Detection of intestinal metaplasia in distal esophagus and esophagogastric junction by enhanced-magnification endoscopy

Detection of intestinal metaplasia in distal esophagus and esophagogastric junction by enhanced-magnification endoscopy Hideki Toyoda, MD, Carlos Rubio, MD, Ragnar Befrits, MD, Norihiro Hamamoto, MD, Yukihiko Adachi, MD, Edgar Jaramillo, MD Stockholm, Sweden

Background: Standard videoendoscopy identifies columnar-lined esophagus but cannot distinguish intestinal metaplasia from other types of epithelium. Enhanced-magnification endoscopy identifies different mucosal pit patterns. A preliminary study suggested that a type 3 pattern is associated with the presence of intestinal metaplasia. This study assesses the value of enhancedmagnification endoscopy for the detection of intestinal metaplasia in the distal esophagus and esophagogastric junction in patients undergoing diagnostic EGD. Methods: Patients undergoing diagnostic endoscopy for upper-GI symptoms underwent enhancedmagnification endoscopy after instillation of 1.5% acetic acid. The enhanced-magnification endoscopy mucosal pattern was classified into 3 types: 1, normal pits; 2, slit-reticular pattern; and 3, gyrus-villous pattern. Preliminary studies indicated that the type 3 pattern was related to intestinal metaplasia. One to 6 biopsies were targeted to areas having a type 3 pattern in columnar-appearing mucosa in the distal esophagus or esophagogastric junction. In the absence of type 3 pattern, one to 8 biopsies were targeted to areas with a type 2 pattern in columnar-appearing mucosa in the distal esophagus or esophagogastric junction. Results: The overall frequency of intestinal metaplasia in the esophagus and esophagogastric junction was 38.8% (26/67 patients). There was a good correlation between the type 3 pattern and intestinal metaplasia in targeted biopsy specimens (sensitivity 88.5%, specificity 90.2%, positive predictive value 85.2%, negative predictive value 92.5%, overall accuracy 90.0%). Conclusions: Enhanced-magnification endoscopy is useful for detection of intestinal metaplasia in distal esophagus and esophagogastric junction. (Gastrointest Endosc 2004;59:15-21.)

Traditionally, Barrett’s esophagus (BE) designates a condition in which normal squamous mucosa is replaced by columnar epithelium, a finding designated columnar-lined esophagus (CLE). The latter may be heterogeneous, being composed of fundic mucosa, cardiac mucosa, or a distinctive epithelial type known as specialized columnar epithelium.1 Only specialized columnar epithelium, defined by the presence of intestinal-type goblet cells, carries an inherent risk of malignancy.2 It has, therefore, been proposed that the diagnosis of BE be reserved for patients with intestinal metaplasia (IM), detected in biopsy specimens from the distal esophagus. Received July 29, 2003. Accepted October 16, 2003. Current affiliations: Department of Gastroenterology and Hepatology, and Department of Pathology and Cytology, Karolinska Hospital, Stockholm, Sweden, Department of Endoscopic Medicine, Mie University School of Medicine, Tsu, Japan. This study was supported by grants from the Karolinska Hospital. Reprint requests: Hideki Toyoda, MD, Department of Endoscopic Medicine, Mie University School of Medicine, 2-174 Edobashi, Tsu, Mie Prefecture 514-8507, Japan. Copyright Ó 2004 by the American Society for Gastrointestinal Endoscopy 0016-5107/$30.00 PII: S0016-5107(03)02527-6 VOLUME 59, NO. 1, 2004

Although CLE can be recognized endoscopically by its velvety red mucosa (when compared with the normal, pale esophageal mucosa), IM in CLE is not identifiable at conventional endoscopy. A technique that improves mucosal visualization would provide a more accurate endoscopic diagnosis of BE. To obtain a better visual recognition of BE, several endoscopic techniques have been used, including chromoendoscopy3,4 and acetic acid instillation,5 either alone or in combination with magnification endoscopy.6-9 Guelrud et al.7 described a technique they named enhanced-magnification endoscopy, which combines magnification endoscopy with instillation of acetic acid. They classified Barrett’s mucosa into 4 patterns: I, round pits; II, reticular (circular or oval pits); III, villous (fine villiform appearance without visible pits); and IV, ridged (thick villi with convoluted, cerebriform appearance without visible pits).7 Guelrud et al.7 found that the rate of detection of IM in pattern III and pattern IV mucosa in patients with previously diagnosed BE were, respectively, 87% and 100%. By using the Guelrud et al.7 classification, our findings in a pilot study (unpublished data) of the rates of detection of IM in pattern III and pattern IV GASTROINTESTINAL ENDOSCOPY

15

H Toyoda, C Rubio, R Befrits, et al.

Magnification endoscopy: detection of intestinal metaplasia in distal esophagus

were, respectively, 40% and 80%. Moreover, it was found that pattern III included both villous and slitlike patterns. Based on these finding, a new classification was devised that has 3 types of mucosal pattern: type 1, normal pits; type 2, slit-reticular pattern; and type 3, gyrus-villous pattern. The results of a preliminary study indicated that the type 3 pattern correlated with IM. The aim of the present study was to determine the sensitivity and specificity of magnification endoscopy for the detection of IM by using this revised pit classification system. PATIENTS AND METHODS Patients A prospective, uncontrolled study of standard and magnification endoscopy in a cohort of patients referred for elective upper endoscopy was conducted over a 5-month period at the Karolinska Hospital. Patients with known BE, esophageal varices, and those being treated with warfarin or antiplatelet medication were excluded. The clinical indications for endoscopy and the presence of reflux symptoms were recorded. The study protocol was approved by the ethics committee of the hospital, and all participating patients gave informed consent. Endoscopic procedure All procedures were performed by using a videoendoscope (GIF-Q160Z; Olympus Optical Co. (Europe), Hamburg, Germany), fitted with a transparent cap, which has the capability of both standard videoendoscopy and image magnification in a continuous range up to 3115. All endoscopic procedures were performed by a single endoscopist (H.T.) with experience in the technique of magnification endoscopy. The esophagus and cardia were evaluated by conventional endoscopy followed by magnification endoscopy. After removing excess air in the esophagus and the fornix, 15 mL of 1.5% acetic acid was sprinkled at low pressure onto the distal esophagus by means of a syringe attached to the accessory channel of the endoscope. There was no attempt to wash away excess acetic acid. After this procedure, both conventional endoscopy and magnification endoscopy were performed. All examinations were videotaped and photographs were obtained. Definitions The EGJ was defined as the distal end of the tubular esophagus coinciding with the proximal margin of the gastric folds.10 The squamocolumnar junction (SCJ) was identified as the demarcation between squamous-appearing and columnar-appearing mucosa. The appearance of the SCJ at conventional endoscopy was recorded and graded according to the ZAP (appearance of the ‘‘Z line’’) classification.11 Grade 0: The Z line is sharp and circular; it may be wave-like because of the mucosal folds, but narrow linear extensions (‘‘tongues’’) or islands of columnar epithelium are not present. Grade I: The Z line is irregular and there are tongue-like projections and/or 16

GASTROINTESTINAL ENDOSCOPY

islands of columnar-appearing epithelium. Grade II: A distinct, obvious narrow linear extension of columnarappearing epithelium less than 3 cm in length is present. Grade III: Distinct linear extensions of columnarappearing epithelium greater than 3 cm in length are present or there is proximal displacement of the Z line more than 3 cm. The presence of endoscopic signs of esophagitis were recorded and graded according to the Los Angeles classification.12 By using magnification endoscopy, columnar epithelium was classified into 3 types: type 1, small round pits of uniform size and shape (‘‘corpus’’ type) (Fig. 1A); type 2, slitreticular pattern with horizontally elongated mucosal pits (‘‘cardia’’ type) (Fig. 1B); and type 3, gyrus, villous, or mixed gyrus-villous patterns (‘‘IM’’ type) (Fig. 1C-E). Biopsy protocol Biopsy specimens were obtained with standard forceps under magnifying observation. In patients with a normalappearing esophagus and EGJ (ZAP 0), one to two biopsy specimens were obtained from the area of type 3 mucosa at the EGJ. If there was no type 3 mucosa, one to two biopsy specimens were obtained from the type 2 or type 1 mucosa at the EGJ. In patients with suspected CLE (ZAP I-III), targeting of biopsies was determined by the magnification endoscopy pattern. If the pattern was type 1 alone, one biopsy specimen was taken from the type 1 mucosa. For a magnification endoscopy pattern of type 2 without type 3, 4-quadrant specimens were taken 1 cm apart within the suspect segment in the distal esophagus. For magnification pattern type 3, one to 6 biopsy specimens (depending on the size of the area) were taken from the type 3 mucosa in distal esophagus. In addition, separate specimens were obtained from mucosal lesions that were visible before staining, irrespective of their surface pattern. A single specimen was taken from the proximal stomach 2 cm distal to the EGJ to differentiate between IM of clearly gastric origin and IM of possible esophageal origin. All biopsy specimens were labeled and fixed in separate tubes. Histopathologic analysis All biopsy specimens were fixed in 4% formalin and embedded in paraffin; sections were stained with H&E. Intestinal metaplasia was recorded as present or absent, confirmed by staining with alcian blue (pH 2.5) to demonstrate acid mucin within goblet cells and with periodic acid-Schiff stain. The histologic sections were interpreted by an experienced GI pathologist (C.R.) who was unaware of the endoscopic findings. Statistical analysis The Mann-Whitney U test and the chi-square test with the Yates correction were used to compare continuous and non-continuous variables, respectively. Enhancedmagnification endoscopic surface pattern and histopathology were compared by the Spearman rank correlation coefficient. A p value less than 0.05 was considered to indicate statistical significance. Although there was VOLUME 59, NO. 1, 2004

Magnification endoscopy: detection of intestinal metaplasia in distal esophagus

H Toyoda, C Rubio, R Befrits, et al.

Figure 1. Classification of surface pattern of Barrett’s epithelium and esophagogastric junctional mucosa by enhanced magnifying endoscopy. A, Small round pits of uniform size and shape (type 1, corpus). B, Slit (arrows) and reticular (arrowhead) pattern (type 2, cardiac). C, Gyrus pattern (type 3, intestinal metaplasia). D, Villous pattern (type 3, intestinal metaplasia). E, Mixed gyrus and villous pattern (type 3, intestinal metaplasia). Findings illustrated in C, D, and E constitute a single mucosal pattern (type 3).

multiple testing of outcome data arising from individual patients, correction by the Bonferroni method would not have removed significance from any finding, so all p values are presented uncorrected for multiple testing. All analyses were performed with a statistical software program (StatView for Windows, version 5.0; (SAS Institute Inc., SAS Campus Drive, Cary, N.C.). VOLUME 59, NO. 1, 2004

RESULTS Patients A total of 95 patients undergoing elective upper endoscopy between May 2002 to August 2002 were invited to participate in the study and 67 agreed (28 men, 39 women; mean age 60.1 years, range 24-88 GASTROINTESTINAL ENDOSCOPY

17

H Toyoda, C Rubio, R Befrits, et al.

Magnification endoscopy: detection of intestinal metaplasia in distal esophagus

Table 1. Clinical and endoscopic characteristics of patients with and without intestinal metaplasia in distal esophagus and esophagogastric junction Intestinal metaplasia (n = 26) Gender Male Female Mean age (y) Reflux symptoms Hiatal hernia Endoscopic esophagitis ZAP classification 0 (n = 6) I (n = 26) II (n = 28) III (n = 7)

No intestinal metaplasia (n = 41)

Intestinal metaplasia p 0.17

16 (62%) 10 (38%) 67.6 15 (58%) 13 (54%) 8 (35%)

12 (29%) 29 (71%) 55.8 15 (37%) 17 (44%) 9 (22%)

Table 2. Distribution of intestinal metaplasia in type 1, 2, and 3 mucosal patterns at esophagogastric junction and columnar-appearing mucosa proximal to the esophagogastric junction

0.0036 0.1538 0.4575 0.3892

Surface pattern

Positive

Negative

Total

Type 1 and type 2 Type 3

3 23

37 4

40 27

Total

26

41

67

Sensitivity = 88.5%; specificity = 90.2%; positive predictive value = 85.2%; negative predictive value = 92.5%; diagnostic accuracy = 90.0%.

0.0149 0 9/26 (35%) 11/28 (39%) 6/7 (86%)

6/6 17/26 17/28 1/7

(100%) (65%) (61%) (14%)

Table 3. Frequency of intestinal metaplasia in distal esophagus and esophagogastric junction according to ZAP classification and surface pattern Surface pattern Type 1 and type 2 ZAP classification (grade) 0 n=6 I n = 26 II n = 28 III n=7

0/6 1/17 1/16 1/1

(0) (5.9%) (6.3%) (100%)

Type 3 0 8/9 (89%)* 10/12 (83%)y 5/6 (83%)

*p = 0.0001. yp = 0.0002.

Figure 2. Correlation of enhanced-magnification endoscopic surface pattern with histopathologic findings.

years). The major reasons for endoscopy were the following: esophageal reflux symptoms (n = 30), dyspeptic symptom (n = 16), chronic diarrhea (n = 8), malignancy screening of relatives (n = 8), and anemia (n = 5). There was no adverse clinical event caused by instillation of the acetic acid. Intestinal metaplasia was found in the distal esophagus and EGJ in 26 of the 67 (38.8%) patients. In one patient, conventional endoscopy revealed a hiatal hernia and ulcer scar in the distal esophagus, while magnification endoscopy demonstrated a villous pattern in the proximal part of the herniated stomach, and biopsy specimens from this area contained low-grade dysplasia. The clinical and endoscopic characteristics of patients with and without IM in the distal esophagus 18

GASTROINTESTINAL ENDOSCOPY

and EGJ are summarized in Table 1. Patients with IM were significantly older than those without IM (mean age, respectively, 67.6 years vs. 55.8 years; p = 0.0036). There was a statistically significant association between ZAP classification and IM of the distal esophagus and EGJ (p = 0.0149). Although IM was more frequent in men, in patients with reflex symptoms, in those with a hiatus hernia, and in those with endoscopic esophagitis, these differences were not statistically significant. Enhanced-magnification endoscopy patterns All surface patterns in the distal esophagus and EGJ were evaluated easily with enhancedmagnification endoscopy. The correlation between the surface patterns and the type of glands is shown in Figure 2 (r = 0.888, p < 0.001). Of patients with type 2 mucosa in the distal esophagus and EGJ, 7.5% (3/40), 7.5% (3/40), and 85.0% (34/40) had the characteristics of, respectively, IM, fundic glands, and cardiac glands. Of patients with type 3 mucosa in the distal esophagus and EGJ, 85.2% (23/27), 3.7% (1/27), and 11.1% (3/27) had characteristics of, respectively, IM, fundic glands, and cardiac glands. The sensitivity, specificity, positive predictive value, VOLUME 59, NO. 1, 2004

Magnification endoscopy: detection of intestinal metaplasia in distal esophagus

H Toyoda, C Rubio, R Befrits, et al.

Figure 3. A, Conventional endoscopic view of esophagogastric junction after instillation of acetic acid, showing irregular squamocolumnar junction (ZAP grade I) but no distinct linear extension of columnar epithelium (arrow indicates field to be magnified). B, Enhanced-magnification endoscopic view (field indicated by arrow in A), showing type 3 gyrus pattern mucosa (arrows). Biopsy specimens from this area revealed intestinal metaplasia.

Figure 4. A, Conventional endoscopic view (short-segment Barrett’s esophagus) after acetic acid instillation showing short extension of columnar-appearing mucosa on posterior right wall of distal esophagus (ZAP grade II) (arrow indicates field to be magnified). B, Enhanced-magnification endoscopic view (field indicated by arrow in A), showing type 3 (mixed) mucosal pattern (arrow) and type 2 mucosal pattern (arrowhead). Biopsy specimens from the type 3 mucosa revealed intestinal metaplasia.

negative predictive value, and diagnostic accuracy of enhanced-magnification endoscopy for detecting IM in distal esophagus and EGJ were, respectively, 88.5%, 90.2%, 85.2%, 92.5%, and 90.0% (Table 2). The frequency of intestinal metaplasia according to the ZAP classification and surface pattern is shown in Table 3. There was a statistically significant association between a type 3 surface pattern and IM in the patients with ZAP grades I (Fig. 3) and II (Fig. 4). The surface pattern of the proximal stomach (2 cm distal to the EGJ) could not be visualized in 14 patients because of adherent mucus. These 14 patients were excluded from the analysis of the proximal stomach. Of patients with type 1 mucosa in the VOLUME 59, NO. 1, 2004

proximal stomach, 2% (1/52) and 98% (51/52) had characteristics of, respectively, IM and fundic glands (Fig. 2). Although one patient with IM in the proximal stomach had small, narrow extensions (tongues) of columnar-appearing mucosa in the distal esophagus, the columnar-appearing mucosa had a type 2 surface pattern, and the biopsy specimens from the mucosa revealed cardiac glands, not IM. DISCUSSION Magnification endoscopy with acetic acid was devised to improve the diagnosis of BE and to avoid sampling error.7 Recognition of intestinalized epithelium in BE by conventional endoscopy is GASTROINTESTINAL ENDOSCOPY

19

H Toyoda, C Rubio, R Befrits, et al.

Magnification endoscopy: detection of intestinal metaplasia in distal esophagus

extremely difficult. By using enhanced magnifying endoscopy in conjunction with spraying of acetic acid, Guelrud et al.7 classified Barrett’s mucosa into 4 patterns (cf. introduction). These investigators found that the rates of detection of IM in pattern III and pattern IV were, respectively, 87% and 100%.7 They noted that for these patients, a diagnosis of short-segment BE had previously been made. Moreover, their study did not include a control group. In another study, Guelrud et al.13 found that the frequency of pattern III was 50.3% (201/399) in the normal EGJ and that the frequency of IM in pattern III was 60.1%. Our findings in a pilot study of the classification of Guelrud et al.7 were that pattern III included villous and slit-like patterns. Based on this observation, the classification used in the present study was devised; it has been validated by the results of this study. Analysis of enhanced-magnification endoscopy revealed that IM in the distal esophagus and in EGJ is characterized by two clearly recognized patterns of surface architecture: a fine villous pattern and a gyrus pattern, and mixed forms. These two patterns were classified as type 3 by us. By using enhancedmagnification endoscopy, type 3 mucosa had a sensitivity of 88.5%, specificity of 90.2%, and overall accuracy of 88.1% for detection of IM. Endo et al.8 classified Barrett’s mucosa into 5 types by using magnifying endoscopy and a newly developed endoscope with 80-fold magnification capability (GIF-240Z; Olympus), with or without methylene blue staining. These investigators found that tubular and villous types were characteristic of IM. Although patients (n = 30) in their study had a prior diagnosis of BE, the sensitivity of these types was only 52.6% for the detection of IM. One possible explanation for this low sensitivity may be that methylene blue leaves unstained areas that cannot be evaluated by magnification endoscopy. Sharma et al.9 classified Barrett’s mucosa into 3 types by using magnification chromoendoscopy (indigo carmine dye). The yield of IM in targeted biopsy specimens according to the mucosal patterns was 97% (57/62) for ridged/villous and 17% (2/12) for circular. Type 3 in our classification may be similar to the ridged/villous pattern of Sharma et al.9 In our experience, adherent mucus or excessive indigo carmine makes magnification chromoendoscopy of the distal esophagus and EGJ difficult in some patients. These problems were rarely encountered with the use of acetic acid. Enhanced-magnification endoscopy with acetic acid is simple to perform and allows visualization of actual villi and cryptal areas, which are similar to structures observed with a stereoscopic microscope (Fig. 1).7 20

GASTROINTESTINAL ENDOSCOPY

The presence of type 1 and type 2 surface patterns was characteristic of, respectively, fundic epithelium and cardiac epithelium (Fig. 2), neither of which could be recognized by conventional endoscopy. Recognition of the cardia was ambiguous.14 In contrast, these areas were easily recognized by enhancedmagnification endoscopy after exposure of the mucosa to acetic acid. Wallner et al.11 proposed a 4-grade classification of the appearance of the squamocolumnar junction (Z line) with conventional endoscopy. Grades I and II in this ZAP classification were related to the presence of IM. Biopsy specimens from the EGJ classified as ZAP 0 revealed no IM. The present study confirms the result of Wallner et al.11 Enhanced-magnification endoscopy may be more useful in patients with ZAP grade I or II than in patients with ZAP grade 0. Because of the malignant potential of IM in shortsegment Barrett’s mucosa,15 it is essential that biopsies be targeted correctly to mucosal areas of IM. Targeting biopsies with enhanced-magnification endoscopy may decrease sampling error and the number of biopsy specimens needed for correct diagnosis. In the absence of type 3 mucosa, biopsy specimens may not be necessary. Conversely, biopsy specimens should be taken in patients with a normal EGJ who are found to have type 3 mucosa on magnification endoscopy.

REFERENCES 1. Paull A, Trier JS, Dalton MD, Camp RC, Loeb P, Goyal RK. The histologic spectrum of Barrett’s esophagus. N Engl J Med 1976;295:476-80. 2. Haggitt RC. Barrett’s esophagus, dysplasia, and adenocarcinoma. Hum Pathol 1994;25:982-93. 3. Canto MI, Setrakian S, Willis J, Chak A, Petras R, Powe NR, et al. Methylene blue-directed biopsies improve detection of intestinal metaplasia and dysplasia in Barrett’s esophagus. Gastrointest Endosc 2000;51:560-8. 4. Kiesslich R, Hahn M, Herrmann G, Jung M. Screening for specialized columnar epithelium with methylene blue: chromoendoscopy in patients with Barrett’s esophagus and a normal control group. Gastrointest Endosc 2001;53:47-52. 5. Guelrud M, Herrera I. Acetic acid improves identification of remnant islands of Barrett’s epithelium after endoscopic therapy. Gastrointest Endosc 1998;47:512-5. 6. Stevens PD, Lightdale CJ, Green PH, Siegel LM, GarciaCarrasquillo RJ, Rotterdam H. Combined magnification endoscopy with chromoendoscopy for the evaluation of Barrett’s esophagus. Gastrointest Endosc 1994;40:747-9. 7. Guelrud M, Herrera I, Essenfeld H, Castro J. Enhanced magnification endoscopy: a new technique to identify specialized intestinal metaplasia in Barrett’s esophagus. Gastrointest Endosc 2001;53:559-65. 8. Endo T, Awakawa T, Takahashi H, Arimura Y, Itoh F, Yamashita K, et al. Classification of Barrett’s epithelium by magnifying endoscopy. Gastrointest Endosc 2002;55: 641-7. VOLUME 59, NO. 1, 2004

Magnification endoscopy: detection of intestinal metaplasia in distal esophagus

9. Sharma P, Weston AP, Topalovski M, Cherian R, Bhattacharyya A, Sampliner RE. Magnification chromoendoscopy for the detection of intestinal metaplasia and dysplasia in Barrett’s oesophagus. Gut 2003;52:24-7. 10. McClave SA, Boyce HW Jr, Gottfried MR. Early diagnosis of columnar-lined esophagus: a new endoscopic diagnostic criterion. Gastrointest Endosc 1987;33:413-6. 11. Wallner B, Sylvan A, Stenling R, Janunger KG. The esophageal Z-line appearance correlates to the prevalence of intestinal metaplasia. Scand J Gastroenterol 2000;35: 17-22. 12. Armstrong D, Bennett JR, Blum AL, Dent J, De Dombal FT, Galmiche JP, et al. The endoscopic assessment of esophagitis:

VOLUME 59, NO. 1, 2004

H Toyoda, C Rubio, R Befrits, et al.

a progress report on observer agreement. Gastroenterology 1996;111:85-92. 13. Guelrud M, Herrera I, Essenfeld H, Castro J, Antonioli DA. Intestinal metaplasia of the gastric cardia: a prospective study with enhanced magnification endoscopy. Am J Gastroenterol 2002;97:584-9. 14. Spechler SJ. The role of gastric carditis in metaplasia and neoplasia at the gastroesophageal junction. Gastroenterology 1999;117:218-28. 15. Rudolph RE, Vaughan TL, Storer BE, Haggitt RC, Rabinovitch PS, Levine DS, Reid BJ. Effect of segment length on risk for neoplastic progression in patients with Barrett esophagus. Ann Intern Med 2000;132:612-20.

GASTROINTESTINAL ENDOSCOPY

21