论文部分内容阅读
Purpose:PET visualization method is routinely used in clinical practice,however,it has high interobserver variations and relies on the experience of nuclear medicine physician.PET automated threshold delineation methods have been proposed.This study aimed to compare Gross tumor volumes(GTVs)defined by Computed tomography(CT)based image,automated and visual assessment of 18F-FDG Positron emission tomography(PET)images.Methods:Eight non-small-cell lung cancer(NSCLC)patients who underwent PET/CT imaging were enrolled into this study.CT-based GTV definition was performed by the radiation oncologist.The automated PET target volume delineation methods(Standardized uptake value(SUV)> 2.5(PET2.5)and PET 40%SUVmax(PET40)defined by RT image Suite,Siemens healthcare)and PET visual assessment(PETvisual)method were performed by nuclear medicine physician.The GTVs were calculated by Oncentra v.4.3 treatment planning software and the percentage difference of each GTV was analyzed and compared to that of PET visual assessment method.Results:The mean SUVmax for all patients was 7.6(4.18-14.7).Mean GTV volumes were 16.06,14.99,6.05,16.75 cc for CT-based,PET2.5,PET40 and PETvisual,respectively.As compare to PETvisual,the average volume difference were 21.02%(PET2.5),22.71%(PET40)and 2.84%(CT-based).PET target volume deviation by automated methods depended on SUVmax.PET40 demonstrated large target volume deviation at higher SUVmax,contrary to PET2.5,the optimal cutpoint SUVmax threshold was at 7.Conclusions:SUVmax impacted on PET target volume delineation in NSCLC patients.Our preliminary study suggested using 40%SUVmax method when SUVmax < 7 and using SUV>2.5 method when SUVmax ≥ 7.