ORIGINAL ARTICLE

The Accuracy of Confocal Laser Endomicroscopy, Narrow Band Imaging, and Chromoendoscopy for the Detection of Atrophic Gastritis Tao Liu, MD, PhD, Haoxuan Zheng, MD, PhD, Wei Gong, MD, PhD, Chudi Chen, MM, and Bo Jiang, MD, PhD

Objective: The aim of this study was to compare chromoendoscopy (CE), narrow band imaging (NBI), and confocal laser endomicroscopy (CLE) in diagnosing atrophic gastritis. Backgrounds: Atrophic gastritis, especially metaplastic atrophy, has been shown to be a risk factor for gastric cancer. Some advanced endoscopic techniques have been used to diagnose atrophic gastritis. However, it is still difficult to diagnose atrophy with a high degree of accuracy. Study: In total, 253 gastric sites from 87 consecutive patients were examined by NBI, CE, and CLE, and in turn endoscopic diagnoses were made. Histologic diagnoses of biopsies taken from the observed sites served as gold standards. Comparisons were made of the sensitivity, specificity, and accuracy between each endoscopic technique for obtaining a diagnosis atrophic gastritis. Results: NBI was found to be equivalent to CE in classifying gastric pits (k = 0.904). The CLE had a higher sensitivity (P = 0.035), specificity (P = 0.049), and accuracy (P = 0.002) than CE for diagnosing atrophic gastritis. The sensitivity and specificity of CLE for diagnosing nonmetaplastic atrophy were 86.76% and 91.89%, respectively, and for metaplastic atrophy were 91.94% and 96.86%, respectively. Interobserver and intraobserver agreements of CLE for predicting histopathologic gastritis were both high (0.938 and 0.895, respectively). Conclusions: CLE is reliable for real-time assessment of atrophic gastritis and is also able to differentiate metaplastic from nonmetaplastic atrophy. Key Words: confocal laser endomicroscopy, chromoendoscopy, NBI, atrophic gastritis

(J Clin Gastroenterol 2015;49:379–386)

Received for publication October 28, 2013; accepted May 8, 2014. From the Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou, China. T.L. and H.Z. are the cofirst authors and contributed equally. Supported by Nanfang Hospital President Foundation (Grant No. 2011C015), Southern Medical University, China; Nanfang Hospital Fund for Distinguished Young Scholars, Southern Medical University, China (to H.Z.); Science and Technology Star of Pearl River, Guangzhou, China (NO. 2014J2200030); Natural Science Foundation of Guangdong Province, China (Grant No. S2012040006985); Specialized Research Fund for the Doctoral Program of Higher Education (Grant No. 20124433120003), Ministry of Education of China; National Natural Science Foundation of China (Grant No. 81201962). The authors declare that they have nothing to disclose. Reprints: Bo Jiang, MD, PhD, Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou City 510515, China (e-mail: [email protected]). Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved.

J Clin Gastroenterol



Volume 49, Number 5, May/June 2015

G

astritis is defined as histopathologically confirmed inflammation of the gastric mucosa. Nonatrophic and atrophic gastritis are the 2 main categories. Atrophic gastritis is characterized by the mucosa that has loss of the appropriate gland and includes 2 general types, nonmetaplastic and metaplastic.1 Previous studies have shown that atrophic gastritis is a primary risk factor for the onset of intestinal-type gastric cancer.2–6 There is substantial evidence that the metaplastic glands are prone to neoplastic transformation and have the potential for evolving into invasive adenocarcinoma.7–11 Hence, it is important to screen for atrophic during endoscopy. There have been a few studies that have focused on the endoscopic diagnosis of gastritis. Sakaki et al12 first established an endoscopic criteria for the diagnosis of gastritis. Other studies have proposed different criteria.13,14 However, arriving at an endoscopic diagnosis of atrophic gastritis with high accuracy remains a challenge, and distinguishing between metaplastic and nonmetaplastic tissue is even more difficult. Confocal laser endomicroscopy (CLE) has recently been developed and represents a substantial advancement in endoscopic imaging. It consists of a confocal laser microscope built into the distal tip of a conventional white light endoscope. The new device can provide real-time, high magnification, crosssectional images of the gastrointestinal epithelium during routine endoscopy. The greatest advantage of the microendoscope is nearly 1000-fold magnification, which readily permits inspection of single cells of the gastrointestinal tract. The real-time, highly magnified images of the gastrointestinal tract mucosa can permit a histopathologic diagnosis during endoscopy without the need for biopsy. Another advance in endoscopic imaging is narrow band imaging (NBI), which uses a filter to limit the illumination of white light to certain specific wavelengths. It allows more detailed visualization of the mucosal architecture and capillary patterns without the need for staining with dyes. Because the capillary network surrounds crypts, this outlines the shape of the openings of crypts, producing a pit pattern. Chromoendoscopy (CE) is a widely used imaging technique, which needs a dye to be sprayed on the mucosa. This permits observation of the fine mucosal details for greater diagnostic accuracy.15 The aim of our study was to assess and compare the sensitivity, specificity, and accuracy of CLE, NBI, and CE for the diagnosis of atrophic gastritis.

MATERIALS AND METHODS Patients In our endoscopy unit from November 2012 to July 2013, consecutive patients who underwent endoscopy for www.jcge.com |

379

J Clin Gastroenterol

Liu et al

gastrointestinal symptoms were enrolled in the study. Inclusion criteria were: patients with previous or current Helicobacter pylori infection, bile reflux, long-term use of nonsteroidal anti-inflammatory drugs, or confirmed atrophic gastritis. Exclusion criteria were: uncorrected coagulopathy, pregnancy or breast feeding, documented allergy to fluorescein, and advanced gastric cancer. A total of 89 consecutive patients were enrolled. Two patients found to have advanced gastric cancer during endoscopy were excluded. The study protocol was approved by clinical research ethics committee of Nanfang Hospital, and full informed consent was obtained from all study participants.

NBI and CE The NBI system used in the present study had a light source (CLV-260HP) that contained 2 rotating filters, one conventional for white light, and another for NBI. With the NBI filter in place, the broadband white light was split into 2 narrow bands (wavelength of 415 and 540 nm) for illumination of the mucosa. For CE, indigo carmine (0.2%) was used as a nonabsorptive dye that has been shown to provide accurate staining for determining gastric pit classification. GIF-H260Z or GIF-Q240Z (Olympus Medical Systems Co., Tokyo, Japan) magnifying endoscopes were used for CE and NBI. The endoscope was mounted on a chargecoupled device positioned on the tip of the endoscope as the imaging sensing device. With the use of an optical zoom unit, the mucosa was magnified up to 80-fold.

CLE System All patients were also observed using a confocal laser endomicroscope (EC-3870K; OptiScan, Notting Hill, Australia and Pentax Corp., Tokyo, Japan).16 The confocal microscope was built into a conventional endoscope. For CLE, the excitation laser wavelength was 488 nm. The images were generated at a scan rate of 1.6 frames/s (1024 512 pixels) or 0.8 frames/s (1024 1024 pixels) after the fluorescein was excited. The scanning depth ranged from 0 to 250 mm, and the field of view was 475 475 mm with a lateral resolution of 0.7 mm. The optical slice thickness was 7 mm.

Endoscopic Procedures Before endoscopy examination, 20,000 U chymotrypsin (50 mL) was administered orally to remove gastric mucus. Then 1 mL of 2% fluorescein sodium (Baiyunshan Mingxing Pharmaceutical Co., Guangzhou, China) was administered intravenously to test for allergic reactions. Conscious sedation was achieved for each patient by using etomidate and fentanyl, and vital signs were monitored during the entire procedure. First, all patients underwent standard endoscopy using white light examination with GIF-H260Z or GIF-Q240Z endoscopes. After routine observation, 2 sites located in the lesser and greater curvature of the distal antrum (2 cm from the pylorus)1 were chosen for NBI observation. To locate each site accurately, the endoscope lens was placed in contact with the mucosa and suction was used to make a mark on the mucosa by the suction port. The observation sites were consistently selected along a line 12 clock and 5 mm distance from the marker. Two selected sites were magnified and carefully observed using NBI for gastric pits. Whitish mucosa with plaques, patches, or homogenous

380 | www.jcge.com



Volume 49, Number 5, May/June 2015

discoloration on the gastric mucosa, found by white light endoscopy, were also observed. After that, 2 mL indigo carmine (0.2%) was sprayed on the mucosa, and using the markers as guides, the same sites observed by NBI were then magnified and observed by CE. After NBI and CE, the endoscope was removed and fluorescein sodium (5 to 10 mL of 10% solution; Baiyunshan Mingxing Pharmaceutical Co.) was administered intravenously. After approximately 1 minute, the confocal laser endomicroscope was introduced into the stomach by conventional endoscopy. With the guide of the markers, the tip of the confocal laser endomicroscope was positioned at the same sites previously observed by NBI and CE. From the confocal images, diagnoses were made for each site. Each NBI and CE examination was conducted by the same endoscopist (H.Z.) who had at least 5 years of endoscopic experience, and was skilled at NBI, CE, and CLE. CLE was performed by another endoscopist (T.L.) who had completed 100 previous cases, and was blind to the NBI and CE results. All the endoscopists were blinded to the prior diagnosis of gastric atrophy. Confocal images were saved for postprocedural analysis.

Diagnostic Criteria On the basis of previous studies,12,14,17 and our analysis of previously obtained images, gastric pits were classified into the following 5 fundamental types by magnifying endoscopy: type A, round pits, which represented the openings of the fundic glands and only could be seen in the corpus and fundus mucosa with basically normal histopathology; type B, continuous short rod-like pits, which represented the pyloric glands, basically normal histopathology of the mucosa in the gastric antrum; type C, elongated and tortuous pits, mainly seen in the antral mucosa with atrophic gastritis or inflammation; type D, reticular pits with dilated openings, mainly seen in the areas with more severe atrophic gastritis or inflammation; and type E, villous-like pits, seen only in the areas with the intestinal metaplastic type of atrophic gastritis. Types A and B are classified as forms of nonatrophic gastritis, whereas types C, D, and E are classified as forms of atrophic gastritis (Fig. 1). Zhang et al18 described features of nonmetaplastic atrophic gastritis: decreased number of pits, and prominently dilated openings of the glands. According to Guo et al,19 the features of metaplastic atrophic gastritis included the presence of goblet cells, columnar absorptive cells and brush borders, and villiform foveolar epithelium. The diagnosis of atrophic gastritis was based on the presence of nonmetaplastic or metaplastic atrophy determined by CLE; others were diagnosed as nonatrophic gastritis (Fig. 2).

Histologic Assessment All specimens underwent standard histopathology assessment by a gastrointestinal pathologist (C.C.) who was blinded to the CLE, CE, and NBI information. Only the anatomic locations were provided, which was the routine clinical practice in our institution. All types of gastritis were defined based on the updated Sydney System criteria.20 In addition, all biopsies were taken from the same sites observed by NBI, CE, and CLE in the vicinity of the marker.

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

J Clin Gastroenterol



Volume 49, Number 5, May/June 2015

CLE, NBI, and CE for Detecting Atrophic Gastritis

FIGURE 1. The gastric pits classified by CE ( 80) and NBI ( 80). A, Type A is round pits, seen in the normal corpus mucosa with fundic glands. C, Type B is continuous short rod-like pits, seen in the normal gastric antrum with pyloric glands. E, Type C is elongated and tortuous pits, seen in the antral mucosa with atrophic gastritis or chronic gastritis. G, Type D is reticular pits, seen in the areas with more severe inflammation or atrophic gastritis. I, Type E is villous-like pits, seen mainly in the areas with intestinal metaplastic atrophy. (B), (D), (F), (H), and (J) were the corresponding NBI images. CE indicates chromoendoscopy; NBI, narrow band imaging.

Sample Size We analyzed the cases of gastritis that contained 34 sites (20 atrophic, 14 nonatrophic sites). For the primary Copyright

r

outcome measure of comparing the sensitivity of CLE versus CE for detecting atrophic gastritis, we analyzed the 20 atrophic sites confirmed by histopathology to determine

2014 Wolters Kluwer Health, Inc. All rights reserved.

www.jcge.com |

381

Liu et al

J Clin Gastroenterol



Volume 49, Number 5, May/June 2015

FIGURE 2. The CLE images of the normal and atrophic mucosa. A, Normal gastric mucosa with the fundic glands shows round pits (circle). Columnar mucus cells (arrow) are readily identifiable. D, The corresponding cross-sectional histopathology of (A) with fundic glands (left arrow) and columnar mucus cells (right arrow). B, Normal antral mucosa with the pyloric glands shows continuous short rod-like pits (circle). E, The corresponding cross-sectional histopathology of (B) with pyloric galands (arrow). C, Metaplastic atrophy in the antrum shows large and dark goblet cells (arrows). F, Goblet cells (arrows) in the corresponding histopathology of (C). G, Metaplastic atrophy in the antrum shows goblet cells (right arrow), absorptive cells (left arrow), and brush border (upper arrow). J, The corresponding histopathology of (G) with goblet cells (arrows). H, Nonmetaplastic atrophy shows decreased gastric pits and dilated opening of the gastric pits (circle). K, The corresponding histopathology of (H). I, most mucosa with decreased pits and dilated opening of pits are associated with atrophy. This image shows mildly decreased pits and dilated openings (circle) caused by chronic inflammation. L, The corresponding histopathology of (I) shows the mucosa infiltrated mainly of lymphocytes.

382 | www.jcge.com

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

J Clin Gastroenterol



Volume 49, Number 5, May/June 2015

the parameters for sample size calculation. For a power of 90%, and a significance of 0.05, the sample size required was 213 sites. For the secondary outcome measure of the diagnostic value of CLE for detecting metaplastic atrophic gastritis, for a sensitivity of 88.89%, a specificity of 96%, a power of 90%, and a significance of 0.05, the sample size required was 53 sites (38 metaplastic sites and 15 nonmetaplastic sites).

Statistical Analysis Accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated for each method (CLE, CE, and NBI) along with 95% confidence intervals. Comparisons between CLE and CE regarding accuracy, sensitivity, specificity, PPV, and NPV were made using the McNemar test for clustered data.21 The agreement between CE and NBI in the classification of gastric pits was assessed. The interobserver and intraobserver agreement of CLE for diagnosing nonmetaplastic gastritis and metaplastic gastritis were assessed. Statistical analyses were performed by using SPSS13.0 (SPSS, Chicago, IL) for windows.

RESULTS General Characteristics of Patients A total of 87 patients (48 male, 55.17%) were recruited in the study. The median age of the patients was 49.75 years (range, 21 to 75 y). A total of 253 sites of the gastric mucosa were examined. The average number of sites per patient was 2.9, and every site was examined by CLE, CE, and NBI. None of the patients experienced any endoscopic complications. However, transient yellow discoloration of the skin and urine was seen in most of the patients, which resolved within 1 to 2 hours (skin) or within 24 hours (urine).

Histopathology of All the Sites Of the 253 analyzed sites, 22 sites were located in the corpus, 208 in the antrum, and 23 sites in the lesser curvature at the incisura angularis. Histopathologic evaluation confirmed 123 nonatrophic gastritis sites, and 130 atrophic gastritis sites including 68 nonmetaplastic atrophic gastritis sites and 62 metaplastic atrophic gastritis sites. Among the 22 sites in the corpus, none were diagnosed as pseudopyloric metaplastic, which is characterized by replacement of the parietal and chief cells of the oxyntic glands by the mucosecreting epithelia.1 All the cases of metaplastic atrophic gastritis consisted of intestinal metaplastic atrophic gastritis. A total of 7 dysplasia sites were confirmed by histopathology, including 6 metaplastic atrophic gastritis sites, and 1 nonmetaplastic atrophic gastritis site.

CLE, NBI, and CE for Detecting Atrophic Gastritis

Classification of the Gastric Mucosa by CE and NBI Six sites classified as type B by CE were classified as type C by NBI. Eight sites classified as type C by CE were different from NBI, and 4 of them were classified as type B. Four other sites were classified as type D by NBI. Three sites classified as type D by CE were classified as type C by NBI. The 2 endoscopies were in total agreement in classifying types A and E. The 2 endoscopic techniques had a high agreement for the classification of gastric pits: k = 0.904 (95% confidence interval, 0.857-0.945) (Table 1). All the sites classified as type A and most classified as type B (79% by CE, 78% by NBI) were confirmed as nonatrophic gastritis by histopathology. Most of the sites classified as type C (72.15% by CE, 70% by NBI), and most type D (92% by CE, 90.19% by NBI) and all type E were confirmed as atrophic gastritis by histopathology. However, type C and type D do not distinguish between metaplastic atrophic gastritis from nonmetaplastic atrophic gastritis. Sites classified as type C contained nonmetaplastic atrophic gastritis (44 sites by CE, 43 sites by NBI) as well as metaplastic gastritis (13 sites by CE, 13 sites by NBI). Sites classified as type D also contained nonmetaplastic atrophic gastritis (22 sites by both CE and NBI) as well as metaplastic gastritis (24 sites by both CE and NBI). Generally, CE and NBI can detect atrophic gastritis, but cannot distinguish metaplastic atrophic gastritis from nonmetaplastic gastritis (Table 2).

Comparison of CLE Versus CE and NBI in Diagnosing Atrophic Gastritis The sensitivity, specificity, accuracy, and NPV of CLE were higher than those of CE with P-values 0.035, 0.049, 0.002, and 0.038, respectively. However, there was no significant difference between the PPV of CE and CLE. Because of the perfect agreement between CE and NBI, there was no need to make comparison between CLE and NBI (Table 3).

Classifying Atrophic Gastritis by CLE Metaplastic gastritis can be recognized morphologically by the presence of goblet cell–absorptive cells, and villous-like foveolar epithelium, which provided the possibility to differentiate metaplastic atrophic gastritis from nonmetaplastic atrophic gastritis. Of the 68 sites of nonmetaplastic atrophic according to real-time CLE, 59 sites were confirmed by histopathology. The sensitivity, specificity, accuracy, PPV, and NPV were 86.76%, 91.89%, 90.51%, 79.73%, and 94.97%, respectively. Of the 62 sites of metaplastic atrophic gastritis according to real-time CLE, 57 sites were confirmed by histopathology. The sensitivity, specificity, accuracy, PPV,

TABLE 1. Gastric Pits Classified by CE and NBI

NBI CE

A

B

C

D

E

j

A B C D E

18 0 0 0 0

0 94 4 0 0

0 6 71 3 0

0 0 4 47 0

0 0 0 0 6

0.904 (0.857-0.945)

CE indicates chromoendoscopy; NBI, narrow band imaging.

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

www.jcge.com |

383

J Clin Gastroenterol

Liu et al



Volume 49, Number 5, May/June 2015

TABLE 2. The Corresponding Histopathology of Gastric Pit Classification by CE and NBI

Atrophic Gastritis Type Type A CE NBI Type B CE NBI Type C CE NBI Type D CE NBI Type E CE NBI

N

Nonatrophic Gastritis

Nonmetaplastic

Metaplastic

Dysplasia

18 18

18 18

0 0

0 0

0 0

100 98

79 76

3 3

18 19

0 0

79 80

22 24

44 43

13 13

0 0

50 51

4 5

22 22

24 24

4 4

6 6

0 0

6 6

6 6

3 3

Atrophic gastritis includes metaplastic atrophy gastritis and nonmetaplastic atrophy gastritis. CE indicates chromoendoscopy; NBI, narrow band imaging.

In the current study, 3 advanced endoscopic techniques were used to observe the same sites, allowing a comparison of diagnostic values for diagnosing atrophic gastritis. We found that NBI was equivalent to CE in classifying gastric pits, and this was also shown in the colon in a previous study.25 In the current study, all the type A, and most of the type B were nonatrophic gastritis confirmed by histopathology, and only a few type B (21% by CE, 22.4% by NBI) were atrophic gastritis, which are similar to the results of Kim et al26 (type 1 and type 2) and Tanaka et al14 (type I and type II). We also found that most of the cases of atrophic gastritis with type B were intestinal metaplastic atrophic gastritis confirmed by histopathology. This may be because the pyloric glands were replaced by the mild metaplastic glands, while the number and opening of the glands were still normal. According to Kim et al,26 44 of the 68 sites classified as type 3 (mucosa with irregular and coarse granular structure) were atrophic, including 28 cases of intestinal metaplastic atrophy. In the current study, mucosa with the above structures was classified as 2 type C and type D. There were 129 sites classified by CE as type C or type D, 103 sites were atrophic, including 37 with metaplastic atrophy. Mucosa with type D was more likely to be atrophic (72% vs. 92%). Only 6 sites with type E were observed in the current study, and all of them were cases of metaplastic atrophy. This is in agreement with previous

and NPV were 91.94%, 96.86%, 95.65%, 90.48%, and 97.37%, respectively. The intraobserver agreement between real-time and post-CLE for predicting histopathologic gastritis was high (k = 0.938). The interobserver agreement between real-time and post-CLE diagnoses for the 2 different endoscopists was also high (k = 0.895) (Table 4).

DISCUSSION CLE allows simultaneous endoscopic and endomicroscopic examination, permitting in vivo histologic evaluation. Kiesslich et al22 first reported that CLE had perfect accuracy in diagnosing intraepithelial neoplasia and colorectal cancer during ongoing colonoscopy. Since then, many studies have focused on the use of CLE in gastrointestinal tract. According to Hurlstone et al,23 during the surveillance of chronic ulcerative colitis, CLE-targeted biopsies increased the diagnostic yield of intraepithelial neoplasia by 2.5-fold compared with CE-guided biopsies alone. Li et al24 studied 1786 patients and confirmed that real-time CLE diagnosis had a higher sensitivity (88.9%), specificity (99.3%), and accuracy (98.8%) for the diagnosis of gastric superficial cancer/HGIN lesions compared with white light endoscopy. All the studies showed that CLE had great advantage in the diagnosis of gastrointestinal disease because of the real-time in vivo histologic imaging. TABLE 3. The Diagnosis of Atrophic Gastritis by CLE, CE, and NBI

CLE Measure Sensitivity Specificity Accuracy PPV NPV

Fraction [n/N (%)] 120/130 106/123 226/253 120/137 106/116

(92.31) (86.18) (89.33) (87.59) (91.38)

CE 95% CI

85.95%-96.04% 78.51%-91.51% 84.70%-92.73% 80.61%-92.40% 84.33%-95.56%

Fraction [n/N (%)] 109/130 97/123 206/253 109/135 97/118

NBI 95% CI

(83.85) (78.86) (81.42) (80.74) (82.20)

76.13%-89.50% 70.39%-85.49% 75.96%-85.91% 72.87%-86.82% 73.85%-88.40%

Fraction [n/N (%)] 108/130 94/123 202/253 108/137 94/116

(83.08) (76.42) (79.84) (78.83) (81.03)

95% CI

P CLE vs. CE

75.28%-88.87% 67.76%-83.40% 74.25%-84.50% 70.86%-85.15% 72.48%-87.47%

0.035 0.049 0.002 0.122 0.038

Atrophic gastritis includes metaplastic atrophic gastritis and non-metaplastic atrophic gastritis. CE indicates chromoendoscopy; CI, confidence interval; CLE, confocal laser endomicroscopy; NBI, narrow band imaging; NPV, negative predictive value; PPV, positive predictive value.

384 | www.jcge.com

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

J Clin Gastroenterol



Volume 49, Number 5, May/June 2015

CLE, NBI, and CE for Detecting Atrophic Gastritis

TABLE 4. The Diagnosis of Nonmetaplastic Atrophic Gastritis and Metaplastic Atrophic Gastritis With CLE

Observer 1 Fraction [n/N (%)]

Measure

95% CI

Nonmetaplastic atrophic gastritis Sensitivity 59/68 (86.76) 75.86%-93.40% Specificity 170/185 (91.89) 86.73%-95.23% Accuracy 229/253 (90.51) 86.04%-93.70% PPV 59/74 (79.73) 68.47%-87.85% NPV 170/179 (94.97) 90.37%-97.53% Metaplastic atrophic gastritis Sensitivity 57/62 (91.94) 81.45%-96.99% Specificity 185/191 (96.86) 92.97%-98.72% Accuracy 242/253 (95.65) 92.14%-97.7% PPV 57/63 (90.48) 79.76%-96.07% NPV 185/190 (97.37) 93.63%-99.03%

Observer 1 (1 mo Later) Fraction [n/N (%)]

95% CI

Observer 2 Fraction [n/N (%)]

95% CI

58/68 172/185 230/253 58/71 172/182

(85.29) (92.97) (90.91) (81.69) (94.51)

74.15%-92.35% 88.02%-96.05% 86.5%-94.03% 70.36%-89.52% 89.84%-97.18%

56/68 171/185 227/253 56/70 171/183

(82.35) (92.43) (89.72) (80.00) (93.44)

70.81%-90.17% 87.68%-95.64% 85.14%-93.05% 68.39%-88.26% 88.56%-94.41%

58/62 186/191 244/253 58/63 186/190

(93.55) (97.38) (96.44) (92.06) (97.89)

83.50%-97.91% 93.67%-99.03% 93.12%-98.25% 81.73%-97.04% 94.34%-99.32%

59/62 186/191 245/253 59/64 186/189

(95.16) (97.38) (96.84) (92.18) (98.41)

85.62%-98.74% 93.67%-99.03% 93.63%-98.52% 81.99%-97.09% 95.06%-99.59%

j1

j2

0.938

0.895

Observer 1, the endoscopists who performed confocal laser endomicroscopy; Observer 2, another endoscopist who was blinded to the former CLE results; k 1, interobserver agreement; k 2, intraobserver agreement. CI indicates confidence interval; CLE, confocal laser endomicroscopy; NPV, negative predictive value; PPV, positive predictive value.

studies, which also showed that villous-like structures represented intestinal metaplastic atrophy.14,26,27 CLE had a higher sensitivity, specificity, and accuracy than NBI and CE in the diagnosis of atrophic gastritis. Because the features could be easily observed by CLE, such as goblet cells, columnar absorptive cells, and brush border, and villiform foveolar epithelium, metaplastic atrophy could be differentiated from nonmetaplastic. In the current study, the sensitivity of CLE in diagnosis of nonmetaplastic atrophy was similar to a previous study18 (86.76% vs. 83.6%). However, the specificity was lower (91.89% vs. 99.6%). Chronic inflammation in some sites may widen the openings of the glands, which made a reliable assessment of mucosal atrophy impossible. This may explain the misdiagnosis of chronic gastritis as mild atrophy. The specificity of CLE in diagnosing metaplastic atrophy was almost the same as that reported in a previous study19 (96.86% vs. 95.33%). However, the sensitivity was much lower than that of the previous study (91.94% vs. 98.13%). According to Guo et al,19 a few goblet cells and other details were not identified while performing real-time CLE because of the speed at which images were scanned, which led to a lower accuracy of real-time CLE than post-CLE. CLE images were later reassessed by the same endoscopist, which showed a sensitivity of 93.55%. Lim et al28 studied a total of 125 sites in 20 patients by probe-based CLE (pCLE) for detection of metaplasia atrophy. Off-site–based pCLE had superior specificity (94.9% vs. 84.7%, P = 0.031) and accuracy (95.2% vs. 88.0%, P = 0.012) compared with realtime pCLE. A recent study29 demonstrated that the accuracy of CLE for diagnosis of metaplastic atrophy was dependent on the experience of the operator. There are limitations to the current study. Because the NBI and CE were performed by the same endoscopist, it is possible that the CE interpretation was biased by the NBI interpretation, and the agreement between the 2 endoscopic techniques for classification of gastric pits also could have been biased. Another limitation is that only a few sites were examined in each patient because the process of conducting 3 advanced endoscopy examinations was time consuming. To limit the time to 40 minutes, only 2 of the 5 recommended sites for diagnosis of atrophic gastritis by OLGA1 were chosen to be observed in each patient. The sites with Copyright

r

lesions were also observed. Another limitation was the decreased ability to observe the sites in the lesser curvature at the incisura angularis because of decreased flexibility of the CLE. We only managed to examine 23 sites in the lesser curvature at the incisura angularis. Because atrophy is most frequently detected in the angularis incisura,1 this limitation could have led to decreased detection of the atrophic sites in the current study. In summary, the gastric pits examined by NBI and CE can be of value in making a diagnosis of atrophic gastritis, but they cannot distinguish the 2 atrophic types. Owing to the 1000-fold magnification, CLE images reveal cell structure, which is useful in diagnosing and classifying atrophic gastritis. ACKNOWLEDGMENTS This report is based on the work supported in part by departments of endoscopy, pathology, and anesthesiology. The authors thank for their excellent technical assistance and also the nurses for their excellent nursing work. REFERENCES 1. Rugge M, Correa P, Di Mario F, et al. OLGA staging for gastritis: a tutorial. Dig Liver Dis. 2008;40:650–658. 2. Filipe MI, Munoz N, Matko I, et al. Intestinal metaplasia types and the risk of gastric cancer: a cohort study in Slovenia. Int J Cancer. 1994;57:324–329. 3. Sipponen P, Riihela M, Hyvarinen H, et al. Chronic nonatropic (‘superficial’) gastritis increases the risk of gastric carcinoma. A case-control study. Scand J Gastroenterol. 1994; 29:336–340. 4. Miehlke S, Hackelsberger A, Meining A, et al. Severe expression of corpus gastritis is characteristic in gastric cancer patients infected with Helicobacter pylori. Br J Cancer. 1998;78:263–266. 5. Cassaro M, Rugge M, Gutierrez O, et al. Topographic patterns of intestinal metaplasia and gastric cancer. Am J Gastroenterol. 2000;95:1431–1438. 6. Rugge M, Cassaro M, Di Mario F, et al. The long term outcome of gastric non-invasive neoplasia. Gut. 2003;52:1111–1116. 7. Stemmermann GN. Intestinal metaplasia of the stomach. A status report. Cancer. 1994;74:556–564. 8. Garcia SB, Park HS, Novelli M, et al. Field cancerization, clonality, and epithelial stem cells: the spread of mutated clones in epithelial sheets. J Pathol. 1999;187:61–81.

2014 Wolters Kluwer Health, Inc. All rights reserved.

www.jcge.com |

385

J Clin Gastroenterol

Liu et al

9. Niwa T, Tsukamoto T, Toyoda T, et al. Inflammatory processes triggered by Helicobacter pylori infection cause aberrant DNA methylation in gastric epithelial cells. Cancer Res. 2010;70:1430–1440. 10. Vannella L, Lahner E, Osborn J, et al. Risk factors for progression to gastric neoplastic lesions in patients with atrophic gastritis. Aliment Pharmacol Ther. 2010;31:1042–1050. 11. Matsushima K, Isomoto H, Inoue N, et al. MicroRNA signatures in Helicobacter pylori-infected gastric mucosa. Int J Cancer. 2011;128:361–370. 12. Sakaki N, Iida Y, Okazaki Y, et al. Magnifying endoscopic observation of the gastric mucosa, particularly in patients with atrophic gastritis. Endoscopy. 1978;10:269–274. 13. Endo T, Awakawa T, Takahashi H, et al. Classification of Barrett’s epithelium by magnifying endoscopy. Gastrointest Endosc. 2002;55:641–647. 14. Tanaka K, Toyoda H, Kadowaki S, et al. Features of early gastric cancer and gastric adenoma by enhanced-magnification endoscopy. J Gastroenterol. 2006;41:332–338. 15. Fu KI, Sano Y, Kato S, et al. Chromoendoscopy using indigo carmine dye spraying with magnifying observation is the most reliable method for differential diagnosis between non-neoplastic and neoplastic colorectal lesions: a prospective study. Endoscopy. 2004;36:1089–1093. 16. Kiesslich R, Goetz M, Vieth M, et al. Technology insight: confocal laser endoscopy for in vivo diagnosis of colorectal cancer. Nat Clin Pract Oncol. 2007;4:480–490. 17. Tanaka K, Toyoda H, Kadowaki S, et al. Surface pattern classification by enhanced-magnification endoscopy for identifying early gastric cancers. Gastrointest Endosc. 2008;67:430–437. 18. Zhang JN, Li YQ, Zhao YA, et al. Classification of gastric pit patterns by confocal endomicroscopy. Gastrointest Endosc. 2008;67:843–853. 19. Guo YT, Li YQ, Yu T, et al. Diagnosis of gastric intestinal metaplasia with confocal laser endomicroscopy in vivo: a prospective study. Endoscopy. 2008;40:547–553.

386 | www.jcge.com



Volume 49, Number 5, May/June 2015

20. Dixon MF, Genta RM, Yardley JH, et al. Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. Am J Surg Pathol. 1996;20:1161–1181. 21. Obuchowski NA. On the comparison of correlated proportions for clustered data. Stat Med. 1998;17:1495–1507. 22. Kiesslich R, Burg J, Vieth M, et al. Confocal laser endoscopy for diagnosing intraepithelial neoplasias and colorectal cancer in vivo. Gastroenterology. 2004;127:706–713. 23. Hurlstone DP, Kiesslich R, Thomson M, et al. Confocal chromoscopic endomicroscopy is superior to chromoscopy alone for the detection and characterisation of intraepithelial neoplasia in chronic ulcerative colitis. Gut. 2008;57: 196–204. 24. Li WB, Zuo XL, Li CQ, et al. Diagnostic value of confocal laser endomicroscopy for gastric superficial cancerous lesions. Gut. 2011;60:299–306. 25. Chiu HM, Chang CY, Chen CC, et al. A prospective comparative study of narrow-band imaging, chromoendoscopy, and conventional colonoscopy in the diagnosis of colorectal neoplasia. Gut. 2007;56:373–379. 26. Kim S, Haruma K, Ito M, et al. Magnifying gastroendoscopy for diagnosis of histologic gastritis in the gastric antrum. Dig Liver Dis. 2004;36:286–291. 27. Guelrud M, Herrera I, Essenfeld H, et al. Enhanced magnification endoscopy: a new technique to identify specialized intestinal metaplasia in Barrett’s esophagus. Gastrointest Endosc. 2001;53:559–565. 28. Lim LG, Yeoh KG, Srivastava S, et al. Comparison of probebased confocal endomicroscopy with virtual chromoendoscopy and white-light endoscopy for diagnosis of gastric intestinal metaplasia. Surg Endosc. 2013;27:4649–4655. 29. Lim LG, Yeoh KG, Salto-Tellez M, et al. Experienced versus inexperienced confocal endoscopists in the diagnosis of gastric adenocarcinoma and intestinal metaplasia on confocal images. Gastrointest Endosc. 2011;73:1141–1147.

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

The accuracy of confocal laser endomicroscopy, narrow band imaging, and chromoendoscopy for the detection of atrophic gastritis.

The aim of this study was to compare chromoendoscopy (CE), narrow band imaging (NBI), and confocal laser endomicroscopy (CLE) in diagnosing atrophic g...
490KB Sizes 5 Downloads 12 Views