Comparison of chromoendoscopy and conventional endoscopy in the detection of premalignant gastric lesions

1Nemazee Hospital, Shiraz University of Medical Sciences; 2Bandar Abbass University of Medical Science, Bandar Abbas, Iran Correspondence: Dr Ahad Eshraghian, Gastroenterohepatology Research Centre of Shiraz University of Medical Sciences, Jahrom University of Medical Science, PO Box 71345-1744, Shiraz, Iran. Telephone 98-711-627-6212, fax 98-711-627-6212, e-mail eshraghiana@yahoo.com Received for publication August 7, 2008. Accepted December 1, 2008 G cancer is one of the most common malignant diseases of the gastrointestinal tract with significant geographical, ethnic and socioeconomic variations in distribution (1). Substantial interest has arisen in recent years in the diagnosis and management of early gastric cancer and premalignant lesions of the gastric mucosa because of the high cure rate achieved with treatment of these lesions, compared with the poor prognosis of full-blown cancer (2-5). The generally accepted pathological sequence that leads to gastric adenocarcinoma is superficial gastritis followed by atrophic gastritis, intestinal metaplasia and increasing grades of dysplasia resulting in cancer (6). Intestinal metaplasia is considered a potentially reversible change (7,8). Diagnosis and localization of intestinal metaplasia and early gastric cancer is problematic because of the lack of any telltale gross endoscopic signs, even with multiple random biopsies. There is no guarantee that the area of abnormality will be included in sampling. The problem with conventional endoscopic diagnosis of these lesions is that, despite the detection of abnormalities (ie, original artiCle

early gastric cancer, dysplasia or incomplete intestinal metaplasia) with routine biopsies, finding the exact location of previously sampled sites for endoscopic or surgical treatment or resampling is extremely difficult (9).
Methylene blue staining was originally described by Ida et al (9) to improve the diagnosis of early gastric cancer.Methylene blue dye is taken up by actively absorbing tissues such as the small intestine or colon epithelia.It does not stain nonabsorptive epithelia such as stratified squamous epithelium of the esophagus or normal gastric mucosa (10).
Methylene blue has recently been used to highlight subtle mucosal changes in the small intestine and colon, and to positively stain metaplastic absorptive epithelia such as intestinal type metaplasia in the stomach (11).
Although it is generally believed that chromoendoscopy can increase the rate of diagnosis of early cancers and precancerous lesions by enhancing visualization, no study has directly compared the accuracy and sensitivity of this method with conventional endoscopy and systematic biopsies.It should also be noted that this method is a time-consuming and costly procedure compared with upper endoscopy and systematic random biopsies; therefore, reliable evidence to show its efficacy is necessary before it can be recommended for use.
Because the prevalence of gastric cancer is high in Saudi Arabia, and premalignant lesions of the gastric mucosa are not infrequently encountered in our practice (12), the problems of localization and biopsy of these lesions are daily challenges.The present study was designed to compare the accuracy and sensitivity of chromoendoscopy with conventional endoscopy and systematic biopsies for locating premalignant lesions of the gastric mucosa.

Patients
The present prospective study was performed in the Department of Gastroenterology at the Shiraz University of Medical Sciences, Shiraz, Iran.
During a period of one year between September 2006 and September 2007, 84 patients were diagnosed with intestinal metaplasia.Of these, 33 patients in whom previous endoscopic biopsies had shown intestinal metaplasia were enrolled in the study.
Inclusion criteria were an age of 18 to 65 years and a biopsyconfirmed intestinal metaplasia in a recent (less than six months) upper gastrointestinal endoscopy.
Exclusion criteria were coagulopathy, pre-existing or newly diagnosed gastric cancer, grossly visible mucosal lesions in a previous endoscopy, recent upper gastrointestinal bleeding or surgery leading to anatomical changes of the stomach, and comorbid diseases that may affect tolerance to upper endoscopy.
Patients were informed of the possible risks and benefits of participation in the study and written consent was obtained from each patient before enrollment.The Shiraz University of Medical Science's ethics committee approved the study.Repeat endoscopy was not only used in the study, but also was part of each patient's routine follow-up for neoplasia.
Demographic data including sex and age, any history of previous gastric endoscopic lesions, previous pathology and/or surgery of stomach, family history of gastric cancer, drug history, and clinical gastrointestinal signs and symptoms were recorded.
Patients underwent the endoscopy procedure using the Olympus GF-Q160 endoscope (Olympus, Japan).After standard sedation with 5 mg of intravenous midazolam, additional sedation was given if needed during the procedure.Continuous monitoring of pulse and oxygen saturation was performed during the procedure.

Conventional endoscopy
Each procedure was performed in two phases: the first phase was a conventional endoscopy with a meticulous search of the gastric mucosa for grossly visible abnormalities.Washing with tap water and the use of anti-foam agents was performed when needed.All grossly visible lesions were sampled first; however, if no gross lesion was found, six random biopsies were taken (two from the antrum, two from the body and two from the fundus) according to the Sydney protocol (13).

Chromoendoscopy
The second phase was performed during the same endoscopic session without withdrawing the endoscope.First, all biopsy sites were vigorously washed with tap water and residual blood or fluid was removed with suction.The surface to be examined was sprayed with 20 mL of 10% N-acetyl cystein (Aurum Pharmaceuticals Ltd, UK) using a specialized spray catheter (Endoflex, GmbH Germany) and all air was suctioned afterward.The N-acetyl cystein was allowed to remain in contact with the gastric surface for 3 min.Thereafter, 30 mL of 5% methylene blue solution (Aguattant Ltd, UK) was sprayed onto the gastric mucosa and was allowed to remain in contact for 3 min followed by extensive irrigation with tap water.Multiple biopsies were taken from all blue-stained spots regardless of whether biopsies were taken from the same spot during the conventional upper endoscopy.

Histopathology
The mean elapsed time for the complete procedure was 20 min.There were no complications related to the procedure.Biopsies were labelled with alphabetic codes only and referred to an expert pathologist blind to the patient's name or procedure type.Biopsies that showed intestinal metaplasia were considered to be positive.To control for the possible effect of biopsy artefacts on visualization during the first phase, and staining or biopsies at the second phase endoscopies, all procedures were supervised by a gastroenterologist who was not a coauthor of the present study and was blind to the study protocol.

Statistical analysis
Data were analyzed by two-tail Fisher's exact and c 2 tests using SPSS software version 12.0 (SPSS Inc, Japan).A value of P<0.05 was considered to be statistically significant.

RESULTS
Of 33 patients, 22 were men and 11 were women, with a mean (± SD) age of 49.73±13.10years.Chromoendoscopy yeilded 40 positive biopsies compared with 16 positive biopsies when the conventional endoscopy method was used.The number of positive biopsies from the antrum, body and fundus was 18, 15 and seven, respectively, using chromoendoscopy, and 10, four and two positive biopsies, respectively, from the same sites using conventional endoscopy.
The results from chromoendoscopy and the conventional method were concordant in 24 patients (72.7%) (13 patients had intestinal metaplasia detected by both methods and 11 had negative results using both methods).In the remaining nine cases, the results from the two methods were discordant.Eight patients had positive results for intestinal metaplasia with chromoendoscopy while their pathology reports did not show any evidence of intestinal metaplasia using the conventional method.One patient had a positive biopsy detected by the conventional method while chromoendoscopy failed to detect any abnormality (Figure 1).The difference between chromoendoscopy-based biopsies versus conventional biopsies for the detection of intestinal metaplasia was statistically significant (P=0.003).The calculated sensitivity (considering all enrolled cases as established positive cases) was 63.6% for chromoendoscopy versus 42.4% for conventional endoscopy.
To assess the possible difference in efficacy of the two methods in the determination of the anatomical extent of histological abnormalities, the stomach was divided into three anatomical segments: the antrum, body and fundus.Chromoendoscopy was significantly more effective in demonstrating the involvement of additional anatomical segments (P=0.002)(Table 1).
Even in the group of patients with concordant positive results with both methods (13 patients), chromoendoscopy remained significantly more effective in demonstrating involvement of additional anatomical segments (P=0.01)(Table 2).
If superiority is defined as either the recognition of a histological abnormality in a patient with a negative result using the other method, or the identification of additional involved anatomical segments, chromoendoscopy was superior in 16 patients (48.5%), while conventional endoscopy was superior in one patient (3%), and both were equally effective in 16 patients (48.5%) (Figure 2).

DiSCUSSiON
Detection of early cancers or precancerous lesions in the stomach may lead to better prognosis and survival (14).However, establishment of an early detection or screening program is difficult, even in high-risk groups, because of the low validity and reliability of conventional endoscopic techniques for the detection of subtle histological changes.The same problems are encountered when determination of the extent of histological abnormalities is required to plan a surgical resection or endoscopic ablation therapy (15,16).
Staining of the gastric mucosa with methylene blue to identify areas of nonacid-producing, mostly absorptive mucosa (ie, chromoendoscopy) has been used for a relatively long time (10,12,13,17,18).
Chromoendoscopy is relatively simple and uses normal endoscopic equipment.However, its superiority over conventional endoscopy is rarely or never subjected to trial.The present study clearly demonstrated that in comparison with conventional endoscopy, chromoendoscopy is a superior method for the detection of at least one of the premalignant gastric lesions (intestinal metaplasia).We showed that staining with methylene blue is helpful for the accurate delineation of the anatomical extent of histological abnormalities in the stomach.This ability may be particularly helpful when considering recent trends in the nonsurgical (ie, endoscopic) treatment of these lesions.The accurate mapping of abnormal areas is of the utmost importance to allow for confident and reliable treatments when using noninvasive methods.
Based on the results of the present study, chromoendoscopy using methylene blue is an effective method for the detection and accurate delineation of at least some of the precancerous lesions of the stomach.Chromoendoscopy can be recommended when a follow-up endoscopy for confirmation of an initial abnormal result or mapping of abnormal area(s) is planned.Although no case of dysplasia or cancer was encountered in the present study (which may limit the application of the results to these more serious lesions), the superiority of chromoendoscopy for the detection of one of the premalignant gastric lesions (ie, intestinal metaplasia) is promising.
It should be noted that there may be a slight risk of bias toward chromoendoscopy because during chromoendoscopy, the endoscopist was already aware of the results of the conventional endoscopy.However, this risk should be minimal considering that biopsies during the chromoendoscopy phase were strictly limited to the stained areas and all procedures were supervised by an independent observer.
Because only patients with previous positive biopsy results were included, the specificity of these procedures cannot be calculated based on data from the current study.The calculated sensitivity should also be regarded as an approximation, and further studies including consecutive patients without a previous diagnosis of metaplastic lesion are needed for the above mentioned purposes.
Further studies with a larger number of patients with more diverse risks for other gastric histological abnormalities, ideally including a wider spectrum of histological changes, are needed to completely elucidate the role of chromoendoscopy in screening or follow-up of patients after surgical or endoscopic treatment.