INTRODUCTION — Barrett's esophagus (BE) is the most important risk factor for esophageal adenocarcinoma, the incidence of which has been rising rapidly over the past few decades. Standard endoscopic imaging provides little detail of the mucosal surface, making it impossible to distinguish specialized intestinal epithelium from gastric-type metaplasia or to recognize dysplastic epithelium.

Better imaging modalities have the potential to improve detection of BE and surveillance for dysplasia and cancer. Many new endoscopic techniques continue to be developed including magnification endoscopy, chromoendoscopy, optical coherence tomography, narrow band imaging, and autofluorescence endoscopy, but none is used routinely in clinical practice [1].

This topic review will summarize experience with autofluorescence endoscopy. General approaches to diagnosis and management of BE are discussed separately. (See "Barrett's esophagus: Epidemiology, clinical manifestations, and diagnosis" and "Barrett's esophagus: Surveillance and management".)

LIGHT-TISSUE INTERACTION AND AUTOFLUORESCENCE — Several interactions are possible when a photon of light (excitation light) comes into contact with tissue:

●The photon may be reflected immediately.

●It may pass through the tissue for a certain distance to be absorbed by chromophores (defined below) (figure 1).

●A small proportion of the excitation light causes a light-tissue interaction resulting in a change in its wavelength, a phenomenon known as autofluorescence.

Reflection and absorption are the main light-tissue interactions that are relevant for conventional endoscopy. Barrett's esophagus (BE) has a reddish appearance since most of the red light is reflected while most of the green and blue light is absorbed by chromophores in the mucosa and submucosa.

Chromophores — Chromophores absorb photons without emission of fluorescence. The chromophores can absorb both excitation and fluorescence light. The absorption depends strongly on the wavelength of the light and may significantly modify the in vivo autofluorescence spectrum observed at the tissue surface. The main chromophore in gastrointestinal tissues in the visible wavelength range (400 to 700 nm) is hemoglobin, in its oxygenated or deoxygenated form [2,3].

Fluorophores — Fluorophores are biological substances that emit fluorescent light (which has a long wavelength) when exposed to light of a short wavelength (usually ultraviolet or blue light). The most important fluorophores responsible for tissue autofluorescence include collagen, reduced-nicotinamide adenine dinucleotide (NADH), aromatic amino acids, and porphyrins [2]. Each group of fluorophores has characteristic ranges of optimal excitation and emission wavelengths, and the emission bandwidths of many different fluorophores are broad and overlapping [3]. Although several studies have investigated the utility of tissue autofluorescence in BE, the most relevant fluorophores responsible for autofluorescence in BE are unknown.

Autofluorescence for distinguishing tissue types — The autofluorescence spectra of normal, metaplastic, and dysplastic tissues differ because of variation of the types, concentrations, and microdistribution of the constituent fluorophores and chromophores. However, the extent to which the differences can accurately distinguish among these epithelial states is incompletely understood (see below).

No marked differences in the intensity and microdistribution of epithelial autofluorescence between nondysplastic BE and high-grade intraepithelial neoplasia (HGIN) was observed in one report evaluating autofluorescence in ex vivo specimens from patients with BE [4]. By contrast, autofluorescence is generally decreased in dysplastic BE when examined in vivo. This suggests that differences in autofluorescence are not caused by specific changes in endogenous fluorophores within the epithelium but probably by alterations in tissue architecture and hemoglobin content that influence autofluorescence when seen in vivo.

The decreased autofluorescence seen in vivo in patients with dysplastic BE is probably caused by three mechanisms [2]:

●Nuclei have little or no autofluorescence compared with cytoplasm, and dysplastic cells have an increase in the nuclear-to-cytoplasm ratio.

●Dysplasia may be accompanied by thickening of the mucosal layer, thereby preventing light from penetrating into the submucosal layer. Collagen, located in the submucosal layer, is the dominant fluorophore in the visible (400 to 700 nm) wavelength range in the gastrointestinal tract. Thus, increased mucosal thickness in early neoplastic lesions leads to attenuation in submucosal tissue autofluorescence.

●An increase in the local concentration of tissue hemoglobin (especially in the mucosal layer) occurs during neoplastic transformation, thereby attenuating autofluorescence.

Other studies have suggested that findings on autofluorescence combined with other characteristics could help improve detection of tissue types. In one such report, early neoplasia in autofluorescence-positive areas of BE was independently associated with autofluorescence intensity, a distance of less than 1 cm from the gastric folds, and a different appearance on white light endoscopy [5].

AUTOFLUORESCENCE SPECTROSCOPY AND ENDOSCOPY — Two in vivo autofluorescence-based endoscopic techniques have been investigated for detecting early neoplasia in Barrett's esophagus (BE):

●Light-induced fluorescence spectroscopy (LIFS)

●Autofluorescence endoscopy

Light-induced fluorescence spectroscopy — In LIFS, a small light probe is passed through the working channel of an endoscope and brought into close contact with an area of interest. The light probe consists of a central delivery fiber (connected to a light source, which is often a laser) and multiple surrounding detector fibers for collection of the fluorescent light (connected to a spectrograph for depiction of the fluorescent spectrum).

LIFS can accurately distinguish BE with high-grade intraepithelial neoplasia (HGIN) from nondysplastic BE [6-10]. However, an important drawback is that it only samples a small area of mucosa, making it impractical as a surveillance tool. A new development is the Optical Biopsy System, which uses an optical fiber integrated in a regular biopsy forceps allowing realtime spectroscopy during BE surveillance [11].

Autofluorescence endoscopy — Autofluorescence endoscopy incorporates a real-time wide-angle view that allows the endoscopist to switch back and forth between standard white light imaging and autofluorescence endoscopy [2].

Fiberoptic fluorescence endoscopy — The first prototype fluorescence endoscopic imaging device (LIFE-II system, Xillix Technologies Corp, Richmond, BC, Canada) consists of a fiberoptic endoscope with an attached camera module that permits fast switching between white light endoscopy and light-induced fluorescence endoscopy (LIFE). The LIFE mode is excited through a metal halide light source with a blue light bandpass filter (400 to 450 nm), and the system is equipped with two intensified charged coupled devices (CCDs) integrated in the camera module for detection of green (490 to 560 nm) and red (>630 nm) tissue autofluorescence, together with an image processor for displaying the autofluorescence image on a monitor. A composite image is formed, with dysplastic lesions displayed as a brick-red color against a green/cyan-colored normal background (picture 1) [12]. However, some reports noted only occasional recognition of HGIN within BE using LIFE and found that BE itself can produce false-positive red fluorescence [13,14].

Ultimately, randomized controlled trials were performed comparing fiberoptic LIFE with standard white light video endoscopy (SVE) [15,16]:

●One randomized trial compared a fiberoptic LIFE system with SVE in 187 patients with BE [16]. In the first phase of the trial, patients were assigned to be examined with LIFE or SVE. In this phase, LIFE did not show a significantly higher yield for detecting early neoplasia compared with SVE (12 versus 5 percent, p = 0.40). In addition, 11 of 19 lesions (58 percent) that contained early neoplasia were not suspicious on LIFE and were only detected on random biopsies.

In the second phase of the trial, patients underwent a second endoscopy with LIFE when they had previously undergone SVE and vice versa. Approximately one-third of the group failed to undergo a second endoscopy due to treatment of early neoplasia detected on the first endoscopy, declination of further participation, or noncompliance with the study protocol. Again, no significant difference in detection of early neoplasia was found between the fluorescence endoscopy and SVE.

●Another randomized crossover trial compared LIFE with SVE for the detection of early neoplasia in BE [15]. Forty-seven patients with known BE were randomly assigned to undergo either LIFE or SVE as the first procedure, followed by crossover to the other technique after four to six weeks. Each procedure was performed by one of two experienced endoscopists who were randomly allocated to either LIFE or SVE for each patient and were blinded to the findings of the other examination.

After both endoscopies, 5 patients showed no dysplasia, 10 were indefinite for dysplasia, 19 had low-grade intraepithelial neoplasia (LGIN), and 13 had HGIN or early carcinoma. Targeted biopsies taken with both LIFE and SVE each detected 8 of 13 patients (62 percent) with histologically confirmed HGIN/early carcinoma. When random biopsies were also considered, LIFE and SVE detected 69 and 85 percent of HGIN/early carcinoma, respectively.

These initial trials suggest that the prototype devices used thus far do not provide additional value in the detection of early neoplasia in BE. There are three possible explanations for this:

●First, the white light images from the fiberoptic fluorescence endoscopes were poorer in quality than with current video endoscopes, in part because of the use of fiber optics. As a result, subtle topographical mucosal features may have been overlooked (picture 2).

●Second, the fluorescence endoscopy imaging systems used have relied on either total autofluorescence or the combination of detecting green (490 to 560 nm) and red (>630 nm) autofluorescence. These algorithms may not be optimized for BE. Further refinements of emission bandwidths specifically selected for autofluorescence imaging in BE are in progress.

●Third, the endoscopists were highly experienced, so the impact of the additional information obtained by LIFE used adjunctively with SVE may be less than for less experienced individuals.

Video autofluorescence imaging — Video autofluorescence imaging (AFI) addresses most of the limitations of fiberoptic fluorescence endoscopy. AFI incorporates a high-resolution video endoscope and a fluorescence imaging modality [17]. A prototype system was developed that is based on a light source with sequential red-green-and-blue (RGB) illumination. White light from a Xenon lamp is passed through a rotary RGB filter that separates the white light into the colors red, green, and blue that are used to sequentially illuminate the mucosa. The red, green, and blue reflected light is detected separately by a monochromatic CCD placed at the tip of the endoscope, and the three images are integrated into a single high-quality color image by the video processor that is synchronized with the rotation speed of the RGB-filter.

The system uses a special high-resolution video endoscope (GIFQ240FY, Olympus, Tokyo, Japan) that has two high-quality CCDs at its tip: one for white light endoscopy as explained above and a second one for autofluorescence imaging. In the AFI-mode, the autofluorescence image is composed of three components: (1) total autofluorescence after blue light excitation, (2) green reflectance, and (3) red reflectance [17].

This system produces excellent white light images and high-quality autofluorescence images in contrast to earlier prototypes that used fiber optic endoscopes (picture 3) [17]. Nondysplastic BE appears green on AFI, whereas suspicious areas appear blue/violet (picture 4).

The first uncontrolled studies of AFI in patients with BE showed promising results for improving the detection of areas with high-grade intraepithelial neoplasia (HGIN) and early carcinoma in high-risk patients [17-19]. An important drawback of AFI in these studies was high false-positive rates, ranging between 40 and 80 percent [17-20]. However, studies suggest that by combining AFI with another state-of-the-art imaging technique (eg, narrow band imaging), this false-positive rate can be reduced to 10 to 26 percent [18-20]. (See "Barrett's esophagus: Evaluation with optical chromoscopy".)

An endoscopy system is commercially available that incorporates high-resolution white light endoscopy, AFI, and narrow band imaging: endoscopic tri-modal imaging (ETMI) (Lucera, Olympus, Tokyo, Japan). An international multicenter uncontrolled feasibility study with this system showed that ETMI increased the targeted detection of HGIN and early carcinoma from 53 to 90 percent [20].

Subsequently, two multicenter randomized crossover studies were performed comparing ETMI with standard white light video endoscopy (SVE) for the detection of early neoplasia in two different settings: patients with inconspicuous early neoplasia referred to a tertiary referral center for endoscopic treatment, and patients with confirmed low-grade intraepithelial neoplasia (LGIN) undergoing surveillance in a general practice setting [21,22]. In both studies, patients were assigned to undergo ETMI or SVE first and to subsequently undergo a second procedure with the other technique a few weeks later. White light endoscopy and AFI were used for the detection of suspicious lesions. Biopsy samples were taken from suspicious areas and at random.

●Tertiary referral setting – In 87 patients who underwent both procedures, ETMI detected significantly more patients with suspicious areas containing HGIN or early carcinoma compared with SVE (36 versus 24 patients) [21]. When comparing overall detection, including targeted detection and random sampling, there was a trend toward ETMI detecting more patients with HGIN and early carcinoma (46 versus 40 patients), but the difference did not reach statistical significance [21].

●General practice setting – In 99 patients, ETMI detected significantly more dysplastic lesions (including LGIN) than SVE (49 versus 29 lesions), but it did not detect significantly more lesions containing HGIN or early carcinoma (16 versus 14 lesions). In addition, when analyzed per patient (instead of per lesion), there was no significant difference in the overall detection of dysplasia between ETMI and SVE (60 versus 54 patients) [22].

A third-generation AFI was developed in order to reduce the false-positive rate by changing the algorithm. However, no difference in false-positive rate was found in a pilot study comparing the second-generation AFI used in ETMI and the third-generation AFI [23].

Other studies have demonstrated that the interobserver agreement between endoscopists for AFI results to be fair to moderate [19,24].

A meta-analysis of five studies with 371 patients with BE found that among the 211 patients undergoing surveillance, 39 (18 percent) had HGIN or early carcinoma [25]. In five patients, the HGIN or early carcinoma was only detected with AFI, increasing the diagnostic yield from 16 percent to 18 percent (incremental diagnostic yield of 2 percent). In addition, the use of AFI led to a change in endoscopic therapy in only 6 of the 371 patients (2 percent).

From these results, we conclude that AFI currently does not have a role in the routine detection of HGIN and early carcinoma in patients with BE given the low impact on diagnosis of HGIN and early cancer, as well as the low impact on therapeutic decision making. Overall, these results indicate that AFI and ETMI are not ready to replace the use of random biopsy protocols for the detection of early neoplasia in patients with BE.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Barrett's esophagus".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Barrett's esophagus (The Basics)")

●Beyond the Basics topics (see "Patient education: Barrett's esophagus (Beyond the Basics)")

SUMMARY — The role of autofluorescence imaging (AFI) in Barrett's esophagus (BE) has not yet been defined. Autofluorescence represents one of several new approaches being evaluated in the hope of improving detection and surveillance of BE. Autofluorescence may improve the detection of early neoplasia in BE by drawing the endoscopist's attention to abnormal areas, after which additional imaging techniques or targeted biopsies can be used. However, the clinical impact of detecting these areas with AFI appears to be low.