Detection of high-grade dysplasia in Barrett\'s esophagus by 5-aminolevulinic acid (ALA) induced protoporphyrin ix (PpIX) fluorescence spectroscopy

Detection of high-grade dysplasia in Barrett’s esophagus by spectroscopy measurement of 5-aminolevulinic acidinduced protoporphyrin IX fluorescence Stephan Brand, MD, Thomas D. Wang, MD, PhD, Kevin T. Schomacker, PhD, John M. Poneros, MD, Gregory Y. Lauwers, MD, Carolyn C. Compton, MD, PhD, Marcos C. Pedrosa, MD, MPH, Norman S. Nishioka, MD Boston, Massachusetts

Background: Preliminary studies with qualitative detection methods suggest that 5-aminolevulinic acid-induced protoporphyrin IX fluorescence might improve the detection of dysplastic Barrett’s epithelium. This study used quantitative methods to determine whether aminolevulinic acidinduced protoporphyrin IX fluorescence can differentiate between Barrett’s mucosa with and without dysplasia. Methods: Patients were given 10 mg/kg of aminolevulinic acid orally 3 hours before endoscopy. Quantitative fluorescence spectra were acquired by using a nitrogen-pumped dye laser (λ 400 nm) spectrograph system. The protoporphyrin IX fluorescence intensity at 635 nm was compared with the histopathologic diagnosis for mucosal biopsy specimens taken immediately after the fluorescence measurements. Results: Ninety-seven spectra were obtained from 20 patients. The mean (± standard error) standardized protoporphyrin IX fluorescence intensity was significantly greater (p < 0.05) for high-grade dysplastic Barrett’s epithelium (0.29 ± 0.07, n = 13) than for nondysplastic Barrett’s epithelium (0.11 ± 0.02, n = 43). By using protoporphyrin IX fluorescence alone, high-grade dysplasia was distinguished from nondysplastic tissue types with 77% sensitivity and 71% specificity. Decreased autofluorescence was particularly found in nodular high-grade dysplasia. By using the fluorescence intensity ratio of 635 nm/480 nm, nodular high-grade dysplasia could be differentiated from nondysplastic tissue with 100% sensitivity and 100% specificity. Conclusion: Protoporphyrin IX fluorescence may be useful for identifying areas of high-grade dysplasia in Barrett’s esophagus and for targeting of biopsies. (Gastrointest Endosc 2002;56:479-87.)

Barrett’s esophagus is a major risk factor for the development of esophageal adenocarcinoma. It is estimated that patients with Barrett’s esophagus are 30 to 125 times more likely to have esophageal adenocarcinoma develop compared with the general population. Although the exact cancer risk of Barrett’s esophagus is still controversial,1,2 the estimated incidence of adenocarcinoma in patients with Barrett’s esophagus is between 0.2% to 2.1% Received December 21, 2001. For revision February 20, 2002. Accepted April 28, 2002. Current affiliations: Gastrointestinal Unit, Wellman Laboratories of Photomedicine, and Department of Pathology, Massachusetts General Hospital, Harvard Medical School, and Section of Gastroenterology, VA Boston Healthcare System, Boston University School of Medicine, Boston, Massachusetts. Presented in part as an oral presentation at the annual meeting of the American Gastroenterological Association, May 21-24, 2000, San Diego, California (Gastroenterology 2000;118:A193). Dr. Poneros was supported by an NIH Training grant (T32 DK07191). Reprint requests: Norman S. Nishioka, MD, Massachusetts General Hospital, Wellman Laboratories of Photomedicine, BAR703, 50 Blossom St., Boston, MA 02114. Copyright © 2002 by the American Society for Gastrointestinal Endoscopy 0016-5107/2002/$35.00 + 0 37/1/128172 doi:10.1067/mge.2002.128172 VOLUME 56, NO. 4, 2002

per year.3-5 In addition, the incidence of adenocarcinoma of the esophagus and gastric cardia has increased more rapidly during the last 2 decades than any other cancer in the United States.6,7 The dramatic rise in the incidence of esophageal adenocarcinoma has increased awareness of Barrett’s esophagus as a preneoplastic lesion. Current surveillance strategies that typically specify 4-quadrant biopsies per 2 cm length of Barrett’s epithelium are expensive, time consuming, and limited by sampling error. Because early detection of dysplasia, especially high-grade dysplasia (HGD), could reduce the mortality from esophageal adenocarcinoma,8 better techniques for detecting dysplasia within Barrett’s epithelium are needed. Preliminary studies suggest that protoporphyrin IX (PpIX) fluorescence resulting from exogenously administered 5-aminolevulinic acid (5-ALA) may improve the detection of dysplastic mucosa in the GI tract.9-13 ALA-induced PpIX fluorescence has also been used successfully as a marker for dysplasia in other organ systems including the bladder,14,15 oral mucosa,16 skin,17 lungs,18 brain,19 and endometrium.20 ALA is a natural precursor of heme, which induces the formation of endogenous PpIX. The GASTROINTESTINAL ENDOSCOPY

479

S Brand, T Wang, K Schomacker, et al.

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

administration of exogenous ALA results in the accumulation of PpIX in tissue due to feedback inhibition of the final step of the heme biosynthetic cycle. Enzymatic differences in dysplastic tissue (e.g., decreased ferrochelatase activity) can lead to an increase in PpIX fluorescence in certain organs.21,22 The preferential concentration of PpIX in dysplastic cells has also been used for therapeutic purposes (photodynamic therapy).23-28 To date, diagnostic studies in patients with Barrett’s esophagus with ALA have been limited by the qualitative nature of the method used to assess fluorescence.13 In most studies an image-intensified endoscope was used to identify areas of PpIX accumulation. However, this technique is subject to numerous errors including (1) the inability to separate autofluorescence from PpIX fluorescence, (2) nonuniform optical excitation of tissue fluorescence, (3) distortion by the imaging system, and (4) sensitivity to the angle at which the tissue is viewed. Before concluding that dysplastic Barrett’s tissue produces more ALA-induced PpIX fluorescence than nondysplastic tissue, quantitative fluorescence measurements of tissue are required. The aim of this study was to quantitate PpIX fluorescence in Barrett’s epithelium with and without dysplasia. PATIENTS AND METHODS Patients Patients with histopathologically confirmed Barrett’s esophagus undergoing routine surveillance endoscopy were asked to participate. All patients were interviewed and underwent physical examination before enrollment. Informed consent was obtained from all patients. A complete blood cell count and biochemical parameters of liver function were obtained. Patients with a history of liver disease, elevated serum levels of liver enzyme, porphyria, or hypersensitivity to porphyrins were excluded. A solution of 10 mg/kg of 5-ALA powder (DUSA Pharmaceuticals, Inc., Valhalla, N.Y.) dissolved in 10 mL of orange juice was given orally 3 hours before endoscopy. Twenty-four hours after the endoscopy, biochemical tests of liver function were obtained to assess the influence of ALA on serum transaminases. Patients with elevated liver enzymes had these measured weekly until the enzyme levels returned to normal. All patients were instructed to avoid direct exposure of the skin to sunlight for 24 hours after ingestion of ALA. The protocol and informed consent form were approved by the human research committees of our hospitals. The use of ALA in this study was approved by the Food and Drug Administration. Endoscopic fluorescence imaging All patients underwent conventional endoscopy. In addition, fluorescence endoscopy was performed using a blue light source (D-light. Karl Storz, Tuttlingen, Germany) and a fluorescence imaging system (SAFE-1000, Pentax, Tokyo, Japan). White light illumination for conventional endos480

GASTROINTESTINAL ENDOSCOPY

copy was provided by a standard Xenon light source (Karl Storz). For fluorescence excitation, a bandpass filter with a spectral transmission range between 350 and 450 nm was rotated in front of the white light source. Both the white and fluorescence excitation light were delivered to tissue through the illumination fibers of the endoscope. A prism located at the proximal end of the fiberoptic endoscope deflected the white light image to a color CCD camera (KS162, Panasonic, Osaka, Japan). During fluorescence imaging, the prism was moved so that the endoscopic image was directed through a bandpass filter (spectral range 600-700 nm) and then to an intensified camera (C3510, Hamamatsu Photonics, Hamamatsu, Japan). The resulting gray scale image was displayed on the green channel of an RGB monitor. The green channel was used because the human eye is most sensitive at green wavelengths. The entire procedure was recorded on S-VHS videotape. Endoscopic laser-induced fluorescence spectroscopy Spectroscopic measurements were obtained immediately before taking a biopsy specimen from mucosal areas that by their appearance raised a suspicion of dysplasia by white light (e.g., nodules) or fluorescence (e.g., increased fluorescence intensity) endoscopy. In patients without endoscopically evident findings suggestive of dysplasia within the Barrett’s epithelium, 4-quadrant biopsy specimens were taken at 2-cm intervals, but fluorescence measurements were made at only 1 site per level. Quantitative fluorescence measurements were made with a laser-induced fluorescence (LIF) spectrometer. Fluorescence excitation was accomplished with 400-nm light from a nitrogen pumped dye laser (models VSL337ND and DLM220, Laser Sciences Inc., Franklin, Mass.). The 400 nm laser pulses (3 ns in duration) were coupled by means of a dichroic mirror (08MLQ003/413, Melles Griot, Irvine Calif.) into a single 0.6-mm core diameter optical fiber (SFS600/660N, Fiberguide Industries, Stirling, N.J.) that was used to deliver and collect light from tissue. The optical fiber was passed through the accessory channel of the endoscope and brought into gentle contact with the mucosal surface. Immediately after spectral acquisition, the probe was removed and a slight demarcation remained on the tissue for 30 to 60 seconds as a result of contact made by the probe. This endoscopically apparent indentation was used as a guide for taking a biopsy specimen from the same site at which spectra were acquired. Repeated spectroscopic measurements on the same mucosal area showed a good reproducibility. In all cases, the first spectroscopic measurement performed was used in the data analysis. Excessive backscattered light was removed by a filter (LL450, Corion Inc., Holliston, Mass.). The fluorescence signal was spectrally dispersed by a spectrograph (MonospeciS, Thermo Jarrell Ash Corp., Waltham, Mass.) and detected with a gated, intensified photodiode array (model 1456B-990-HQ, EG&G Princeton Applied Research, Princeton, N.J.) and optical multichannel analyzer system (model 1471A, EG&G Princeton Applied Research, Princeton, N.J.). The spectrograph was equipped with a 25µm wide input slit and a 150-grooves/mm grating blazed at VOLUME 56, NO. 4, 2002

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

S Brand, T Wang, K Schomacker, et al.

450 nm. A 1-µs gate width was used. Ten laser pulses were averaged for each spectrum. Emission spectra from 450 nm to 800 nm were measured and stored on a personal computer by running OMA-Vision spectroscopy software (EG&G Princeton Applied Research, Princeton, N.J.). Overall system response A background spectrum (fiber in air) was obtained and subtracted from the tissue spectra. All spectra were corrected for system response by dividing the observed spectra by a system response curve that was determined by measuring the spectral intensity of a calibrated 1 kW quartz-halogen lamp (model 200A, Optronic Laboratories Inc., Orlando, Fla.). Day-to-day variations in laser energy and fiber-coupling efficiency were accounted for by measuring the fluorescence of a laser dye standard with a peak optical density of 0.25 (DCM, Exciton Inc., Dayton, Ohio) before the procedure in all patients. All spectra were calibrated by dividing the intensity of the measured spectrum by the maximum intensity value obtained for the laser dye.

A

Histopathologic evaluation All biopsy specimens were fixed in formalin, embedded in paraffin, and sectioned. Tissue sections were stained with hematoxylin and eosin and histologically classified into one of the following categories: HGD, low-grade dysplasia (LGD), indefinite for dysplasia, nondysplastic Barrett’s epithelium, gastric tissue, cardia type of tissue, and squamous epithelium. Dysplasia was diagnosed by using established criteria.29 All biopsies were also histopathologically classified by the grade of inflammation with the following categories: 1, no inflammation; 2, mild inflammation; 3, moderate inflammation; 4, severe inflammation. Two independent senior GI pathologists reached consensus on all of the histopathologic diagnoses of dysplasia. The pathologists were blinded to the results of the spectroscopic measurements. Data analysis Each fluorescence spectrum contains fluorescence emissions due to autofluorescence (broad peak at approximately 480 nm) as well as PpIX fluorescence, which is characterized by peaks at 635 nm and 705 nm. Because there is a direct correlation between the 635 nm peak and the 705 nm peak, only the 635 nm peak was used for the PpIX fluorescence calculations. The autofluorescence contribution to the 635 nm peak PpIX fluorescence intensity was corrected for using a previously described method17 with minor modifications. The resulting value was referred to as the standardized PpIX fluorescence intensity (I 635-standard) and was calculated as follows (Fig. 1): 1.

I

635-standard

= I635 – I750 – Y

in which I635 is the fluorescence intensity at 635 nm and I750 is the fluorescence intensity at 750 nm. Y is calculated as follows: 2.

Y = (750 – 635)*X/(750 – 600)

VOLUME 56, NO. 4, 2002

B Figure 1. A, B, Spectra illustrating calculation of autofluorescence contribution to the 635 nm peak PpIX fluorescence intensity. The calculation is as follows: I 635-standard = I635 – (0.7666 *I600) – (0.2333*I 750) (See text for details).

3.

X = I600 – I750

By using equations 1, 2, and 3, I plified to: 4.

I

635-standard

635-standard

can be sim-

= I635 – (0.7666*I600) (0.2333* I750)

For each histopathologic category the mean value of I 635-standard was calculated and compared with the mean value of I 635-standard for the HGD category. Significance was determined with the student t test. To determine the effect of inflammation on PpIX fluorescence a multivariate regression analysis using the “least squares” method was performed. The inflammation score and the histopathology score were used as independent variables and the PpIX fluorescence value as dependent variable.

RESULTS Spectroscopic measurements Twenty patients (18 men, 2 women; mean age 67.8 years, range 31-81 years) with histopathologically confirmed Barrett’s esophagus undergoing surveillance endoscopy were studied. Spectroscopic measurements were made at 97 mucosal sites. Histopathologic analysis revealed HGD in 13 sites, GASTROINTESTINAL ENDOSCOPY

481

S Brand, T Wang, K Schomacker, et al.

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

A

A

B

B Figure 2. A, Mean standardized PpIX fluorescence (corrected for tissue autofluorescence) measured spectroscopically at 635 nm and mean standard error for each histologic category. B, Box plots show minimum, maximum, median, 25th, and 75th percentile of standardized PpIX fluorescence. Mean standardized PpIX fluorescence in HGD is greater than in nondysplastic tissue.

LGD in 9, nondysplastic Barrett’s esophagus in 43 (including 12 sites of indefinite for dysplasia), cardia tissue in 24, and gastric tissue in 8 sites. The cardia and gastric samples were included in the analysis because such samples are frequently found among samples gathered in patients with Barrett’s esophagus. The 13 biopsy specimens with HGD were taken from 5 different patients. As shown in Figure 2, the mean standardized PpIX fluorescence intensity (I 635-standard) increased with increasing degrees of dysplasia. The PpIX fluorescence at 635 nm was significantly greater in HGD than in nondysplastic Barrett’s epithelium, cardia, and gastric tissue (p < 0.05). By using linear discriminant analysis with I 635-standard as the only variable, a fluorescence intensity threshold of 0.117 was identified as differentiating sites with HGD from nondysplastic sites with high sensitivity and specificity. With this threshold, 10 of 13 samples (77%) with HGD were correctly classified and 60 of 84 (71%) of biopsy spec482

GASTROINTESTINAL ENDOSCOPY

Figure 3. HGD is associated with decreased tissue autofluorescence. Comparison of mean fluorescence intensity values of Barrett’s esophagus with HGD (n = 13, black line) and without HGD or LGD (n = 43, gray line). A, Absolute standardized fluorescence intensity values. B, Fluorescence intensity values normalized to PpIX peak value. There is an overlap of both curves between the two PpIX peaks at 635 and 705 nm because of the normalization process. Left y-axis for HGD, right y-axis for nondysplastic Barrett’s esophagus.

imens without HGD were correctly classified as nondysplastic. This classification threshold correctly classified HGD in 4 of 5 patients with HGD (80%). All patients with more than one site of HGD were correctly classified. The patient incorrectly classified by this method had only a single focus of HGD. Five of the 13 specimens with HGD were taken from endoscopically visible nodules whereas the other 8 specimens with HGD were from non-nodular, unremarkable Barrett’s mucosa. All specimens with nodular HGD (100%) were correctly classified by using fluorescence endoscopy and 5 of 8 non-nodular areas of HGD were identified by fluorescence endoscopy (56%). All nodular areas with HGD exhibited increased PpIX fluorescence and were correctly classified as premalignant by using the established fluorescence threshold of 0.117, whereas 5 of the 8 areas with flat HGD were correctly classified by this method. Two of the 5 areas with flat HGD were obtained by taking random biopsies and would therefore have been identified by 4-quadrant biopsy. Three areas of flat HGD were outside the random biopsy sampling areas and were only diagnosed by fluorescence endoscopy and fluorescence spectroscopy (Table 1). VOLUME 56, NO. 4, 2002

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

S Brand, T Wang, K Schomacker, et al.

Table 1. Comparison of conventional and fluorescence endoscopy for the detection of high-grade dysplasia in patients with Barrett’s esophagus HGD diagnosed by

Nodular HGD Flat HGD Total HGD detected by single method Total HGD detected by combined methods: conventional vs. fluorescence endoscopy

White light endoscopy

Random biopsy sampling

Fluorescence endoscopy

Fluorescence spectroscopy

5/5 (100%) 0.8 (0%) 5/13 (38%)

0/5 (0%) 5/8* (62%) 5/13 (38%)

5/5 (100%) 5/8* (62%) 10/13 (77%)

5/5 (100%) 5/8* (62%) 10/13 (77%)

10/13 (77%)

10/13 (77%)

*Two of the 5 areas with flat HGD diagnosed by fluorescence endoscopy overlapped with random biopsies. This means that these areas have been identified first by fluorescence endoscopy but they were within the area of a (random) 4-quadrant biopsy.

Multiple regression analysis including the possibility of an effect of inflammation on PpIX fluorescence was performed. In this analysis PpIX fluorescence intensity was dependent on the grade of dysplasia (p < 0.0001) but not dependent on the inflammation score (p = 0.31). However, the majority of specimens exhibited no inflammation (n = 34) or only mild inflammation (n = 49). Twelve specimens contained areas with moderate inflammation and only 2 showed signs of marked inflammation. Interestingly, the 2 specimens with the highest inflammation score (1 indefinite for dysplasia and 1 cardia tissue) had the greatest PpIX fluorescence in their tissue category. In addition to the increased PpIX fluorescence, HGD was characterized by a decreased autofluorescence peak at 480 nm when compared with nondysplastic Barrett’s esophagus. Although the absolute difference in the autofluorescence peak at 480 nm was small (Fig. 3A), normalization of the spectra to the PpIX peak fluorescence revealed that the relative autofluorescence peak of nondysplastic Barrett’s epithelium was 4 times higher than that with HGD (Fig. 3B). The quotient of fluorescence intensity at 635 nm divided by fluorescence intensity at 480 nm in Barrett’s epithelium with HGD was significantly greater than in all other tissue types (p < 0.05; Fig. 4). However, nodules with HGD exhibited much lower autofluorescence than areas with HGD outside of nodules. The 5 biopsy specimens with the highest overall 635mn/480nm quotient were all specimens with HGD within nodules and had the lowest autofluorescence intensity of all measured HGD areas. Therefore, by using a standardized fluorescence intensity threshold ratio of 44.3, nodular HGD could be differentiated from all other tissue types with 100% sensitivity and 100% specificity. VOLUME 56, NO. 4, 2002

Fluorescence imaging The correlation between the quantitative fluorescence measurement and the appearance on fluorescence endoscopy was poor. Compared with surrounding areas, increased fluorescence was seen in all HGD sites correctly diagnosed by spectroscopy alone. However, in many instances a high background autofluorescence level diminished the contrast between areas of increased and decreased PpIX fluorescence. Fluorescence imaging was sensitive to the angle at which the tissue was viewed and resulted in false positive images with the fluorescence imaging system. Side effects Systemic administration of ALA has been associated with side effects such as skin photosensitivity (limited to 24 hours), nausea, vomiting and transient increase in liver enzymes. In the present study, the most common side effect was nausea, observed in 6 patients (30%), followed by vomiting (15%). Serum levels of liver enzymes were transiently elevated in 3 patients (15%) but returned in all patients to normal values within 1 week. One patient who did not completely avoid exposure to sunlight during this interval developed a mild sunburn. DISCUSSION The purpose of this study was to determine whether ALA-induced PpIX fluorescence measured in vivo using a quantitative spectroscopic method can distinguish nondysplastic from dysplastic Barrett’s epithelium. In studies of PpIX fluorescence for detection of dysplasia in Barrett’s esophagus, only qualitative measurements of PpIX fluorescence were made.9,12,13 The present study demonstrates for the first time that quantitatively measured PpIX fluorescence intensity in Barrett’s epithelium with GASTROINTESTINAL ENDOSCOPY

483

S Brand, T Wang, K Schomacker, et al.

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

A

B Figure 4. A, Mean standardized fluorescence intensity quotient 635 nm/480 nm for each histologic category. B, Box plots showing minimum, maximum, median, 25th and 75th percentile of standardized fluorescence intensity quotient 635 nm/480 nm. Mean standardized fluorescence intensity quotient 635 nm/480 nm in HGD is greater than in all other tissue categories.

HGD is significantly greater than in Barrett’s tissue without dysplasia. HGD was identified with a sensitivity of 77% and specificity of 71%. PpIX fluorescence imaging has been used in studies with ALA for the identification of dysplasia. Gossner et al.12 found improved sensitivity for detection of dysplasia when biopsy sites were selected based on the intensity of PpIX fluorescence (69%) compared to random sampling (sensitivity 16%) in patients with Barrett’s esophagus. Endlicher et al.13 used the same technique to detect dysplasia in Barrett’s mucosa by using different ALA dosages and application regimens (orally and topically). The best discrimination of HGD from nondysplastic Barrett’s epithelium was found with dosages of 10 mg/kg (sensitivity 80%, specificity 56%) and 20 mg/kg (sensitivity 100%, specificity 51%). An ALA dose of 5 mg/kg ALA was too low for detection of dysplasia whereas 30 mg/kg ALA resulted in 100% sen484

GASTROINTESTINAL ENDOSCOPY

sitivity but was associated with a decrease in specificity.13 Based on these data, and because side effects of ALA are dose-dependent,30 a dose of 10 mg/kg ALA was chosen for the present study. Wagnieres et al.31 found an excellent negative predictive value (95%) for the diagnosis of dysplasia in patients with Barrett’s esophagus for ALA-induced PpIX fluorescence, although the positive predictive value by using this qualitative technique was only 30%. Although these studies demonstrated the feasibility of surveillance of Barrett’s epithelium with ALA-induced fluorescence endoscopy, they relied on qualitative assessment of PpIX fluorescence by the endoscopist. Assessing fluorescence intensity by qualitative means may have an inherent bias because areas with the greatest intensities may be examined preferentially. In addition, there may be observer bias because the endoscopist is aware of the findings by standard endoscopy. However, quantitative spectroscopic fluorescence systems can screen only small mucosal areas for dysplastic lesions. Therefore, for ultimate clinical use improved fluorescence imaging or a combined fluorescence imaging/spectroscopy system would be needed. Spectroscopic measurements with tissue autofluorescence without ALA-induced PpIX fluorescence have been used for the detection of dysplasia in the GI tract.32,33 Panjehpour et al.34 used tissue autofluorescence spectroscopy to detect diffuse HGD in patients with Barrett’s esophagus with a sensitivity of 90%. However, only 28% of the specimens with LGD and focal HGD were spectroscopically classified as HGD with this technique. The results of the present study suggest that increased mucosal thickness, as seen in nodular HGD, is associated with a decrease in autofluorescence. This might be explained by the fact that the majority of tissue autofluorescence originates in the submucosa. Increased nodularity, which is characterized by increased mucosal thickness, results in a decrease of autofluorescence. By using the standardized fluorescence intensity quotient of 635 nm/480 nm nodular HGD could be differentiated from other tissue types with 100% sensitivity and 100% specificity. Diffuse HGD confers a 5-fold greater risk for the development of esophageal adenocarcinoma than focal HGD, whereas the risk for the development of cancer for nodular HGD is 4-fold higher than that for non-nodular HGD.35 Data from the present study suggest that the PpIX-to-autofluorescence-ratio measurement might identify areas with a higher risk of subsequent cancer development with high sensitivity and specificity. This study confirms the high frequency of HGD in areas of Barrett’s epithelium with increased nodularity, which can be identiVOLUME 56, NO. 4, 2002

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

fied without ALA-induced fluorescence. However, 5 areas of HGD without nodularity displayed an increased ALA fluorescence and were identified spectroscopically as HGD. Twenty-five percent of the patients studied had at least one focus of HGD. This relatively high frequency of dysplasia is likely related to the select nature of the study population. Only patients with histopathologically confirmed Barrett’s esophagus were included and nearly one third of these (n = 6) had at least one focus of HGD or LGD at a previous endoscopy. By using fluorescence endoscopy, 3 additional sites of HGD (not found by random sampling) were detected, further increasing the number of dysplastic biopsies. Autofluorescence imaging without spectroscopic measurements for the detection of dysplasia has also been studied in Barrett’s esophagus36 and the colon.36-39 However, autofluorescence is limited by the low contrast between normal and dysplastic epithelium. In the present study, the difference in fluorescence between nondysplastic Barrett’s epithelium and HGD appeared to be enhanced by the addition of ALA. Another limitation of autofluorescence imaging is the false-positive rate caused by blood from biopsy sites,39 which may limit the use of autofluorescence imaging methods in Barrett’s esophagus. Blood is less likely to influence quantitative PpIX fluorescence measurements because light at 635 nm is transmitted through blood better than light at 500 nm (autofluorescence). In an animal model of colitis, inflammation was found to increase PpIX fluorescence, leading to false-positive results as measured by a qualitative fluorescence imaging system.40 Therefore, the potential influence of inflammation on PpIX fluorescence was analyzed in the present study by using quantitative measurements. PpIX fluorescence intensity was not found to be increased in tissue with mild and moderate inflammation. The two tissue samples with the highest degree of inflammation had the greatest PpIX fluorescence within their tissue categories, suggesting that detection of dysplasia by PpIX fluorescence in areas of severe inflammation might lead to false-positive results. However, this observation has to be confirmed in a larger study group. The interval between ALA administration and endoscopy may also influence PpIX fluorescence intensity. In the present study, spectroscopy was performed in all patients 3 hours after ingestion of ALA. Peak PpIX levels in esophageal tissue are reached within 4 to 6 hours30,41 and the PpIX peak is reached earlier in Barrett’s epithelium (4.6 h) than in squamous epithelium (6.5 h).41 Thus, perVOLUME 56, NO. 4, 2002

S Brand, T Wang, K Schomacker, et al.

forming spectroscopic measurements more than 5 hours after administration of ALA might decrease the differentiation between normal esophageal tissue and Barrett’s epithelium, but whether it affects the detection of dysplasia is unknown. In addition, lower doses of ALA (2 and 10 mg/kg) result in higher PpIX fluorescence in Barrett’s epithelium compared with squamous epithelium, but this difference was not apparent when high ALA doses (30 mg/kg) were used.42 The importance of this observation for the detection of dysplasia has not been established. ALA has several advantages by comparison to other exogenous fluorescent markers. It has a short period of photosensitivity (less than 48 hours) and can be administered orally a few hours before the examination. ALA is generally described as a relatively safe drug with few side effects such as phototoxicity, reversible liver enzyme abnormalities, nausea, and vomiting. 24,27,30 These side effects appear to be dose dependent.30 Although no serious side effects were noted in the present study, a significant number of patients (30%) complained of nausea and 15% of patients had at least one episode of vomiting. Fluorescence endoscopy adds little additional time to standard endoscopy because the fluorescence endoscopy devices currently available have a white light source that allows switching between white light and fluorescence endoscopy without changing endoscopes. Fluorescence spectroscopy can also be readily applied by passing the optical fiber through the accessory channel of a standard endoscope and activating the laser-induced fluorescence device with a foot switch. By using an endoscope with 2 accessory channels, the spectroscopic fiber can be left inside one of the channels, thereby eliminating the need to remove the fiber before taking biopsy specimens. By using this method, a spectroscopic measurement can be made in less than 10 seconds. Furthermore, if biopsy specimens are taken only from spectroscopically abnormal areas, the total number of specimens might be reduced, leading to a reduction in total time required for the procedure. In the present study, in which a specimen was obtained from all sites measured spectroscopically, the fluorescence measurements added approximately 10 minutes to the procedure. Surveillance of patients with Barrett’s esophagus with ALAinduced fluorescence may provide a more precise method for detection of HGD and may reduce sampling error. Although 3 additional sites of flat HGD were found by fluorescence endoscopy, the results of fluorescence endoscopy were comparable with those of conventional endoscopy (77% sensitivity for detection of HGD in both groups). These data suggest GASTROINTESTINAL ENDOSCOPY

485

S Brand, T Wang, K Schomacker, et al.

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

that the combination of fluorescence endoscopy and random biopsies may improve the detection of HGD. However, a randomized, prospective trial is necessary to determine conclusively whether fluorescence endoscopy has a diagnostic advantage over the current surveillance strategy by using random biopsies.

13.

14.

DISCLOSURE The 5-ALA used in this study was provided by DUSA Pharmaceuticals, Inc., Valhalla, N.Y., and the endoscopic fluorescence imaging system was provided by Pentax Precision Instruments. ACKNOWLEDGMENTS

15.

16.

The authors thank William Puricelli, RN for assistance in conducting the study. REFERENCES 1. Shaheen NJ, Crosby MA, Bozymski EM, Sandler RS. Is there publication bias in the reporting of cancer risk in Barrett’s esophagus? Gastroenterology 2000;119:333-8. 2. Schnell TG, Sontag SJ, Chejfec G, Aranha G, Metz A, O’Connell S, Seidel UJ, Sonnenberg A. Long-term nonsurgical management of Barrett’s esophagus with high-grade dysplasia. Gastroenterology 2001;120:1607-19. 3. Cameron AJ, Ott BJ, Payne WS. The incidence of adenocarcinoma in columnar-lined (Barrett’s) esophagus. N Engl J Med 1985;313:857-9. 4. Hameeteman W, Tytgat GN, Houthoff HJ, van den Tweel JG. Barrett’s esophagus: development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:1249-56. 5. Skinner D. The incidence of cancer in Barrett’s esophagus varies according to series. In: Giuli R, McCallum R, editors. Benign lesions of the esophagus and cancer. Answer to 210 questions. New York: Springer-Verlag; 1989. p. 764-5. 6. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF, Jr. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991;265:1287-9. 7. Pera M, Cameron AJ, Trastek VF, Carpenter HA, Zinsmeister AR. Increasing incidence of adenocarcinoma of the esophagus and esophagogastric junction. Gastroenterology 1993;104: 510-3. 8. van Sandick JW, van Lansehot JJ, Kuiken BW, Tytgat GN, Offerhaus GJ, Obertop H. Impact of endoscopic biopsy surveillance of Barrett’s oesophagus on pathological stage and clinical outcome of Barrett’s carcinoma. Gut 1998;43:216-22. 9. Messmann H, Knuchel R, Baumler W, Holstege A, Scholmerich J. Endoscopic fluorescence detection of dysplasia in patients with Barrett’s esophagus, ulcerative colitis, or adenomatous polyps after 5-aminolevulinie acid-induced protoporphyrin IX sensitization. Gastrointest Endosc 1999;49: 97-101. 10. Mayinger B, Reh H, Hochberger J, Hahn EG. Endoscopic photodynamic diagnosis: oral aminolevulinic acid is a marker of GI cancer and dysplastic lesions. Gastrointest Endosc 1999; 50:242-6. 11. Endlicher E, Knuechel R, Schoelmerich J, Messmann H. Endoscopic photodynamic diagnosis: oral aminolevulinic acid is a marker of GI cancer and dysplastic lesions (ALA) [abstract]. Gastrointest Endosc 1999;49:AB117. 12. Gossner L, Sroka R, Stepp H, May A, Stolte M, Ell C. Photodynamic diagnosis versus random biopsies for dysplasia and invisible mucosa]. Cancer in Barrett’s esophagus—a 486

GASTROINTESTINAL ENDOSCOPY

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

prospective randomized trial [abstract]. Gastroenterology 1999;116:A175. Endlicher E, Knuechel R, Hauser T, Szeimies RM, Scholmerich J, Messmann H. Endoscopic fluorescence detection of low and high grade dysplasia in Barrett’s oesophagus using systemic or local 5-aminolaevulinic acid sensitisation. Gut 2001;48:314-9. Kriegmair M, Baumgartner R, Knuechel R, Steinbach P, Ehsan A, Lumper W, et al. Fluorescence photodetection of neoplastic urothelial lesions following intravesical instillation of 5-aminolevulinic acid. Urology 1994;44:836-41. Koenig F, McGovern FJ, Larne R, Enquist H, Schomacker KT, Deutsch TF. Diagnosis of bladder carcinoma using protoporphyrin IX fluorescence induced by 5-aminolaevnlinic acid. BJU Int 1999;83:129-35. Leunig A, Rick K, Stepp H, Gutmann R, AIwin G, Baumgartner R, Feyh J. Fluorescence imaging and spectroscopy of 5-arminolevulinic acid induced protoporphyrmn IX for the detection of neoplastic lesions in the oral cavity. Am J Surg 1996;172:674-7. af Klinteberg C, Enejder AM, Wang I, Andersson-Engels S, Svanberg S, Svanberg K. Kinetic fluorescence studies of 5aminolaevulinic acid-induced protoporphyrin IX accumulation in basal cell carcinomas. S Photochem Photobiol B 1999;49:120-8. Baumgartner R, Huber RM, Schulz H, Stepp H, Rick K, Gamarra F, et al. Inhalation of 5-aminolevulinie acid: a new technique for fluorescence detection of early stage lung cancer. J Photochem Photobiol B 1996;36:169-74. Stummer W, Stocker S, Wagner S, Stepp H, Fritsch C, Goetz C, et al. Intraoperative detection of malignant gliomas by 5aminolevulinic acid-induced porphyrin fluorescence. Neurosurgery 1998;42:5l8-26. Gannon MJ, Johnson N, Roberts DJ, Holroyd JA, Vernon DJ, Brown SB, et al. Photosensitization of the endometrium with topical 5-anunolevulinic acid. Am S Obstet Gynecol 1995;173: 1826-8. Hinnen P, de Rooij FW, van Velthuysen ML, Edixhoven A, van Hillegersberg R, Tilanus HW, et al. Biochemical basis of 5aminolaevuliie acid-induced protoporphyrin IX accumulation: a study in patients with (pre)malignant lesions of the oesophagus. Br J Cancer 1998;78:679-82. Peng Q, Berg K, Moan J, Kongshaug M, Nesland SM. 5Aminolevulinic acid-based photodynamic therapy: principles and experimental research. Photochem Photobiol 1997;65: 235-51. Barr H, Shepherd NA, Dix A, Roberts DJ, Tan WC, Krasner N. Eradication of high-grade dysplasia in columnar-lined (Barrett’s) oesophagus by photodynamic therapy with endogenously generated protoporphyrin IX. Lancet 1996;348: 584-5. Gossner L, Stolte M, Sroka R, Rick K, May A, Hahn EQ, et al. Photodynamic ablation of high-grade dysplasia and early cancer in Barrett’s esophagus by means of 5-aminolevulinic acid. Gastroenterology 1998;114:448-55. Grant WE, Hopper C, MacRobert AJ, Speight PM, Bown SG. Photodynamic therapy of oral cancer: photosensitisation with systemic aminolaevulinic acid. Lancet 1993;342:147-8. Mlkvy P, Messmann H, Regula J, Conio M, Pauer M, Millson CE, et al. Photodynamic therapy for gastrointestinal tumors using three photosensitizers—ALA induced PPIX, Photofrin and MTHPC. A pilot study. Neoplasma 1998;45:157-61. Regula J, MacRobert AJ, Gorchein A, Buonaccorsi GA, Thorpe SM, Spencer GM, et al. Photosensitisation and photodynamic therapy of oesophageal, duodenal, and colorectal tumours VOLUME 56, NO. 4, 2002

Detection of high-grade dysplasia in Barrett’s esophagus by protoporphyrin IX

28.

29.

30.

31.

32.

33.

34.

35.

using 5 aminolaevulinic acid induced protoporphyrin TX—a pilot study. Gut 1995;36:67-75. Loh CS, MacRobert AJ, Bedwell J, Regula J, Krasner N, Bown SG. Oral versus intravenous administration of 5-aminolaevulinic acid for photodynamic therapy. Br J Cancer 1993;68:41-51. Reid BJ, Haggitt RC, Rubin CE, Roth G, Surawicz CM, Van Belle G, et al. Observer variation in the diagnosis of dysplasia in Barrett’s esophagus. Hum Pathol 1988;19:166-78. Ackroyd R, Brown N, Vernon D, Roberts D, Stephenson T, Marcus S, et al. 5-Aminolevulinic acid photosensitization of dysplastic Barrett’s esophagus: a pharmacokinetic study. Photochem Photobiol 1999;70:656-62. Wagnieres G, Jornod P, Stepinac T, Grosjean P, Lange N, van den Bergh H, et al. Tissue fluorescence spectroscopy: application to the endoscopic detection of dysplasia and early cancer in Barrett’s esophagus after sensitization with 5-aminolevulinic acid [abstract.] Endoscopy 1999;31:E18. Cothren RM, Richards-Kortum R, Sivak MV Jr, Fitzmaurice M, Rava RP, Boyce GA, Doxtader M, et al. Gastrointestinal tissue diagnosis by laser-induced fluorescence spectroscopy at endoscopy. Gastrointest Endosc 1990;36:105-11. Schomacker KT, Frisoli JK, Compton CC, Flotte TJ, Richter JM, Deutsch TF, et al. Ultraviolet laser-induced fluorescence of colonic polyps. Gastroenterology 1992;102:1155-60. Panjehpour M, Overholt BF, Vo-Dinh T, Haggitt RC, Edwards DH, Buckley FP 3rd. Endoscopic fluorescence detection of high-grade dysplasia in Barrett’s esophagus. Gastroenterology 1996;111:93-101. Buttar NS, Wang KK, Sebo TJ, Riehle DM, Krishnadath KR,

VOLUME 56, NO. 4, 2002

36.

37.

38.

39.

40.

41.

42.

S Brand, T Wang, K Schomacker, et al.

Lutzke LS, et al. Extent of high-grade dysplasia in Barrett’s esophagus correlates with risk of adenocarcinoma. Gastroenterology 2001;120:1630-9. Haringsma J, Tytgat GN, Yano H, Iishi H, Tatsuta M, Ogihara T, et al. Autofluorescence endoscopy: feasibility of detection of GI neoplasms unapparent to white light endoscopy with an evolving technology. Gastrointest Endosc 2001;53:642-50. Wang TD, Van Dam J, Crawford JM, Preisinger EA, Wang Y, Feld MS. Fluorescence endoscopic imaging of human colonic adenomas. Gastroenterology 1996;111:1182-91. Wang TD, Crawford JM, Feld MS, Wang Y, Itzkan I, Van Dam J. In vivo identification of colonic dysplasia using fluorescence endoscopic imaging. Gastrointest Endosc 1999;49:447-55. Brand S, Stepp H, Ochsenkuhn T, Baumgartner R, Baretton G, Holl J, et al. Detection of colonic dysplasia by light-induced fluorescence endoscopy: a pilot study. Int J Colorectal Dis 1999;14:63-8. Messmann H, Kullmann F, Wild T, Knuchel-Clarke R, Ruschoff I, Gross V, et al. Detection, of dysplastic lesions by fluorescence in a model of colitis in rats after previous photosensitization with 5-aminolaevulinic acid. Endoscopy 1998;30:333-8. Hinnen P, de Rooij FW, Terlouw EM, Edixhoven A, van Dekken H, van Hillegersberg R, et al. Porphyrin biosynthesis in human Barrett’s oesophagus and adenocarcinoma after ingestion of 5-aminolaevulinic acid. Br S Cancer 2000;83:539-43. Saidi R, Wong Kee Song L, DaCosta R, Wilson B, Lilge L, Kost J, et al. Fluorescence studies of the selectivity of 5-aminolevulinic acid-induced protoporphyrin IX in Barrett’s esophagus [abstract]. Gastroenterology 2000;118:A230.

GASTROINTESTINAL ENDOSCOPY

487