A multicenter validation of an endoscopic classification with narrow band imaging for gastric precancerous and cancerous lesions

BACKGROUND AND STUDY AIM

The reliability and external validity of narrow band imaging (NBI) in the stomach have not been described consistently. The aim of the current study was to describe and estimate the accuracy and reliability of a simplified classification system for NBI in the diagnosis of gastric lesions.

METHODS

Consecutive patients undergoing NBI endoscopy at two reference centers (n=85, 33% with dysplasia) were included in two studies. In total, 224 different areas were biopsied and recorded onto video. In the derivation study, previously described NBI features were analyzed in order to develop a simplified classification. In the validation study the accuracy and reliability of this classification were estimated among three groups of endoscopists with different levels of expertise in NBI.

RESULTS

The reliability/accuracy results from the derivation study allowed the creation of a simplified NBI classification. In the validation study, "regular vessels with circular mucosa" (pattern A) was associated with normal histology (accuracy 83%; 95% confidence interval [CI] 75 %-90%); "tubulo-villous mucosa" (pattern B) was associated with intestinal metaplasia (accuracy 84%; 95CI 77%-91%; positive likelihood ratio [LR+]=4.75); and "irregular vessels and mucosa" (pattern C) was associated with dysplasia (accuracy 95%; 95CI 90%-99%; LR+=44.33). The reproducibility of these patterns was high (k=0.62). "Light-blue crest" was moderately reliable (k=0.49) but specific (87%) for intestinal metaplasia. A variable vascular density (additional pattern+) was the best feature for Helicobacter pylori gastritis (accuracy 70%; 95CI 59%-80%) but showed only fair reliability (k=0.38). Non-experienced endoscopists presented lower agreement (k=0.6 vs. k=0.75) and accuracy (74% vs. 86%) than international experts/experienced endoscopists.

CONCLUSION

A simplified NBI classification is accurate and reliable for the diagnosis of intestinal metaplasia and dysplasia. The classification should be further assessed and validated on a per-patient assessment of NBI, and by comparing NBI with other imaging technologies.

Management of precancerous conditions and lesions in the stomach (MAPS): guideline from the European Society of Gastrointestinal Endoscopy (ESGE), European Helicobacter Study Group (EHSG), European Society of Pathology (ESP), and the Sociedade Portuguesa de Endoscopia Digestiva (SPED)

Atrophic gastritis, intestinal metaplasia, and epithelial dysplasia of the stomach are common and are associated with an increased risk for gastric cancer. In the absence of guidelines, there is wide disparity in the management of patients with these premalignant conditions. The European Society of Gastrointestinal Endoscopy (ESGE), the European Helicobacter Study Group (EHSG), the European Society of Pathology (ESP) and the Sociedade Portuguesa de Endoscopia Digestiva (SPED) have therefore combined efforts to develop evidence-based guidelines on the management of patients with precancerous conditions and lesions of the stomach (termed MAPS). A multidisciplinary group of 63 experts from 24 countries developed these recommendations by means of repeat online voting and a meeting in June 2011 in Porto, Portugal. The recommendations emphasize the increased cancer risk in patients with gastric atrophy and metaplasia, and the need for adequate staging in the case of high grade dysplasia, and they focus on treatment and surveillance indications and methods.

Segmentation for classification of gastroenterology images

Automatic classification of cancer lesions in tissues observed using gastroenterology imaging is a non-trivial pattern recognition task involving filtering, segmentation, feature extraction and classification. In this paper we measure the impact of a variety of segmentation algorithms (mean shift, normalized cuts, level-sets) on the automatic classification performance of gastric tissue into three classes: cancerous, pre-cancerous and normal. Classification uses a combination of color (hue-saturation histograms) and texture (local binary patterns) features, applied to two distinct imaging modalities: chromoendoscopy and narrow-band imaging. Results show that mean-shift obtains an interesting performance for both scenarios producing low classification degradations (6%), full image classification is highly inaccurate reinforcing the importance of segmentation research for Gastroenterology, and confirm that Patch Index is an interesting measure of the classification potential of small to medium segmented regions.

Quality reporting of endoscopic diagnostic studies in gastrointestinal journals: where do we stand on the use of the STARD and CONSORT statements?

BACKGROUND AND STUDY AIMS

Many papers have been published in the field of diagnostic endoscopy in the last few years. However, there are no reports on their quality. The aim of this study was to evaluate quality in recently published endoscopic articles.

MATERIALS AND METHODS

The study reviewed published articles on diagnostic endoscopy from 1998 to 2008. Quality was assessed and independently quantified by two observers using the STARD (STandards for the Reporting of Diagnostic accuracy studies) and CONSORT (Consolidated Standards for Reporting of Trials) statements. The interobserver proportion of agreement and kappa coefficient were estimated.

RESULTS

A total of 120 articles comprising 10 randomized controlled trials and 110 diagnostic accuracy studies were evaluated. Most studies related to colonic polyp detection (30 %) or evaluation of Barrett's esophagus (29 %). Chromoscopy (45 %), fluorescence (21 %), and narrow-band imaging (14 %) were the technologies most often evaluated. The mean number of items (i. e. standard requirements) fulfilled by the randomized controlled trials was 15.7 +/- 2.2 out of 22 while for the diagnostic accuracy studies it was 12.2 +/- 3.6 out of 25. Reporting of study results was complete in 90 % of the randomized controlled trials, but only 65 % of the diagnostic accuracy studies presented a cross-table of results. The global proportion of agreement between observers was 97 % in randomized controlled trials and 95 % in diagnostic accuracy studies.

CONCLUSIONS

Recent publications in diagnostic endoscopy achieve only medium quality according to the available statements. It seems that it would be useful for authors, reviewers, and editors to be familiar with and apply these statements. The development of a specific checklist for diagnostic endoscopy publications might be helpful toward achieving better quality reporting in the future.

Towards more adequate colour histograms for in-body images

Although there is a growing number of scientific papers describing classification of in-body images, most of it is based on traditional colour histograms. In this paper we explain why these might not be the most adequate visual features for in-body image classification. Based on a colour dynamic range maximization criterion, we propose a methodology for creating more adequate colour histograms, testing it on a vital-stained magnification endoscopy scenario.

"Downhill" varices. A rare cause of esophageal hemorrhage

"Downhill" varices or upper esophageal varices are a rare cause of proximal digestive tract hemorrhage with only 16 cases described in the literature. In our series, hemorrhage due to "downhill" varices represents 0.1% of all acute esophageal variceal bleeding. Their etiology differs from that of the usual "uphill" varices secondary to portal hypertension, and the clinical management should be directed to vascular obstruction if present. We report a case of an 89-year-old male with hemorrhagic "downhill" varices not associated, as usually, with superior vena cava obstruction or compression, but with severe pulmonary hypertension and drug-related hemorrhagic risk factors, whose removal proved sufficient to prevent rebleeding.