| Objective:Esophageal fistula(EF)is a disastrous complication of radiation therapy in patient s with esophageal carcinoma(EC).Its pathogenesis is complex,and several factors are involved.This study aims to establish a clinical prediction model for predicting the risk of radiotherapy-related esophageal fistula in esophageal cancer patients by combining demographic characteristics,tumor characteristics,treatment-related data,and hematol ogical data,to screen out high-risk groups of esophageal fistula and provide recommen dations for clinical decision-making.Methods:A retrospective nested case-control study was performed,in which EC patients rec eiving radiotherapy from 2014 to 2020 in the First Affiliated Hospital of Anhui Medica l University were enrolled.All clinical data were retrospectively collected from electro nic medical records.Patients with EF were screened according to inclusion and exclusi on criteria,and the cases and controls were matched by the age,gender,and diagnosis t ime at a ratio of 1:3.According to independent risk factors screened by univariate and multivariate logistic regression analysis,a nomogram for predicting the risk of EF in es ophageal cancer patients receiving radiotherapy was constructed.The following metho ds including receiver operating characteristic(ROC)curve,C-index,calibration curves,Brier score,and Decision Curve Analysis(DCA)were adopted to evaluate this nomog ram.Results:1.During the study period,1183 cases who had undergone radical radiotherapy we re identified in our database,and 81 had developed EF.After matching by the age,gend er,and diagnosis time,a total of 324 EC patients,including those 81 EF cases and 243 c ontrols,were enrolled for the subsequent analyses.Of these patients who developed EF,12 patients(14.8%)suffered perforation during RT,while 69 patients(85.2%)developed t his complication after RT.The median intervals between the end of radiotherapy and the EF onset were 4.6 months.Among all the 81 cases with EF,46 cases developed esopha gomediastinal fistula,28 cases developed esophagotracheal fistula,2 cases developed es ophago-arterio fistulas and 5 cases suffered both esophagomediastinal and esophagotrac heal fistula.Managements of fistula included nutrient canal in 61 patients(75.3%),esop hageal stent in 16 patients(19.7%)and parenteral nutrition in 4 patients(4.9%).2.Univariate analysis revealed that body mass index(BMI)<20 kg/m2(P=0.023),T stage(P<0.001),N stage(P=0.049),NLR(neutrophil to lymphocyteratio,NLR)(P=0.012),hemoglobin(P=0.018),prealbumin(P=0.003),re-radiotherapy(P=0.014),ulcerative E C(P<0.001),stenosis before radiotherapy(P<0.001),length of gross tumor volume(GT V)(P<0.001),maximum diameter of GTV(P<0.001)were significantly correlated with t he occurrence of EF.The other clinical parameters including age,albumin,tumor locati on,total dose,single dose,GTV volume,chemotherapy,treatment response,smoking hi story,diabetes history,and hypertension history were not significant for their associatio n with EF.Multivariate analysis showed that T stage,stenosis before radiotherapy,ulcer ative EC,prealbumin,maximum diameter of GTV,and NLR remained significant,whic h indicated these clinical characteristics were independent risk factors for the occurrenc e of EF.3.A nomogram incorporating the aforementioned six parameters was constructed b ased on the results of the multivariate logistic regression analysis.And a dynamic nomo gram was established,which can dynamically predict EF on the website.The area under the ROC curve(AUC)was 0.848(95%CI 0.901-0.895),and further evaluation of this model showed that the C-index was 0.847,while the bias-corrected C-index after intern al validation was 0.833,indicating a high prediction accuracy of this nomogram.The Br ier score was 0.127.The calibration curves presented well concordance and the DCA re vealed promising clinical application.Conclusions:We determined T stage,stenosis before radiotherapy,ulcerative EC,prealbumin,m aximum diameter of GTV,and NLR as the independent risk factors of EF based on a sin gle-center retrospective nested case-control study.A nomogram was accordingly constru cted and visualized to facilitate the prediction of EF risk related to radiotherapy.This cal culable tool is promisingly applied in clinical practice to participate in determining indi vidual therapeutic strategies for EC patients. |