004;  99 1419-1426
• 5 Sharma P, Falk G W, Weston A P. et al . Dysplasia and cancer in a large multicenter cohort of patients with Barrett’s esophagus.  Clin Gastroenterol Hepatol. 2006;  4 566-572
• 6 Data from Surveillance, Epidemiology and End results (SEER) program, 1973 – 1992. 
• 7 Fountoulakis A, Zafirellis K D, Dolan K. et al . Effect of surveillance of Barrett’s oesophagus on the clinical outcome of oesophageal cancer.  Br J Surg. 2004;  91 997-1003
• 8 Streitz Jr. J M, Andrews Jr. C W, Ellis Jr. F H. Endoscopic surveillance of Barrett’s esophagus. Does it help?.  J Thorac Cardiovasc Surg. 1993;  105 383-387
• 9 Peters J H, Clark G W, Ireland A P. et al . Outcome of adenocarcinoma arising in Barrett’s esophagus in endoscopically surveyed and nonsurveyed patients.  J Thorac Cardiovasc Surg. 1994;  108 813-821
• 10 van Sandick J W, van Lanschot J J, Kuiken B W. et al . Impact of endoscopic biopsy surveillance of Barrett’s oesophagus on pathological stage and clinical outcome of Barrett’s carcinoma.  Gut. 1998;  43 216-222
• 11 Incarbone R, Bonavina L, Saino G. et al . Outcome of esophageal adenocarcinoma detected during endoscopic biopsy surveillance for Barrett’s esophagus.  Surg Endosc. 2002;  16 263-266
• 12 Ferguson M K, Durkin A. Long-term survival after esophagectomy for Barrett’s adenocarcinoma in endoscopically surveyed and nonsurveyed patients.  J Gastrointest Surg. 2002;  6 29-35
• 13 Peters F P, Curvers W L, Rosmolen W D. et al . Surveillance history of endoscopically treated patients with early Barrett’s neoplasia: nonadherence to the Seattle biopsy protocol leads to sampling error.  Dis Esophagus. 2008;  21 475-479
• 14 Kiesslich R, Gossner L, Goetz M. et al . In vivo histology of Barrett’s esophagus and associated neoplasia by confocal laser endomicroscopy.  Clin Gastroenterol Hepatol. 2006;  4 979-987
• 15 Leung K K, Maru D, Abraham S. et al . Optical EMR: confocal endomicroscopy-targeted EMR of focal high-grade dysplasia in Barrett’s esophagus.  Gastrointest Endosc. 2009;  69 170-172
• 16 Becker V, Vieth M, Bajbouj M. et al . Confocal laser scanning fluorescence microscopy for in vivo determination of microvessel density in Barrett’s esophagus.  Endoscopy. 2008;  40 888-891
• 17 Guo Y T, Li Y Q, Yu T. et al . Diagnosis of gastric intestinal metaplasia with confocal laser endomicroscopy in vivo: a prospective study.  Endoscopy. 2008;  40 547-553
• 18 Liu H, Li Y Q, Yu T. et al . Confocal laser endomicroscopy for superficial esophageal squamous cell carcinoma.  Endoscopy. 2009;  41 99-106
• 19 Dunbar K B, Okolo 3rd P, Montgomery E. et al . Confocal laser endomicroscopy in Barrett’s esophagus and endoscopically inapparent Barrett’s neoplasia: a prospective, randomized, double-blind, controlled, crossover trial.  Gastrointest Endosc. 2009;  70 645-654
• 20 Pohl H, Rosch T, Vieth M. et al . Miniprobe confocal laser microscopy for the detection of invisible neoplasia in patients with Barrett’s oesophagus.  Gut. 2008;  57 1648-1653
• 21 Wallace M B, Sharma P, Lightdale C. et al . Preliminary accuracy and interobserver agreement for the detection of intraepithelial neoplasia in Barrett’s esophagus with probe-based confocal laser endomicroscopy.  Gastrointest Endosc. 2010;  , . epub ahead of print
• 22 Wallace M B, Meining A, Canto M I. et al . The safety of intravenous fluorescein for confocal laser endomicroscopy in the gastrointestinal tract.  Aliment Pharmacol Ther. 2010;  31 548-552
• 23 Wallace M B, Fockens P. Probe-based confocal laser endomicroscopy.  Gastroenterology. 2009;  136 1509-1513
• 24 Bajbouj M, Vieth M, Rösch T. et al . Probe-based confocal laser endomicroscopy compared with standard four-quadrant biopsy for evaluation of neoplasia in Barrett’s esophagus.  Endoscopy. 2010;  42 435-440
• 25 Wolfsen H C, Crook J E, Krishna M. et al . Prospective, controlled tandem endoscopy study of narrow band imaging for dysplasia detection in Barrett’s Esophagus.  Gastroenterology. 2008;  135 24-31
• 26 Curvers W L, Singh R, Song L M. et al . Endoscopic tri-modal imaging for detection of early neoplasia in Barrett’s oesophagus: a multi-centre feasibility study using high-resolution endoscopy, autofluorescence imaging and narrow band imaging incorporated in one endoscopy system.  Gut. 2008;  57 167-172
• 27 Kara M A, Bergman J J. Autofluorescence imaging and narrow-band imaging for the detection of early neoplasia in patients with Barrett’s esophagus.  Endoscopy. 2006;  38 627-631