Introduction: Accurate delineation of targets and organs at risk (OAR) is required to ensure treatment efficacy and minimize risk of normal tissue toxicity with radiotherapy. Therefore, we evaluated the impacts of computed tomography (CT) slice thickness and reconstruction methods on the volume and dose evaluations of targets and OAR. 
Patients and methods: Eleven CT datasets from patients with thoracic cancer were included. 3D images with a slice thickness of 2 mm (2–CT) were created automatically. Images of other slice thickness (4–CT, 6–CT, 8–CT, 10–CT) were reconstructed manually by the selected 2D images using two methods; internal tumor information and external CT Reference markers. Structures and plans on 2–CT images, as a reference data, were copied to the reconstructed images. 
Results: The maximum error of volume was 84.6% for the smallest target in 10–CT, and the maximum error (≥20 cm³) was 10.1%, 14.8% for the two reconstruction methods, internal tumor information and external CT Reference, respectively. Changes in conformity index for a target of <20 cm³ were 5.4% and 17.5% in 8–CT. Changes on V₃₀ and V₄₀ of the heart were considerable. In the internal tumor information method, volumes of hearts decreased by 3.2% in 6–CT, while V₃₀ and V₄₀ increased by 18.4% and 46.6%.
Conclusion: The image reconstruction method by internal tumor information was less affected by slice thickness than the image reconstruction method by external CT Reference markers. This study suggested that before positioning scanning, the largest section through the target should be determined and the optimal slice thickness should be estimated.
Keywords: computed tomography, slice thickness, thoracic cancer, dose, image reconstruction