Performance of thoracic radiotherapy may be assisted by the use of thoracoabdominal flat immobilization devices (TAFIDs) and integrated cervicothoracic immobilization devices (ICTIDs). This study was performed to compare setup errors of TAFIDs and ICTIDs. Forty-four patients with lung cancer were retrospectively reviewed; 22 patients were immobilized with a TAFID and 22 with an ICTID. In total, 343 cone-beam computed tomography images of these patients were collected for radiotherapy setup. The 3-dimensional setup errors and the displacement of the acromioclavicular joint against the supraclavicular region were calculated. An independent-samples t-test and rank-sum test were used for statistical analyses. The translational setup errors of the TAFID group vs ICTID group in the left–right (LR), superior–inferior (SI), and anterior–posterior (AP) directions were 0.14 ± 0.17 vs 0.14 ± 0.16 cm (p = 0.364), 0.23 ± 0.26 vs 0.15 ± 0.15 cm (p = 0.000), and 0.16 ± 0.15 vs 0.12 ± 0.14 cm (p = 0.049), respectively. The relative displacement of the acromioclavicular joint against the supraclavicular joint in the LR, SI, and AP directions were 0.10 ± 0.12 vs 0.09 ± 0.10 cm (p = 0.176), 0.13 ± 0.13 vs 0.11 ± 0.12 cm (p = 0.083), and 0.17 ± 0.16 vs 0.12 ± 0.11 cm (p = 0.001), respectively. The overall displacement of the supraclavicular region was 0.28 ± 0.19 vs 0.23 ± 0.15 cm (p < 0.001). The recommended planning target volume margins in the LR, SI, and AP directions were 0.46 vs 0.74 cm, 0.51 vs 0.47 cm, and 0.49 vs 0.41 cm, respectively. For patients with lung cancer, using an ICTID can reduce setup errors in the SI direction and displacements of the acromioclavicular joint and supraclavicular region compared with a TAFID. Therefore, an ICTID is preferred for patients with lung cancer with supraclavicular target volume.