Diagnosis Of Colorectal Polyps And Atrophic Gastritis With Confocal Laser Endomicroscopy

Part 1 The Comparison between Confocal Laser Endomicroscopy and Chromoendoscopy for Diagnosing Colorectal PolypsBackground:Colorectal cancer is the third most common cancer, and the fourth most common cancer caused death in the world. There are more than 1.2 million new cases each year, and caused 600,000 death in the world. With the improvement of surgery and chemotherapy, more and more people received effective treatment, however, the 5-year survival rate is still low, which is close to 65% in developed country and less than 50% in developing country. Therefore, to detect the cancer in early stage is the most important thing.Colorectal cancer progressed through two mainly ways. The first one is "adenoma-carcinoma" classical sequence, which also the is most important sequence. APC gene mutations are an early event in the multistep process of colorectal cancer formation and occur in more than 70% of colorectal adenomas. The adenoma-carcinoma sequence is further promoted by activating mutations of the KRAS oncogene and inactivating mutations of the TP53 tumour suppressor gene. A lot prospective or retrospective studies have confirmed that people suffering from adenomatous polyps have a higher risk of progressing to colorectal cancer. The second sequence is "serrated adenoma-cancer". WHO (2010) new classification for digestive tumor distinguished sessile serrated adenoma/ polyp(SSA/P) from hyperplastic polyps, and defined it as one of the pre-cancer lesions. There are multiple mutations during SSA/P progression, which includes BRAF, KRAS mutation, CpG island methylation (CMP).In a long-term prospective study, the researchers found that colonoscopic removal of adenomatous polyps prevents death from colorectal cancer. Compared to general population, people with adenomas resection had a 53% reduction in mortality from colorectal cancer. It also suggests colon endoscopy play a important role.There is still some problem. A lot polyps can be detected during colonoscopy which include more than 50% hyperplastic polyps, however, adenomas cannot be distinguished from hyperplastic polyps by white-light colonscopy. Removal of all polyps, especially the lots of small hyperplastic polyps is unnecessary, which also cause related complications and waste of resources. The best way to solve this problem is an accurate pathological diagnosis in vivo. Furthermore, "resect and discard" strategy also requires accurate pathological diagnosis in vivo.In summary, either the distinguishing adenomas from non-adenomas or "resect and discard" strategy both need a complete and accurate pathological diagnosis in vivo. Confocal laser endomicroscopy is a new developed technology, which can provide an "optical biopsy" in vivo, chromoendoscopy as a mature technology also enables pathological diagnosis in vivo.Hence, the purpose of this study is to evaluate the diagnostic value of Confocal laser endomicroscopy and chromoendoscopy through a single-center, prospective, paired design, double-blind, controlled trial.Method1. PatientsInclusion criteria:patients more than 18 years old, with polyps found by colonscopy. Exclusion criteria:(1) has a serious heart or lung failure; (2) coagulation disorders caused by blood disease or other disease; (3)severe liver and kidney dysfunction; (4) pregnancy or breast-feeding women, or prepared for pregnancy; (5) suffering from mental illness;(6) allergy to fluorescein sodium;(7) patients who don’t agree to involve in clinical trials.2. Endoscopy Confocal laser endomicroscopy(EC3870-CIK,Pentax). Magnifying colonscopy (EC-590,Fujinon)3. Endoscopic ProceduresPatients included in the study accepted both confocal laser endomicroscopy and chromoendoscopy examination. Random numbers were used to decide which one the patients accepted first. Confocal laser endomicroscopy and chromoendoscopy were performed by different endoscopists. Diagnosis were made according to their observations, pathology served as the gold standard.4. Outcome MeasurementsThe primary outcome measurements:sensitivity and specificity of confocal laser endomicroscopy and compared with chromoendoscopy for diagnosing adenomatous polyp. The second outcome measurements:the value of confocal laser endomicroscopy for classifying adenomas and grading intraepithelial neoplasia.5. Diagnostic Criteria5.1 Criteria for Diagnosing Adenomas by Confocal Laser EndomicroscopyAdenomas:Ridged-lined irregular epithelial layer with loss of crypts and goblet cells; irregular cell architecture with littleor no mucin. SSA/P:more dilated crypt openings and more serrated than the hyperplastic polyps, the middle of the glands is wider than hyperplastic polyps.5.2 Criteria for Classifying Adenomas by Confocal Laser EndomicroscopyAccording to pathological classification criteria, we made a classification criteria: tubular adenoma was diagnosed if most of confocal images presented tubular structure; villous adenoma was diagnosed if most of confocal images presented villous structure; tubular villous adenoma was diagnosed if the two structure both existed.5.3 Criteria for Grading Intraepithelial Neoplasia by Confocal Laser EndomicroscopyWe made a scoring system, which included the gland morphology, gland epithelial stratification and vascular structures, each parameters score 1-3.5.4 Criteria for Diagnosing Adenomas by ChromoendoscopyKudo’s pit pattern criteria:pit Ⅰ normal mucosa; pit Ⅱ hyperplasic polyp; pit Ⅱ-O SSA/P; adenomatous polyps for pit ⅢL, Ⅲs, Ⅳ, Ⅴi, Ⅴn.6. Histologic AssessmentAll polyps were removed after the observations by two endoscopy, pathology examinations were then followed. In addition, the serrated lesions were diagnosised according to WHO (2010) new classification for digestive tumor.7. Statistical AnalysisAccuracy, sensitivity, specificity were estimated for each method (CLE, CE) along with 95% confidence intervals. Comparisons between CLE and CE regarding accuracy, sensitivity, specificity were made using the McNemar test for clustered data. Comparison between two independent sample were made using Pearson chi-square test. Statistical analyses were performed by using SPSS18.0(SPSS, Chicago, IL) for windows.Result1. General Characteristics of PatientsA total of 121 patients with 161 polyps were include during January 2014 to October 2014, which included 86 males and 35 females. The polyps location were as follow:four at ascending colon, eight at transverse colon,15 at descending colon,60 at sigmoid colon,74 at rectum. These polyps contained 107 adenomas and 54 non-adenomas.2. Comparison between Chromoendoscopy Pit Pattern and Pathological FindingsAccording to chromoendoscopy, there are 7 pit Ⅰ including one inflammatory polyps,1 juvenile polyps and 5 hyperplastic polyps; 57 pit Ⅱ including one SSA/P,17 tubular adenomas and 33 hyperplastic polyps; Seven pit Ⅱ-O including 3 SSA/P and 4 hyperplastic polyps; As for pit ⅢL, most of them are adenomas except four inflammatory polyps; pit Ⅳs, pit Ⅴi are both one cases and both high grade intraepithelial neoplasia; There are one pit Vn cases and are colon cancer according to pathology.3. Comparison between Confocal Laser Endomicroscopy and Pathological FindingsAll the polyps are diagnosed as adenoma or non-adenoma by confocal laser endomicroscopy according to Mainz criteria. There are 54 non-adenomatous polyps (confocal endoscopy), and 5 of them are misdiagnosed including one SSA/P (pathology) and four tubular adenomas (pathology).There are 107 adenomatous polyps (confocal endoscopy),and 5 of them are misdiagnosed including one SSA/P (pathology) and 4 inflammatory polyps (pathology).4. Comparison between Confocal Laser Endomicroscopy and Chromoendoscopy for Diagnosing AdenomasThe sensitivity, specificity, accuracy, NPV and PPV of chromoendoscopy are 83.18%,85.19%,83.85%,71.88%and 91.75% respectively; The sensitivity, specificity, accuracy, NPV and PPV of confocal laser endomicroscopy are 95.33%, 90.74%,93.79%,90.74%,95.33% respectively. The sensitivity, accuracy and NPV of confocal laser endomicroscopy are higher than those of chromoendoscopy (P values were 0.001,0.041,0.01 respectively), however, specificity and PPV have no difference between two endoscopy.5. Comparison between Confocal Laser Endomicroscopy and Chromoendoscopy for Diagnosing SSA/PAccording to Pit II-O by chromoendoscopy, the sensitivity, specificity and accuracy for diagnosing SSA/P are 60%,90.48%,87.23% respectively. According to the characters of SSA/P we developed, the sensitivity, specificity and accuracy of confocal laser endomocroscopy for diagnosing SSA/P are 80%、97.62%、95.74% respectively.6. Classification of Adenomas by Confocal Laser EndomicroscopyAccording to the criteria for classifying adenomas we developed, the accuracy of endoscopists 1 are 88.54%(85/96), the agreement between endoscopists 1 and pathology were 0.724; the accuracy of endoscopists 2 are 83.33%(80/96), the agreement between endoscopists 2 and pathology were 0.610;7. Grading Intraepithelial Neoplasia by Confocal Laser EndomicroscopyWe developed a scoring system, ROC curve analysis suggest 8 as the cut-off, which means that the diagnosis of high-grade intraepithelial neoplasia will be made if the score≥8, pathology served as golden stander. The sensitivity, specificity and accuracy of endoscopists 1 for predicting high-grade intraepithelial neoplasia were 88.24%,67.06%,70.59% respectively. The sensitivity, specificity and accuracy of endoscopists 2 were 88.24%,64.71%,68.63% respectively.8. Risk Factors of the Misdiagnosis by EndoscopyAs for confocal laser endomicroscopy, the misdiagnosis rate has no difference(p=0.121) between diameter ≤5mm group (12.20%,5/41) and diameter> 5mm group(4.3%,5/115). As for chromoendoscopy, the misdiagnosis rate is higher in diameter ≤5mm group(28.26%,13/46) than that in diameter> 5mm group (11.3%,13/115)(p=0.008).Conclusion:1. The sensitivity, accuracy and NPV of confocal laser endomicroscopy are higher than those of chromoendoscopy, however, specificity and PPV have no difference between two endoscopy.2. Confocal laser endomicroscopy and chromoendoscopy both can be used toclassify serrated lesions, but the accuracy need to be further confirmed.3. Confocal laser endomicroscopy can be used to classify adenomas.4. The sensitivity of confocal laser endomicroscopy for grading intraepithelial neoplasia is acceptable however, the specificity is low.5. The misdiagnosis rate of confocal laser endomicroscopy is not related to the diameter of polyp, however, The misdiagnosis rate of chromoendoscopy is more higher in small polypsPart 2 The Accuracy of Confocal Laser Endomicroscopy, Narrow Band Imaging, and Chromoendoscopy for the Detection of Atrophic GastritisBackground:Gastric Cancer (GC) is one of the most common cancer, and the second most common cancer caused death in the world. With the improvement of surgery and chemotherapy, more and more people received effective treatment, however, the 5-year survival rate is still low. Therefore, to detect the cancer in early stage and precancerous lesions is an effective methods to prevent advanced GC and reduce death caused by GC.Chronic gastritis is a chronic inflammation lesions of gastric mucosa, which character by mucous infiltrated mainly with lymphocytes and plasma cells, neutrophils and eosinophils can also exist. There are two mainly categories:the chronic non-atrophic gastritis and chronic atrophic gastritis.Atrophic gastritis, especially metaplastic atrophic gastritis is defined as precancerous lesions. Detection of atrophic gastritis and subsequently interventions are the effective methods to prevent gastric cancer. Endoscopic biopsy is an important methods to detect atrophic gastritis, however, due to the low detection rate for atrophic gastritis under white light endoscopy, the positive biopsy rate is also low.Confocal laser endomicroscopy is a new developed technology, which can provide an "optical biopsy" in vivo, chromoendoscopy as a mature technology also enables pathological diagnosis in vivo.Hence, the purpose of this study is to evaluate the diagnostic value of Confocal laser endomicroscopy, chromoendoscopy and NBI through a single-center, prospective, paired design, double-blind, controlled trial. All the patients included will accept examinations of CLE,CE and NBI.Method:1. PatientsInclusion criteria:(1) patients more than 18 years old; (2) previous or present H. pylori infection;(3)previous atrophic gastritis; (4)others with upper gastrointestinal symptoms, such as bile reflux, long-term use of non-steroidal anti-inflammatory drugs.Exclusion criteria:(1)serious heart and lung disorder; (2) the coagulation dysfunction; (3) uncorrected digestive bleeding; (4) advanced gastric cancer; (5) serious liver and kidney dysfunction;(6) pregnancy or breast-feeding women, or prepared for pregnancy; (7) suffering from mental illness; (8) allergy to fluorescein sodium; (9) patients who don’t agree to involve in clinical trials.2. EndoscopyConfocal laser endomicroscopy(EC3870-CIK,Pentax). Magnifying colonscopy (GIF-H260Z or GIF-Q240Z, Olympus).3. Endoscopic ProceduresThis study is designed to evaluate diagnostic value of confocal laser endoscopy, chromoendoscopy and NBI for diagnosing atrophic gastritis. Patients included will also accept the examinations of confocal laser endomicroscopy, chromoendoscopy and NBI. Pathology serves as the golden standard.NBI is the first endoscopy to be performed, then followed by chromoendoscopy, at last is confocal laser endomicroscopy. NBI and chromoendoscopy is performed by the same endoscopist using GIF-H260 or GIF-Q240Z, confocal laser endomicroscopy is performed by the other endoscopist using Pentax EC3870-CIK.4.Outcome MeasurementsThe primary outcome measurements:sensitivity and specificity of confocal laser endomicroscopy for diagnosing atrophic gastritis. The second outcome measurements:the value of confocal laser endomicroscopy for classifying atrophic gastritis.5. Diagnostic Criteria5.1 Criteria for Diagnosing Atrophic Gastritis by Chromoendoscopy and NBItype A, small round pits, normal mucous in gastric body or fundus. type B, continuous short rod-like pits, normal antral mucosa. type C, elongated pits, atrophic gastritis, type D reticular pits with dilated openings, atrophic gastritis. type E villous-like pits, metaplastic atrophic gastritis.5.2 Criteria for Diagnosing Atrophic Gastritis by CLENon-metaplastic atrophy features:the number of pit decreasing, and pits prominent dilated; metaplastic atrophy features:villious-like pits with goblet cells, columnar absorptive cells and brush border.6. Histologic AssessmentAll specimens underwent standard histopathology assessment by a experienced gastrointestinal pathologists. All types of gastritis were defined based on the updated Sydney System criteria.7. Statistical AnalysisAccuracy, sensitivity, specificity were estimated for each method (CLE, CE and NBI) along with 95% confidence intervals. Comparisons between CLE and CE regarding accuracy, sensitivity, specificity were made using the McNemar test for clustered data. Comparison between two independent sample were made using Pearson chi-square test. Statistical analyses were performed by using SPSS18.0(SPSS, Chicago, IL) for windows.Result:1. General Characteristics of PatientsA total of 86 patients with 253 sites were recruited from November 2012 to July 2013, which includs 48 males and 38 females, the mean age is 49 years (21-75). In a total of 253 observation site,123 sites are non-atrophic gastritis,130 sites are atrophic gastritis including 68 non-metaplastic atrophy and 62 metaplastic atrophy.2. The Agreement between CE and NBI for Classifying Pit PatternThere is excellent agreement between Chromoendoscopy and NBI for Classifying Pit Pattern (kappa=0.904,95% CI 0.857-0.945). In details, there is totally agreement of type A and type E and only slightly difference of type B, type C, type D.3.Comparison between NBI Pit Pattern and Pathological FindingsAll the Type A and most of the Type B (78%) according to NBI are non-atrophic gastritis. There are 80 Type C according to NBI, which includes 24 non-atrophic gastritis and 56 atrophic gastritis (43non-metaplastic atrophy,13metaplastic atrophy). There are 51 Type D including 22 non-metaplastic atrophy and 24 metaplastic atrophy; All the 6 type E according to NBI are metaplastic atrophy.4.Comparison between CE Pit Pattern and Pathological FindingsThe result is similar to that of NBI. All the type A and most of the type B (79%) according to chromoendoscopy is non-atrophic gastritis. There are 79 type C including 57 atrophic gastritis, and there are 50 type D including 4 non-atrophic gastritis and 46 atrophic gastritis.5.Comparison between CLE, CE and NBI for Diagnosing Atrophic GastritisThe sensitivity, specificity and accuracy of CLE for diagnosing atrophic gastritis are 92.31%,86.18%,89.33% respectively. The sensitivity, specificity and accuracy of CE are 83.85%,78.86%,81.42% respectively. The sensitivity, specificity and accuracy of NBI are 83.08%、76.42%、79.84% respectively. The sensitivity, specificity, accuracy and NPV of CLE are higher that of CE (P value are 0.035,0.049,0.002,0.038 respectively), the NPV has no difference between CLE and CE.6.Classification of Atrophic Gastritis by CLEThe sensitivity, specificity and accuracy of CLE for diagnosing non-metaplastic atrophy are 86.76%,91.89%,90.51% respectively. The sensitivity, specificity and accuracy for diagnosing metaplastic atrophy are 91.94%、96.86%、95.65% respectively. There is excellent interobserver agreement between two endoscopists (kappa= 0.895).Conclusion:1. Chromoendoscopy has excellent agreement with NBI for classifying gastric pits.2. Chromoendoscopy and NBI can predict atrophic gastritis according to gastric pits, but can not classify atrophic gastritis.3. The sensitivity, specificity and accuracy of confocal laser endomicroscopy are higher than that of chromoendoscopy for diagnosing atrophic gastritis.4. Confocal laser endomicroscopy can distinguish metaplastic atrophic gastritis from non-metaplastic atrophic gastritis.