Surveillance Strategies Of Individuals With Esophageal Low Grade Intraepithelial Neoplasia And Negative Findings After Endoscopy Screening

ObjectivesIn order to optimize the surveillance strategy for individuals with esophageal low-grade intraepithelial neoplasia(LGIN)and negative findings after a single endoscopy screening,based on a multicenter,population-based EC screening cohort in rural China,this study aimed to profile their baseline endoscopy findings,estimate the long-term risk of esophageal cancer(EC)incidence and EC mortality,evaluate the progression and regression patterns and evaluate the endoscopy surveillance effectiveness in patients with LGIN,and to identify cost-effectiveness management strategies.Materials and Methods1.In this part,we included participants aged 40-69 years who were estimated high-risk individuals for EC and underwent endoscopy examinations during 2007-2016 in 13 screening centers in Jiangsu Province,Anhui Province,Shandong Province,and Henan Province.The proportions of negative endoscopy findings,mild dysplasia,moderate dysplasia,severe dysplasia/carcinoma in situ(SD/CIS),and EC were estimated.The generalized and linear regression models were used to evaluate the individual and systematic factors affecting the detection rate of severe dysplasia and above(SDA).2.In this part,individuals diagnosed with negative endoscopy findings,mild dysplasia,and moderate dysplasia in 2007-2012 were included and were prospectively followed up until December 31,2021.We plotted cumulative incidence function curves to estimate the cumulative EC incidence and mortality.Sub-distribution hazard models were used to calculate the adjusted hazard ratios(HRs)and 95%confidence interval(CI).For individuals with negative endoscopy findings,we calculated the standardized incidence ratio(SIR)and standardized mortality ratio(SMR)based on the EC incidence and EC mortality in rural China in 2010.3.This part included patients with LGIN diagnosed in 2010-2016 and offered endoscopic surveillance until December 31,2021.Rates of progression,remaining,and regression were estimated,and Kaplan-Meier analyses were used to estimate the cumulative incidence of SDA.4.This part included patients with LGIN diagnosed in 2007-2016;all participants were followed up until December 31,2021.Participants were divided into two groups:those who did not undergo surveillance endoscopy(non-surveillance group)and those who had undergone at least one surveillance endoscopy(surveillance group).Incidence of esophageal squamous cell carcinoma(ESCC),tumor stages,and treatments in ESCC cases were compared between these two groups.HRs and 95%CIs were calculated using Cox proportional hazards models.5.From the healthcare perspective,a Markov model was constructed to compare the cost-effectiveness of 30 endoscopy surveillance strategies(under different surveillance intervals for individuals with negative endoscopy findings,mild dysplasia,and moderate dysplasia)with no surveillance.The primary outcome was the incremental costeffectiveness ratio(ICER),incremental costs per quality-adjusted life-year(QALY)gained.The gross domestic product(GDP)per capita of China in 2021(12,551 U.S.$)was adopted as the threshold of willingness to pay(WTP).Univariate sensitivity analyses and probabilistic sensitivity analyses were performed to estimate the robustness.Results1.Among 120,759 eligible individuals from 13 screening centers in 2007-2016,51,495 were males(42.64%),and 69,264 were females(57.36%),with a mean age of 54.67 years.In endoscopic screening,115,785 individuals with negative endoscopy findings,3669 individuals with mild dysplasia,640 individuals with moderate dysplasia,423 individuals with SD/CIS,and 242 individuals with EC were diagnosed,respectively.The proportions of negative endoscopy findings,mild dysplasia,moderate dysplasia,SD/CIS,and EC gradually decreased,as they were 95.88%,3.04%,0.53%,0.35%,and 0.20%.Multivariate logistic analysis showed that individuals with male sex,older age,low socioeconomic status,cigarette smoking,disease history of the digestive system,and a family history of EC were associated with a higher detection rate of SDA.In addition,spearman correlation and linear regression analysis showed that population-based EC incidence of screening centers was positively correlated with the detection rate of SDA.2.A total of 40,977 individuals diagnosed with negative endoscopy findings,1562 individuals diagnosed with mild dysplasia,and 288 individuals diagnosed with moderate dysplasia were included in this part.The median follow-up time was 10.62 years(IQR:9.36-11.62),which accounted for a total of 453555.75 person-years.During the follow-up period,268 new EC cases were diagnosed,with an EC incidence of 59.09 per 100,000 person-years;128 EC deaths were observed,with EC mortality of 28.18 per 100,000 person-years.EC incidence and EC mortality gradually increased in the groups of negative endoscopy findings,mild dysplasia,and moderate dysplasia,as the incidence rates of 46.54 per 100,000 person-years,234.69 per 100,000 person-years,919.50 per 100,000 personyears,and mortality rates of 24.17 per 100,000 person-years,89.51 per 100,000 personyears,and 261.52 per 100,000 person-years.After adjusting for potential factors,the multivariate analysis showed a positive association between endoscopy findings and EC incidence and EC mortality.Compared with individuals with negative endoscopy findings,patients with mild dysplasia(HR=3.52,95%CI:2.47-5.00)and moderate dysplasia(HR=13.18,95%CI:8.60-20.17)had a higher risk of developing EC;patients with mild dysplasia(HR=2.43,95%CI:1.40-4.22)and moderate dysplasia(HR=6.46,95%CI:3.0913.51)also had a higher risk of EC deaths.Overall,compared with the general population in rural China,EC incidence and EC mortality were lower in those diagnosed with negative endoscopy findings over the follow-up period of 0-5 or 5.1-10 years.The SIRs in the period of 0-5 years and 5.1-10 years were 0.44(95%CI:0.32-0.60)and 0.74(95%CI:0.58-0.93),as the SMRs were 0.24(95%CI:0.15-0.36)and 0.68(95%CI:0.53-0.86),respectively.3.A total of 1183 patients with LGIN(976 with mild dysplasia and 207 with moderate dysplasia)were included in this part.The median follow-up time was 6.95 years(IQR:5.79-8.30),accounting for 8427.08 person-years.During the follow-up period,88 patients progressed to SDA(7.44%),with an incidence rate of 10.44 per 1000 person-years.The median interval from the progression of LGIN to SDA was 2.39 years(IQR:1.58-4.32).74.47%of patients with LGIN(n=881)experienced regression to non-dysplasia,and 18.09%(n=214)showed no lesion progression.Inpatients with moderate dysplasia,the cumulative incidence of SDA at five years was 15.47%,significantly higher than that(4.01%).in patients with mild dysplasia.In addition,patients with moderate dysplasia had a slightly shorter time(median interval of 1.86 years,IQR:1.27-2.87)to progress to SDA than those with mild dysplasia(median interval of 3.03 years,IQR:1.88-3.67).4.A total of 3258 patients with LGIN were included;1378 patients(42.30%)underwent at least one surveillance endoscopy(surveillance group),and 1880(57.70%)did not undergo any surveillance endoscopy(non-surveillance group).During the follow-up period(median:7.96 years),170 ESCC cases were diagnosed,with an incidence of 6.28 per 1000 person-years.A higher incidence of ESCC(7.07 per 1000 person-years)was observed in the non-surveillance group than in the surveillance group(5.14 per 1000 person-years).Among ESCC cases with the specific tumor stage and treatment information,the percentage of early stage was much higher in the surveillance group(80.77%)than that in the non-surveillance group(40.12%);patients with ESCC in the surveillance group more frequently underwent an endoscopic resection(endoscopic submucosal dissection or endoscopic mucosal resection)compared with those in the non-surveillance group(44.44%vs.14.29%).Patients in the surveillance group had a 31%decreased risk of ESCC incidence(HR=0.69,95%CI:0.50-0.95)compared with those in the non-surveillance group after adjusting for baseline risk factors.In addition,those undergoing surveillance endoscopy within the recommended interval had a 38%lower risk of developing ESCC(HR=0.62,95%CI:0.42-0.92).5.Compared with no endoscopy surveillance in individuals with negative endoscopy findings and LGIN,the ICERs of 30 surveillance strategies ranged from 1607-109,875 U.S.$ per QALY gained,respectively,for those aged 40 years at screening baseline;the ICERs of 26 surveillance strategies were less than the GDP per capita in 2021(12,551 U.S.$/QALY),indicating cost-effectiveness.For those aged 50 or 60 years at the screening baseline,the ICERs of 30 surveillance strategies were all less than the GDP per capita in 2021.Among them,15 surveillance strategies that only endoscopy surveillance for mild dysplasia and moderate dysplasia with surveillance interval of 1-5 years were cost-saving strategies.These strategies could increase QALYs and reduce costs,with ICER<0.At a WTP threshold of the GDP per capita,the optimal strategy for those with negative endoscopy findings,mild dysplasia,and moderate dysplasia was endoscopy surveillance every 10 years,5 years,and 3 years when aged 40-49 years,endoscopy surveillance every 10 years,5 years and 1 year when aged 50-59 years,and without endoscopy surveillance,every 4 years and 1 year when aged 60-69 years.Conclusions1.In endoscopy screening in rural populations,the prevalence of SDA was less than 1%,and most of them were diagnosed with negative endoscopy findings(95.9%)and LGIN(3.6%).Patients with LGIN had a higher risk of EC incidence and EC mortality than those with negative endoscopy findings.Patients with LGIN showed an overall risk of SDA of 1.04%per year;endoscopic surveillance was associated with decreased EC incidence in these patients.For individuals with negative endoscopy findings,the risk of EC incidence or EC mortality during a follow-up time of 10 years was lower than the general population in rural China.2.Surveillance for patients with LGIN and negative endoscopy findings after a single endoscopy screening is cost-effective in rural populations.The age-specific endoscopy surveillance strategy is recommended:for patients with negative endoscopy findings,surveillance every 10 years at the ages of 40-59;for patients with mild dysplasia,surveillance every 5 years at the ages of 40-59 and surveillance every 4 years at ages of 60-69;for patients with moderate dysplasia,surveillance every 3 years at ages of 40-49 and surveillance every 1 year at ages of 50-69.