Objective To improve the differential diagnosis efficiency of xanthogranulomatouscholecystitis(XGC)and gallbladder carcinoma(GBC),the differences between preoperative enhancedCT image features and clinical features in XGC and GBC with liver invasion were compared.Methods 39 cases of XGC and 55 cases of GBC with liver invasion confirmed by pathological examination,those patients were collected as research objects.All patients underwent dual phase enhanced CT scans before cholecystectomy.Logistic regression analysis was conducted to diagnose the independently correla-ted features of XGC,a modelformula for distinguishing XGC and GBC was established by Logit transformation calculus.The prediction efficiency of the model was evaluated by receiver operating characteristiccurve.Results Univariate analysis showed that there were statistically significant differences between XGC and GBC in gender,gallbladder wall thickeningpat-tern,CT enhancement method,the continuity of mucosal line,intramural nodules of gallbladder wall,sandwich biscuit sign,gallstones,peripheral enlargement lymph nodes(P<0.05).Multivariate analysis showed that diffuse thickening of the gall-bladder wall,continuous gallbladder mucosal line,intramural nodules of gallbladder wall and no peripheral enlargement lymph nodes were independently associated with XGC(all P<0.05).Based on the above imaging features independently related to XGC,a Logit transformation algorithm was used to establish a differentiation model formula between XGC and GBC.The AUC was 0.972(95%CI:0.936-1.000),the specificity was 0.909,and the sensitivity was 0.949.Conclu-sion Diffuse thickening of the gallbladderwall,continuous mucosal line,Intramural nodules of gallbladder wallandno Pe-ripheral enlargement lymph nodes are independent related factors for distinguishing XGC from GBC.The prediction model formula established by independent related factors can further improve the diagnostic accuracy of XGC with liver invasion,so as to avoid unnecessary extended resection.
NETL
NSTL
万方数据
