The Diagnositic Value Of¹⁸F-FDG PET/CT Imaging In Epatocellular Carcinoma

Objective To investigate the diagnostic effectiveness and diagnosticcriteria of18F-FDG PET/CT imaging in the diagnosis of hepatocellularcarcinoma.Methods Thirty-four patients who have been confirmed to behepatocellular carcinoma by pathology in our hospital were retrospectivelyanalyzed by their18F-FDG PET/CT imaging findings since September2008to March2014. We set detection index by using qualitative and quantitativeparameters：standard uptake value（SUV），choose the maximum standardizeduptake value SUVmax、the ratio of target and non target area. These twoindicators were obtained from region of interest sketch which wasautomatically got from computer software. At the same time，we selectedfifty healthy subjects with normal blood glucose and without any liver relateddiseases to undergo18F-FDG PET/CT imaging，their results were used asnormal control group. SPSS16.0statistical software was applied for statisticalProcessing. We obtained the value of SUVmax, T/NT mean±standarddeviation from the study group and the control group. Took a two independent samples t test comparison between T/NT and SUVmax values, and usedstatistical software to analyze whether there is a difference between the twogroups. The best diagnostic cut-off points from the values of SUVmax and T/NT were obtained by using receiver operating characteristic (receiveroperating characteristic, ROC) in the diagnosis of HCC. The area under theROC curve (area under the curve, AUC) was evaluated between the twomethods for the accuracy of HCC diagnosis. And then we selectSUVmax≥2.5as a standard value (due to high SUVmax value has significancein diagnosis of HCC) to calculate its diagnostic sensitivity and specificity.Results1、18F-FDG PET/CT imaging shows: HCC group,34patientsshowed varying degrees of FDG uptake in liver.And25patients with varyingdegrees of fatty liver, cirrhosis and liquefaction necrosis at the edge；12patients with portal vein thrombosis、 intrahepatic and distant metastasisformation. The control group of50healthy subjects had no history ofliver-related diseases,18F-FDG PET/CT imaging showed no liver lesions andabnormal FDG uptake.50healthy subjects in the control group showed noobvious high18F-FDG uptake and liver-related diseases. SUVmax values inHCC group and control group were4.69±1.77and2.91±0.63respectively.The difference of SUVmax values between the two groups was statisticallysignificant (P=0.00). T/NT values in HCC group and the control group were1.88±0.76and1.08±0.08respectively，The difference of T/NT valuesbetween the two groups was statistically significant (P=0.00).2、The best diagnostic cut-off point: T/NT best diagnostic cut-off pointdetermined by ROC curve was1.25(Youden index=0.85), SUVmax bestdiagnostic cut-off point was3.75(Youden index=0.59). When T/NT≥1.25was selected as a positive diagnostic criterion，its sensitivity was91.20%, specificity of94%and accuracy of92.86%. When SUVmax≥1.25wasselected as a positive diagnostic criterion, its sensitivity was70.60%,specificity of88%and accuracy of80.95%.3、Comparison of the two methods: Area of T/NT value method under theROC curve was0.98, greater than values of SUVmax which was0.86. Theaccuracy of the T/NT value method was92.86%, higher than values ofSUVmax which was80.95%.4、When we set SUVmax≥2.5which is the boundary of benign andmalignant tumors as the diagnostic criteria, the sensitivity was91.20%(31/34),specificity of24%，accuracy rate of51.19%, positive predictive value of44.92%and negative predictive value of80%and the approximately Youdenindex was0.15.Conclusions1、 In this study, the diagnosis of HCC derives fromsemi-quantitative analysis of indicators of SUVmax and T/NT，theirmean±standard deviation are4.69±1.77,1.88±0.76respectively，which canprovide a reference for subsequent18F-FDG PET/CT imaging in the diagnosisof HCC.2、Semi-quantitative analysis of18F-FDG PET/CT imaging can helpdiagnose patients occupying lesions in the liver, the best diagnostic points of T/NT and SUVmax are1.25and3.75respectively. The sensitivity，specificityand accuracy of the former diagnostic point of HCC are91.20%，94%and92.86%respectively while the sensitivity，specificity and accuracy of the latterone were70.60%,88%and80.95%respectively.3、T/NT ratio method has a higher diagnostic value for HCC than SUVmaxmethod, the former is stronger than the latter in the ability to identify patientswith HCC and has a higher accuracy than the latter. Therefore, T/NT ratio method is more accurate in the diagnosis of HCC and can provide areference for subsequent clinical cases, but when we select different non-targettissue site it will cause a greater difference of T/NT values which affects itsstability.