| Objective:Endoscopic(EGD)biopsy is the gold standard in the diagnosis of esophageal cancer,but traditional endoscopy is expensive and dependent on the endoscopic practitioner,and universal screening in the population is not feasible.In the area of high incidence of esophageal cancer,we urgently need to develop a simple,less traumatic,accurate,safe and acceptable new method for early esophageal cancer screening.Cytosponge combined with artficial intelligence(AI)is a new method for esophageal cancer screening.In this study,Cytosponge,a new non-invasive cell sampling device,will be used to collect samples for patients with high risk of esophageal cancer,make pathological sections of esophageal cells,and analyze pathological reports using AI.To evaluate the accuracy,safety and acceptability of this new diagnostic method for esophageal cancer,taking EGD as the control,in order to provide a new feasible method for the early diagnosis and screening of esophageal cancer.Mthods:This prospective study was conducted in Ganzhou People’s Hospital from July 2022 to February 2023.Subjects meeting the standards were recruited and selected for sampling using Cytosponge method,which was used to make pathological sections of cell samples.AI scanning was used to analyze the sections to identify labeled abnormal cells,and pathological reports were confirmed by 2 pathologists.Subj ects were required to undergo EGD examination within 5 days after the completion of sampling,and endoscopic biopsy was performed for the diagnosis of suspected esophageal lesions.The accuracy,acceptability and safety of the experiment were evaluated by analyzing the questionnaire data before and after Cytosponge experiment,pathological results of Cytosponge and EGD biopsy,and adverse event records.(1)Acceptability:Public evaluation data on Cytosponge experiment were recorded before the experiment to evaluate the acceptability of the public,and survey data of subjects’satisfaction with the experiment process were recorded to evaluate the subj ects’ satisfaction with the experiment process.After the experiment,VAS pain scores of Cytosponge,normal 5 and p ainless EGD were recorded and compared respectively to evaluate the acceptability ofCytos ponge experiment(2)Accuracy:Pathological diagnosis results of Cytosponge combined with artificial intelligence and EGD biopsy results were compared and analyzed,and EGD pathological biopsy was used as the standard to evaluate the accuracy of this diagnostic method.(3)Safety:Adverse events and sample quality of Cytosponge experiment were recorded and analyzed to evaluate the safety of the experiment.Results:A total of 100 eligible subjects were recruited from July 2022 to February 2023.All of them(47 males and 43 females aged 45-75 years)successfully swallowed cytosponge capsules on the first try and subsequently completed endoscopy,including 73 painless gastroscopes and 27 normal gastroscopes.80 patients underwent EGD biopsy and 20 patients refused EGD biopsy.All subjects completed the pre-experiment questionnaire,Cytosponge experiment satisfaction survey and VAS score.(1)The diagnostic result of Cytosponge combined with AI had good diagnostic accuracy;Cytosponge experiments reported 2 cases of squamous cell carcinoma(ESCC)(2.50%),6 cases of low or high grade neoplasia(7.50%)and 9 cases of atypical squamous cell hyperplasia(11.25%).Compared with the pathological report of EGD biopsy,the sensitivity(TPR)and specificity(TNR)of the diagnosis(LGIN+HGIN+ESCC is positive)were 58.3%and 98.5%respectively without the number of labeled abnormal cells grouped,which could eliminate negative cases well.When the value of Cytosponge-AI diagnosis on the number of abnormal labeled cells was ≥5,the diagnostic accuracy was greatly improved with sensitivity of 91.6%and specificity of 94.1%.It can effectively identify severe atypical hyperplasia(ASC)and low-grade neoplasia(LGIN)+high-grade neoplasia(HGIN)+squamous cell carcinoma(ESCC).(2)Cytosponge experiment is acceptable:In general,most of the public believe Cytosponge is an acceptable screening method rather than passive and passive resistance.About 90%of subjects are satisfied with the sampling process.VAS score:Overall acceptance of Cytosponge was satisfactory,with a mean score of 8.23+0.59 and a median score of 8.1,which was higher than that of routine examination(mean 5.6+0.92,median 5.9,-6P<0.001)and lower than that of sedative endoscopy(mean 8.45+0.52,median 8.4).P<.001).(3)Cytosponge experiment is safe:All participants successfully swallowed the sponge capsule on the first time without repeated swallowing or sore throat in 2(2%)subjects,vomiting in 2(2%)subjects,and no adverse events such as rope disconnection and upper gastrointestinal bleeding.Conclusion:Cytosponge combined with artificial intelligence is a feasible,safe and acceptable new diagnostic method for early esophageal cancer screening. |
PDF Full Text Request