| BackgroundHelicobacter pylori is a microaerobic Gram-negative bacillus that can cause gastrointestinal diseases,such as chronic active gastritis,peptic ulcer,gastric mucosa-associated lymphoid tissue lymphoma,and gastric cancer.About 50%of people worldwide have H.pylori infections.Early diagnosis of H.pylori infection and eradication are important for treating chronic active gastritis and reducing the recurrence of peptic ulcer and the occurrence of gastric cancer.Invasive and non-invasive methods are currently used to diagnose H.pylori infection.The main invasive methods are the Rapid Urease Test（RUT）,histopathological diagnosis,endoscopic diagnosis,and molecular analysis;the main non-invasive methods are the Urea Breath Test（UBT）,Stool Antigen Test,and serological examination.In China,the cost of an endoscopic examination is relatively low and this procedure is widely accepted by patients.China also has a high rate of infection by H.pylori and a high incidence of gastric cancer,so if two tests can be combined and completed simultaneously using gastroscopy,H.pylori infection can be treated in a timely manner,precancerous lesions of gastric cancer and early gastric cancer can be detected,and medical costs can be reduced.Although invasive methods using gastroscopy provide these benefits,there are also has some shortcomings,in that the RUT is susceptible to false negatives depending on the load and focal distribution of H.pylori.However,histopathological and molecular biological tests cannot provide real-time results.After H.pylori infection,endoscopy can identify the unique characteristics of the gastric mucosa,and enable the diagnosis of H.pylori infection.Gastric mucosa that is positive for H.pylori typically has diffuse redness,spotty redness,mucosal swelling,with or without white turbid mucus,atrophy,intestinal metaplasia,disappearance of regular arrangement of collecting venules（RAC）,nodular changes,xanthoma,and hyper plastic polyps.Gastric mucosa that is negative for H.pylori typically has a RAC,and may have fundic gland polyposis,a red streak in the gastric antrum and gastric body,and attachment of a former bleeding spot.After H.pylori eradication,there are often localized red areas of various sizes on white mucosa,and redness after erosion healing appears as patchy redness.A previous study that used receiver operating characteristic（ROC）analysis found that use of conventional gastroscopy with white light imaging（WLI）for the diagnosis of H.pylori infection had an area under the curve（AUC）of 0.783,with a total coincidence rate of 78.46%,a sensitivity of 77.44%,a specificity of 79.31%,a positive predictive value（PPV）of 75.74%,and a negative predictive value（NPV）of 80.83%.This previous study was based on the endoscopic observations of 2 experienced endoscopists（each with more than 5 years of endoscopy experience and examination of more than 5000 cases using gastroscopy）.This previous study also reported that the accuracy of diagnosis depended on the experience of the endoscopist,and that diagnosis was time-consuming and adversely affected by fatigue.Therefore,objective,accurate,fast,and feasible endoscopic diagnosis of H.pylori is needed in clinical practice.Several studies reported that artificial intelligence（AI）provided accurate and standardized endoscopic diagnosis of H.pylori infection.Convolutional neural networks（CNNs）are commonly used for analysis of medical images.Deep learning mainly uses convolutional neural network（CNN）,which has been widely used in gastroenterology and is the most popular network architecture in image deep learning.For example,the real-time automatic detection system under colonoscopy built by CNN can increase the detection rate of polyps and adenomas under colonoscopy.The detection rate of polyps increased from 29.10%to 45.02%,and the detection rate of adenomas increased from 20.34%to 29.12%.Currently,the international recommended bismuth quadruple 14-day therapy can achieve about 90%eradication rate of H.pylori,but they still have limitations,such as complex program,high cost,many side effects,poor patient compliance,etc.In addition,the use of multiple antibiotics may aggravate the rise of resistance in the future.Therefore,it is still a clinical demand to find a treatment with high eradication rate,convenient method,few side effects and less influence on drug resistance.In China and the Asia-Pacific region,the resistance rate of H.pylori to amoxicillin（AMO）is very low（0%-5%）[1].As early as the late 1980s,the dual therapy consisting of omeprazole（20mg,1 or 2/day）and amoxicillin（1000mg,2/day）was initially proposed to treat H.pylori infection,but the eradication rate of this regimen varied from 55%to 62%.So far,domestic and foreign reports on the eradication rates of different dual therapies are inconsistent,and the conflicting results of eradication rates may be mainly attributed to the differences in the dose,frequency and course of PPI-amoxicillin.The present study established an endoscopic system for the diagnosis of H.pylori infection based on deep learning and applied it to a clinical setting.In particular,we compared the accuracy of gastroscopic diagnosis of H.pylori infection by endoscopists who were assisted by AI with diagnosis by endoscopists who were unassisted by AI,and then analyzed the possible clinical applications of this AI diagnostic system.After the diagnosis of H.pylori,timely and effective treatment of H.pylori is an important controllable means to treat chronic active gastritis,reduce the recurrence of peptic ulc er,and reduce the occurrence of gastric cancer.We conducted a retrospective study to compare the eradication rate and adverse drug reactions of different courses of high-dose dual therapy in the treatment of H.pylori.At present,the domestic and foreign reports on the eradication rate of different dual therapy are inconsistent,which is mainly attributed to the differences in the dose and duration of PPI-amoxicillin.It has been reported that increasing the dose and frequency of administration can achieve better H.pylori eradication rate,but it is still unclear whether different durations of dual therapy can also achieve satisfactory eradication rate.Methods1.Firstly,7377 images of 639 patients were included to construct an H.pylori gastroscopic image diagnosis system based on deep learning model.2.2080 images of 201 patients were included for model testing,and there was no overlap between test and model building data sets.3.The diagnostic ability of AI-assisted endoscopists was compared with that of non-AI-assisted endoscopists in the diagnosis of H.pylori.The main indicators were accuracy,sensitivity and specificity,and the secondary indicators were positive predictive value and negative predictive value4.Based on the previous two randomized controlled trials,a total of 523 patients were screened from the H.pylori specific outpatient clinic of Daping Hospital from January 2017to April 2019,and 220 patients met the inclusion criteria,including 104 in the 10-day group and 116 in the 14-day group.The primary outcome was H.pylori eradication rate,and the secondary outcomes were the incidence of adverse reactions and patient compliance.Results1.For the diagnosis of H.pylori infection,the CNN model had an accuracy of 89.6%,a sensitivity of 90.9%,and a specificity of 88.9%.The AUC for H.pylori positivity was84.1%（95%CI:73.0%,95.2%）and the AUC for H.pylori negativity was 90.3%（95%CI:82.2%,98.4%）.We also performed separate analyses of diagnosis based on the gastric body alone（upper body,middle body,lower body,and lesser curvature）and the antrum alone.Analysis of the gastric body provided significantly greater sensitivity（86.2%vs.19.8%）and accuracy（76.4%vs.68.4%）,but analysis of the antrum provided significantly greater specificity（93.3%vs.71.7%）;analysis of the antrum provided a significantly lower false-positive rate（6.7%vs.28.3%）,but analysis of the body provided a significantly lower false-negative rate（13.8%vs.80.2%）.Combining data from the body and antrum provided the best diagnostic performance.2.Comparisons of all 14 endoscopists in the AI-assisted and AI-unassisted groups indicated the AI-assisted group had significantly better accuracy（92.8%vs.75.6%,P<0.001）,sensitivity（91.8%vs.78.6%,P=0.032）,and specificity（93.4%vs.74.5%,P<0.001）.We then determined the accuracy of endoscopists in the AI-unassisted group who had different levels of experience,using two different metrics to define“junior”and“senior”status.The accuracy of diagnosis was 85.5%（47/55）for senior endoscopists who were trained in early cancer detection,and was 72.1%（111/154）for junior endoscopists who did not have this training（P=0.047）.The accuracy of diagnosis was 77.3%（75/97）for senior endoscopists who had 5 years of experience or examination of more than 5000cases,and was 73.2%（82/112）for junior endoscopists who did not have this experience（P=0.494）.3.In ITT analysis,the eradication rates of 10-day and 14-day high-dose dual therapy were 78.4%（95%CI 69.6%-87.2%,69/88）and 89.7%（95%CI 83.3%-96.2%,79/88）,respectively（p=0.039）.According to the PP analysis,the eradication rates were 80.0%（95%CI 71.3%-88.7%,68/85）and 92.9%（95%CI 87.4%-98.5%,79/85）,respectively（p=0.014）.The eradication rate in the 14-day group was significantly higher than that in the10-day group.There was no significant difference in adverse reactions between the two groups.Adverse reactions were mild and relieved spontaneously after drug withdrawal.There was no difference in the compliance between the two groups,but the compliance had an impact on the treatment effect.CYP2C19 gene polymorphism has no effect on the eradication rate of dual therapy.The resistance rates of metronidazole,clarithromycin and levofloxacin were high in this area,but the high resistance rates of each drug had no effect on the eradication rate of dual therapy.ConclusionIn conclusion,this study established an AI-assisted H.pylori diagnosis system under gastroscopy by learning the pictures of five parts of the stomach.This is the first time that the system has been applied to clinical practice in China.This system is more valuable for doctors without early cancer training,so it may be more suitable for primary care and areas with high incidence of gastric cancer and high penetration rate of gastroscopy.Second,propensity-matched analysis suggested that 14-day high-dose dual therapy could achieve a higher H.pylori eradication rate than 10-day high-dose dual therapy,and the two treatment courses had a lower incidence of adverse reactions.Therefore,14-day courses of dual therapy with a PPI and amoxicillin are recommended. |
PDF Full Text Request