Pet-CT Clinical Study On Radiotherapy Target Volume Delineation Of Non-small Cell Lung Cancer

Objective:Analyzing the data of newly diagnosed non-small cell lung cancer(NSCLC)patients,and comparing the tumor size on 18 FFDG PET/CT with the size on enhanced CT,Find the optimal delineation method.Methods:From May 2017 to December 2017,19 patients diagnosed with NSCLC by pathology or imaging in Hunan tumor hospital were recruited.According to the 8th AJCC guidelines,all patients underwent whole body or chest 18F-FDG PET-CT and chest enhanced CT examination,then automatically delineate the tumor target area on 18F-FDG PET-CT with different SUV cutoff values,and find the largest cross-sectional area of the imaging,using delineation methods of the different percentage of SUVmax in this level,eyeballing with PET,CT mediastinum window,CT lung window,a total of 15 methods.Surgical lung resection or lung cancer resection was operated.The tumor tissue corresponding to the largest area of imaging was obtained according to the imaging thickness.The tissue was fixed in formalin,then the longest path(Lpath)and area of the measured tissue(Apath)were measured.,and observe the infiltration of the microscope.Using Lpath and Apath as the gold criterion,the reference method for delineating the best target area was preliminarily confirmed.Result:The best reference for SUV cutoff was 3.5.The ICC consistency analysis showed that in the Lpath comparison,the best SUV cutoff was SUV3.5,which is in the highest agreement(ICC=0.880,95% CI: 0.720 to 0.952,P<0.001);also in the Apath comparison,the SUV cutoff value of 3.5 had the highest agreement(ICC=0.773,95% CI: 0.474 to 0.914,P<0.001).Combined with Bland-Altman analysis,in the Lpath comparison,when the SUV was taken as 3.5,the consistency limit was-1.764 to 2.147.And only 1 case exceeded the limit.In the Apath comparison,when the SUV was taken as 3.5,the consistency limit was-9.762 to-11.470,and one case out of the limit,which had a good consistency.The maximum value of microscopic infiltration was 5 mm.There was an independent sample t-test with different SUVmax values under microscopic infiltration and infiltration without microscopy,pvalue was 0.327,no statistical significance.Conclusion:This study primarily has discussed the best delineation method of GTV for NSCLC,using the pathological specimen as the gold criterion.The best reference value has shown that the optimal SUV cutoff value is 3.5.There is no intrinsic link between the SUVmax value and the microscopic infiltration.According to the infiltration data,clinical target volume should be less than 5mm in maximum.