The Standard Uptake Value of 18F-FDG PET/CT Imaging in the Application of Diagnosis of Gastric Cancer
|Course||Medical Imaging and Nuclear Medicine|
|Keywords||18F-FDG PET/CT SUV Gastric Cancer ROC|
Objective:To explore the value of standardized uptake value of 18F-FDG PET/CT in diagnosising gastric cancer.Methods:The 18F-FDG PET/CT imaging and pathologic data (endoscopic biopsy and /or surgery) of 86 cases of gastric lesions were retrospectively analyzed. With PET SUVmax (=2.5) for the quantitative diagnostic criteria, the diagnostic value of PET was analysed. Then ROC curves were used to assess six quantitative indicators (SUVmax, SUVave, four TLR) of PET. which were used to differentially diagnosis gastric cancer and benign lesions (inflammatory lesions),and the best cut-off value was selected.Result:(1) Among the 86 cases,61 cases of gastric cancer patients were found, and the other 25 cases were gastritis. (2)When the SUVmax cut-off value was selected as 2.5 for the 86 cases, the sensitivity of diagnosis of gastric cancer was higher (90.16%) and the specificity and accuracy were both poor (32.00% and Youden index=0.22. respectively). (3)The difference of the SUVmax values between the gastric cancer group and the gastritis group was significantly (U=262.00, P< 0.0001). ROC analysis showed that the area under the curve of SUVmax was largest among six quantitative indicators (SUVmax, SUVave. four TLR) (AUC=0.828.z=7.474,P<0.0001). therefore, SUVmax was chosen as the best quantitative indicator of PET. The up-left cut-off value of the ROC was 6.05, the sensitivity and specificity were 63.93% and 100.00%, respectively. To take into account the sensitivity and specificity, we chose 4.56 as the cut-off value of SUVmax, the sensitivity and specificity were 70.49% and 84.00%, respectively. (4)When 4.5 and 6 were selected as the SUVmax cut-off point, using consistency test, the Kappa values were 0.453 and 0.438. respectively, this means that little difference was found between the both, they were moderately consisted with pathological findings. (5)When gastric wall thickening (above 5mm)and/or mass were selected as the criteria for differential diagnosis of gastric cancer and gastritis, the Kappa value were 0.616. and the sensitivity was higher (98.36%),while the specificity was poor (56.00%) and moderate accuracy (Youden index=0.54) was got. (6)We conducted a joint sequence diagnosis for gastric cancer, which combined the CT scan criteria(gastric wall thickening> 5mm) with the quantitative indicator of PET(SUVmax).When 4.5 and 6 were selected as the SUVmax cut-off point, the diagnostic accuracy of the former (Youden index=0.62) was higher than the latter (Youden index=0.56). the sensitivity and specificity of the former were 70.49% and 92.00%, respectively.Conclusions:(1)ROC analysis showed that the area under the curve of SUVmax was the largest among six quantitative indicators (SUVmax, SUVave, four TLR) (AUC=0.828, z=7.474.P<0.0001). therefore, the best indicator of quantitative analysis is SUVmax of gastric lesion. (2)When 4.5 and 6 were selected as the SUVmax cut-off point, although the sensitivity and specificity of the two differential diagnoses of gastric cancer and gastritis vary, the accuracy of the both has little difference(Youden index=0.56,0.50,respectively). They are moderately consisted with pathological findings(the Kappa values were 0.453 and 0.438,respectively.). (3)When combined the CT scan criteria(gastric wall thickening> 5mm) with the quantitative indicator of PET(SUVmax=4.5), PET/CT in the differential diagnosis of gastric cancer and gastritis achieve a better performance.