Background: Computed tomography (CT)/positron emission tomography (PET) may decide on nodal staging so as to improve radiotherapy arrangement for patients of non-small cell lung cancer (NSCLC). However, if PET/CT can reduce the variability of interobserver in the target amount delineation remains to be answered. The present study assesses how accurate PET/CT is in identification of nodal metastasis as well as nodal staging and its impact on the radiotherapy amount delineation to patients of NSCLC. Methods: In total, 27 patients with NSCLC were recruited and scanned with PET/CT. Sets of PETS as well as CT data could be delivered to the treatment planning system (TPS), followed by the fusing of CT and PET images using TPS. Two radiation oncologists on CT and PET/CT images were employed to contour gross tumor volume (GTV). The contours were outlined together with two radiation oncologists, thus producing a final consensus. Results: Tumor-node-metastasis categories changed on 8/27 patients (30%) on the basis of PET/CT. Radiation on PET and CT integration images led to changes in treatment arrangements on 12/20 patients (60%). In addition, PET/CT produced significant accuracy for nodal staging than CT. The negative and positive predictive values, sensitivity, specificity, and accuracy of PET/CT to detect nodal metastasis were 74.3%, 89.3%, 51.5%, 95.8%, and 87.3% and the corresponding CT data were respectively 45.5%, 87.1%, 45.5%, 87.1%, and 79.2%. PET/CT confers significantly higher specificity, positive predictive value as well as accuracy of CT in detecting nodal metastasis. Conclusions: PET/CT fusion images may influence treatment arrangement. PET/CT is much more precise than CT in nodal staging, and is more detailed and precise than CT in finding nodal metastasis but is less sensitive with a higher false-negative rate.