Vital staining and Barrett s esophagus

Transcription

1 Marcia Irene Canto, MD, MHS Baltimore, Maryland Vital staining or chromoendoscopy refers to staining of endoscopic tissue or topical application of chemical stains or pigments to alter tissue appearances and thereby improve localization, characterization, or diagnosis. 1,2 Chromoendoscopy is an important topic in GI practice because it is an easy, inexpensive, widely available, and safe technique for improving endoscopic diagnosis. With Barrett s esophagus vital staining may improve our ability to detect columnar epithelium and identify patients at risk for cancer. The accurate diagnosis and estimation of length of Barrett s esophagus by visual inspection at endoscopy can be very imprecise, with reported sensitivity as low as 60% and false positivity as high as 31%. Furthermore vital staining may help diagnose severe dysplasia and early cancer and therefore aid in selection of patients for more aggressive endoscopic surveillance, mucosal ablation, or surgery. WHAT WE KNOW NOW Four vital stains have been used in patients with Barrett s esophagus: Lugol s solution, methylene blue, toluidine blue, and indigo carmine (Table 1). Lugol s solution, which contains iodine and potassium iodide, is a vital stain used to improve detection of early squamous cell esophageal cancer (Table 1). A 1.5% to 4% solution is sprayed on the esophagus and within minutes the normal whitish squamous mucosa will change to dark brown or greenish brown as a result of binding of the iodine to glycogen in nonkeratinized squamous epithelium. Cells that are inflamed, dysplastic, or malignant will not stain. Lugol s staining is effective in detecting synchronous esophageal tumors in high risk patients with head and neck cancers. 3 It is used in screening for early squamous cell cancers of the esophagus, 3,4 particularly in endemic areas such as Japan and China. Lugol s solution also has been used for improving the diagnosis of Barrett s esophagus by highlighting From the Division of Gastroenterology, The Johns Hopkins University School of Medicine, Baltimore, Maryland. Reprint requests: Marcia Irene Canto, MD, MHS, 1830 E. Monument St., Room 425, Baltimore, MD Gastrointest Endosc 1999;49:S12-6. Copyright 1999 by the American Society for Gastrointestinal Endoscopy /99/$ /0/95786 columnar metaplastic epithelium, which remains pink or unstained, from squamous mucosa, which will stain dark brown. Woolf et al. 5 studied 11 subjects with a previous diagnosis of columnar lined esophagus and 12 control patients to evaluate the accuracy of endoscopic diagnosis of Barrett s esophagus enhanced by Lugol s staining. Compared with histologic data as the reference measure, the sensitivity, specificity, and accuracy of Lugol s-enhanced endoscopy for diagnosing Barrett s esophagus were 89%, 93%, and 91%, respectively. Toluidine blue is a basic absorptive stain that stains nuclei (Table 1). It has been used to diagnose inflammatory and malignant cells because of their increased nuclear-to-cytoplasmic ratio. 2 Staining is accomplished by applying a 1% acetic acid rinse followed by a 1% aqueous solution of toluidine blue before a final wash with 1% acetic acid. 2,6 Toluidine blue has been used to help diagnose squamous cell esophageal cancer. 7 In 1987 Chobanian et al. 6 reported a sensitivity of 98% and specificity of 80% 6 for diagnosing Barrett s esophagus. Toluidine blue stains columnar-type mucosa deep blue and does not discriminate between the various histologic subtypes of Barrett s esophagus. 6 Indigo carmine is a blue contrast stain and is not absorbed by cells. It pools in the crevices between mucosal projections and highlights subtle mucosal abnormalities (Table 1). Stevens et al. 8 described its use for diagnosing Barrett s esophagus in combination with high magnification endoscopy. A 0.1% solution of indigo carmine was sprayed on columnar mucosa after delineation of the squamocolumnar junction with Lugol s solution. The areas of possible Barrett s esophagus were then examined with the Olympus GIF-200Z endoscope (Olympus Corp, Tokyo, Japan), which provides adjustable image magnification from 10 to 35. Using this technique, Barrett s epithelium was visualized as a slightly raised surface pattern with villiform (similar to small intestinal mucosa). 8 This appearance was correlated with the presence of specialized columnar epithelium in endoscopic biopsy specimens. Furthermore Stevens et al. 8 described a slightly raised lesion with irregular surface that was mildly dysplastic Barrett s epithelium on biopsy. Methylene blue is a vital stain taken up by actively absorbing tissues, such as small intestinal and S12 GASTROINTESTINAL ENDOSCOPY VOLUME 49, NO. 3, PART 2, 1999

2 M Canto Table 1. Vital stains used in Barrett s esophagus Type of stain What is stained What does not stain Clinical use Lugol s solution Stains glycogen- Dysplastic, inflamed, or Squamous cell esophageal cancer; reflux (absorptive stain) containing squamous malignant squamous esophagitis; Barrett s esophagus cells brown cells; columnar mucosa (delineates squamous from columnar) Methylene blue Actively absorbing Squamous and gastric cells, Gastric intestinal metaplasia; Barrett s (absorptive stain) intestinal-type cells gastric-type metaplasia esophagus (stains specialized columnar epithelium and not gastric-type) Toluidine blue Nuclei of columnar Squamous cells Squamous cell carcinoma of the esophagus (absorptive stain) (gastric and intestinal- and oropharynx; Barrett s esophagus (stains type) cells both gastric and intestinal metaplasia) Indigo Carmine* Nonabsorbed blue stain Not applicable Highlights mucosal irregularities in the (contrast stain) pools in crevices between esophagus (Barrett s esophagus) or mucosal projections colon (flat colorectal tumors) *Combined with high-magnification endoscopy. colonic epithelium. It also will stain metaplastic absorptive epithelium, such as intestinal-type metaplasia in the stomach. 9,10 Methylene blue will not stain nonabsorptive epithelium, such as squamous or gastric mucosa, or ectopic gastric metaplasia in the duodenal bulb. Methylene blue has a sensitivity of 94% for intestinal metaplasia in the stomach. 9 It has been used to aid in the diagnosis of early gastric cancer and to highlight subtle mucosal changes in the small intestine (for example, celiac disease) and colon (flat adenomas and carcinomas). Methylene blue staining involves application of a mucolytic agent, followed by dye, followed by washing to remove excess dye. The mucolytic agent helps to remove surface mucus, thereby increasing the uptake of dye into epithelial cells. Japanese investigators first described the technique of methylene blue staining for improving the diagnosis of early gastric cancer using pronase as the mucolytic agent. 10 Fennerty et al. 9 modified the technique by substituting acetylcysteine for pronase, which is not available in the United States. Methylene blue staining in Barrett s esophagus was evaluated in a pilot study involving 14 patients with Barrett s esophagus and 12 control subjects. Canto et al. 11 showed that methylene blue stains intestinal metaplasia (specialized columnar epithelium) in Barrett s esophagus with high accuracy and no side effects. The mean cost of the reagents was below $9.00 and the mean staining time was 5 minutes. Methylene blue also stained specialized columnar epithelium reproducibly in patients who underwent a repeated staining within 4 weeks. Interestingly methylene blue staining correctly diagnosed specialized columnar epithelium in 5 of the 12 control subjects (diagnostic yield of 42%). These control subjects comprised patients without GERD (n = 5), patients with GERD but no Barrett s esophagus (n = 4), and patients with GERD and more than 3 cm of columnar mucosa but no specialized columnar epithelium in prior endoscopic biopsy specimens (n = 3). Methylene blue also has been used for improved diagnosis of intestinal metaplasia at the gastric cardia. Morales et al. 12 obtained 4 random standard biopsy specimens followed by 4 stained biopsy specimens after staining with methylene blue. The sensitivity for intestinal metaplasia at the cardia increased from 38% (for random biopsy alone) to 67% with targeted methylene blue-stained biopsy. The potential of methylene blue-directed biopsy for improving endoscopic surveillance in Barrett s esophagus was suggested in a prospective, controlled, sequential, endoscopic surveillance trial. Canto et al. 13 studied 43 patients with biopsyproven Barrett s esophagus and specialized columnar epithelium. These patients underwent both 4- quadrant jumbo random biopsy and methylene bluedirected jumbo biopsy in a randomized order. Methylene blue-directed biopsy led to identification of a much larger proportion of specialized columnar epithelium in endoscopic biopsy samples compared with those obtained by random biopsy (p = ). This was particularly evident in patients with limited-segment (54% versus 94%) and longsegment Barrett s esophagus (72% versus 92%). Despite fewer biopsies per patient, methylene bluedirected biopsy diagnosed dysplasia or cancer in significantly more biopsy samples (12% versus 6%, p = 0.004) and more patients (44% versus 28%, p = 0.03) than random biopsy. It diagnosed 5 more patients with low grade dysplasia, 1 more with high grade dysplasia, and 1 more with cancer. The estimated average cost of diagnosing a cancer using methylene blue-directed biopsy was less than half that of random biopsy. VOLUME 49, NO. 3, PART 2, 1999 GASTROINTESTINAL ENDOSCOPY S13

3 M Canto Vital staining and Barrett s esophagus The methylene blue staining characteristics of dysplastic and malignant Barrett s esophagus were studied prospectively in parallel in vivo and ex vivo experiments by Canto et al.. 14 The study involved 551 biopsy samples from 47 patients and 48 sections from 5 surgical specimens. In the ex vivo study, esophagectomy specimens with high grade dysplasia or early adenocarcinoma were mapped, photographed, and sampled before and after methylene blue staining. In a concurrent in vivo study, methylene blue staining was performed and stain intensity was characterized as dark blue, moderate blue, light blue, and unstained on the basis of a standardized ordinal scale. For patients with diffusely staining Barrett s esophagus, stain heterogeneity (the degree of variation in the intensity of methylene blue staining) also was classified as absent, mild, moderate, or marked. Pathologists estimated the proportion of specialized columnar epithelium in each specimen and graded dysplasia. Univariate and multivariate analyses were performed to determine the association of staining properties with the risk for high grade dysplasia and adenocarcinoma. The accuracy of ex vivo and in vivo methylene blue staining for specialized columnar epithelium was 87% and 90%, respectively. Accuracy was influenced by the length of Barrett s esophagus, the location of the biopsy specimen in relation to the gastroesophageal junction, and the presence of esophagitis or dysplasia. Macroscopic and microscopic ulceration reduced the accuracy of staining. The intensity of staining was significantly associated with the grade of dysplasia; 92% of endoscopic biopsy samples with high grade dysplasia or cancer were unstained or light blue, compared with 82% with low grade dysplasia and 38% with no dysplasia. Furthermore the presence of moderate to marked stain heterogeneity was present in all patients with severe dysplasia or adenocarcinoma, compared with 21% of patients with low grade dysplasia and 3% without dysplasia. In both the ex vivo and in vivo studies, light to absent staining (p = 0.01) and moderate to marked heterogeneity (p = 0.01) were significantly associated with high grade dysplasia or cancer in a multivariate model, which included the presence of an endoscopic lesion and ulceration. Increased heterogeneity and decreased methylene blue stain intensity are significant independent predictors of high grade dysplasia or cancer and may help direct biopsies. The sensitivity and specificity of vital staining is known to be significantly affected by the presence of ulcers and esophagitis. 7,11,14 Dye can bind nonspecifically to exudate or fail to stain areas of denuded epithelium. This was clearly shown in our ex vivo experiments on esophagectomy specimens. 14 It therefore is generally recommended that a careful endoscopic examination be performed before chromoendoscopy to note the presence and location of erosions and ulcers. Vital staining is best performed after aggressive treatment of esophagitis with proton pump inhibitors, because healing may be accompanied by development of specialized columnar epithelium. 11 WHAT WE NEED TO KNOW Which vital stain should be used for the diagnosis of Barrett s esophagus? Lugol s staining delineates the border between squamous and nonsquamous cells by staining the former dark brown. It functions as a negative stain for Barrett s esophagus. Toluidine blue stains columnar mucosa blue but does not discriminate between gastric and intestinal-type metaplasia in Barrett s esophagus. Indigo carmine highlights the villiform appearance of intestinal metaplasia in Barrett s esophagus but requires high magnification endoscopy. In contrast methylene blue appears to be highly accurate for selective staining of specialized columnar epithelium, which defines Barrett s esophagus and is the histologic subtype most at risk for developing adenocarcinoma. Unlike methylene blue, Lugol s and toluidine blue staining have not been shown to improve the diagnosis of dysplasia or cancer. The relative performance (accuracy, ease of use, cost) of the various vital stains for diagnosing Barrett s esophagus and associated dysplasia or cancer have not been formally compared in any study. Is methylene blue-directed biopsy useful for diagnosing short segment Barrett s esophagus and intestinal metaplasia in the cardia? What is the cancer risk of patients with short segment Barrett s esophagus and intestinal metaplasia of the cardia? The randomized sequential study by Canto et al. 13 showed no difference in the proportion of specialized columnar epithelium using random biopsy versus methylene blue-directed biopsy, but only 8 patients were studied in this group. It is not clear from the study by Morales et al. 12 whether increased diagnostic yield for intestinal metaplasia was the result of a greater number of biopsies performed (4 specimens obtained by random biopsy followed by another 4 after methylene blue staining) or the targeting of biopsy samples made possible by methylene blue. Furthermore intestinal metaplasia at the gastroesophageal junction and cardia appears to be a common finding in several prospective studies that obtained biopsy specimens randomly from these areas (prevalence of about 21%). The clinical significance and relative risk for adenocarcinoma S14 GASTROINTESTINAL ENDOSCOPY VOLUME 49, NO. 3, PART 2, 1999

4 M Canto need to be evaluated in large longitudinal cohort studies before routine or methylene blue-directed biopsy can become standard practice. Why do dysplastic and malignant cells stain differently with methylene blue? What is the mechanism for methylene blue absorption into cells? Methylene blue appears to be absorbed through the cell membrane into the cytoplasm but the exact mechanism for absorption of methylene blue into cells is unknown. Our preliminary fluorescence microscopy experiments in esophagectomy specimens with Barrett s esophagus support this mechanism for absorption. One possible explanation for the differential staining of dysplastic and nondysplastic Barrett s esophagus is the decrease in cytoplasm (increase in nuclear to cytoplasmic ratio) and the decrease in the number of goblet cells that are characteristic of dysplastic epithelium. This theory needs to be verified, because the technique of diagnosing dysplastic and malignant cells could be further improved if decreased uptake of dye is truly caused by specific ultrastructural or physiologic differences. What is the clinical utility and cost-effectiveness of vital staining for improved diagnosis of Barrett s esophagus? Prospective studies performed by other investigators and practicing physicians in the community should verify the potential for methylene bluedirected biopsy to decrease the number of biopsies performed and increase diagnostic yield. Methylene blue-directed biopsy could significantly improve endoscopic practice if it could reliably and accurately help physicians select patients with endoscopically inapparent severe dysplasia or early cancer at a relatively lower cost. WHAT STUDIES SHOULD BE CONSIDERED IN THE FUTURE Future studies of the various vital stains could be improved by the routine use of magnification endoscopy, which could help improve the precision of biopsies and possibly enhance discrimination of subtle mucosal abnormalities that may be highlighted by the dyes. A study is needed to further evaluate the clinical utility and cost-effectiveness of various techniques for endoscopic surveillance (including methylene blue-directed biopsy) when used in the community and in academic centers to diagnose Barrett s esophagus and associated dysplasia or cancer. A prospective, randomized, controlled, clinical trial could address these issues. Training on the technique and interpretation of methylene blue staining would be required of all participating investigators. Patients with dysplasia or suspected but endoscopically inapparent early adenocarcinoma in Barrett s esophagus would be enrolled at multiple centers and randomized to different methods of performing endoscopic surveillance: standard random biopsy, 4-quadrant jumbo random biopsy (every 2 cm), and methylene blue-directed jumbo biopsy. Random biopsy using standard forceps (with no formal biopsy protocol) needs to be compared with the other 2 endoscopic techniques because it is possibly the most commonly practiced method of performing endoscopic surveillance in the community. Patients who are referred to academic centers for possible mucosal ablation of dysplasia would also be eligible. Patients would be stratified according to the estimated length of Barrett s esophagus. At least 3 expert GI pathologists would serve as study pathologists to evaluate endoscopic biopsy samples in a blinded fashion. The primary trial endpoints would be the proportion of biopsy samples and patients with dysplasia or cancer and the mean and incremental costs of diagnosing a patient with high grade dysplasia or cancer using each type of surveillance technique. Study results would be analyzed by length of Barrett s esophagus and by the setting (academic center versus community practice). The expected sample size for the study would be based on the expected difference in prevalence of high grade dysplasia or cancer using the different endoscopic surveillance techniques. In our randomized sequential trial, which included patients in an academic surveillance program, the prevalence of dysplasia in biopsies was 6% for 4-quadrant random biopsy and 12% for methylene-blue directed biopsy. 13 In the same study, the prevalence of any dysplasia or cancer in all patients was 28% versus 44% for 4-quadrant jumbo random biopsy and methylene blue-directed biopsy, respectively. 13 The prevalence of dysplasia or cancer in patients undergoing standard random biopsy is unknown but can be estimated to be approximately 15%. Hence 208 patients would be required in each group to detect a significant difference in the prevalence of any dysplasia or cancer between any of the endoscopic techniques, with an alpha error of 5% and 90% power. Alternatively if the more important endpoint was the mean cost of diagnosing a cancer in Barrett s esophagus, fewer patients per surveillance group would be required, because our cost comparison of random, 4-quadrant jumbo biopsy, and methylene blue-directed biopsy 13 resulted in a more than 50% decrease in cost (mainly as a result of fewer biopsies) using the latter technique. The associated cost differences would be calculated with respect to the diagnostic yield (number of VOLUME 49, NO. 3, PART 2, 1999 GASTROINTESTINAL ENDOSCOPY S15

5 M Canto Vital staining and Barrett s esophagus cancers or high grade dysplasias diagnosed per surveillance technique). The incremental cost for diagnosing an early esophageal cancer in Barrett s esophagus would be reported using the perspective of the payer as well as society. Sensitivity analyses would be performed by varying the prevalence of Barrett s esophagus, the length of Barrett s esophagus, and the cost of the endoscopy, biopsy, and staining. The incremental costs would be compared with other reported costs of cancer screening or endoscopic surveillance, such as breast cancer screening or colorectal cancer screening using flexible sigmoidoscopy or colonoscopy. REFERENCES 1. Fennerty MB. Tissue staining. Gastrointest Clin N Am 1994;4: Technology Assessment Committee. Endoscopic tissue staining and tattooing. Technology assessment status evaluation. American Society for Gastrointestinal Endoscopy, October, Okumura T, Aruga H, Inohara H, et al. Endoscopic examination of the upper gastrointestinal tract for the presence of second primary cancers in head and neck cancer patients. Acta Otolaryngol Suppl (Stockh) 1993;501: Sugimachi K, Kitamura K, Baba K, et al. Endoscopic diagnosis of early carcinoma of the esophagus using Lugol s solution. Gastrointest Endosc 1992;38(6): Woolf GM, Riddell RH, Irvine EJ, Hunt RH. A study to examine agreement between endoscopy and histology for the diagnosis of columnar lined (Barrett s) esophagus. Gastrointest Endosc 1989;35(6): Chobanian SJ, Cattau EL, Winters C, et al. In vivo staining with toluidine blue as an adjunct to the endoscopic detection of Barrett s esophagus. Gastrointest Endosc 1987;33(2): Hix WR, Wilson WR. Toluidine blue staining of the esophagus. A useful adjunct in the panendoscopic evaluation of patients with squamous cell carcinoma of the head and neck. Arch Otolaryngol Head Neck Surg 1987;113(8): Stevens PD, Lightdale CJ, Green PHR, et al. Combined magnification endoscopy with chromoendoscopy for the evaluation of Barrett s esophagus. Gastrointest Endosc 1994;40: Fennerty MB, Sampliner RE, McGee DL, et al. Intestinal metaplasia of the stomach: identification by a selective mucosal staining technique [comments]. Gastrointest Endosc 1992;38(6): Ida K, Hashimoto Y, Kawai K. In vivo staining of gastric mucosa: its application to endoscopic diagnosis of intestinal metaplasia. Endoscopy 1975;7: Canto MI, Setrakian S, Petras RE, et al. Methylene blue selectively stains intestinal metaplasia in Barrett s esophagus. Gastrointest Endosc 1996;44(1): Morales T, Bhattacharyya A, Camargo E, Johnson C, Sampliner R. Methylene blue staining for intestinal metaplasia of the gastric cardia with follow-up for dysplasia. Gastrointest Endosc 1998;48: Canto MI, Setrakian S, Willis J, et al. Methylene blue-directed biopsy for improved detection of intestinal metaplasia and dysplasia in Barrett s esophagus: a controlled sequential trial [abstract]. Gastrointest Endosc 1996;43: Canto MI, Setrakian S, Petras RE, et al. Methylene blue staining of dysplastic and nondysplastic Barrett s esophagus: an in vivo and ex vivo study [abstract]. Gastrointest Endosc 1996;43:164. S16 GASTROINTESTINAL ENDOSCOPY VOLUME 49, NO. 3, PART 2, 1999