Abstract: Objective To compare the setup errors provided by the chest and abdomen flat frame fixation device and integral cervico-thoracic fixation device in supraclavicular regions of locally advanced esophageal cancer patients. Methods The cone beam computed tomography (CBCT) images of 50 patients with locally advanced esophageal cancer and received radiotherapy in supraclavicular region were retrospectively analyzed. The chest and abdomen flat frame fixation device (group A) or the integral cervicothoracic fixation device (group B) were used respectively with 25 in each.There were 175 CBCT images in group A and 230 CBCT images in group B. The setup errors of the two groups were analyzed, and the movement and displacement of acromioclavicular joint in three-dimensions were measured. The differences were compared by independent sample t test and rank sum test. Results The setup errors were significantly different between group A and group B. The P value in X axis (left-right) and Z axis (anterior-posterior) were both less than 0.001 (P＜0.001);The P value in pitch was less than 0.001 (P＜0.001)and roll was less than 0.05 (P＜0.05). The movement amplitude of acromioclavicular joint (ΔX, ΔY, ΔZ) of groups A and B were (0.10±0.10)cm and (0.08±0.08)cm in X axis, (0.14±0.13)cm and (0.11±0.11)cm in Y axis, (0.16±0.12)cm and (0.11±0.12)cm in Z axis. The three-dimensional displacement of groups A and B were (0.27±0.14)cm and (0.20±0.14)cm, (P＜0.001). Conclusions If the influence of rotation angle is well controlled, the application of the integral cervicothoracic fixation device is better to immobilize the acromioclavicular joint and reproduce the setup position for locally advanced esophageal cancer patients. Special attention should be paid for the influence of rotation angle on the location accuracy of target.

Key words: locally advanced esophageal cancer, acromioclavicular joint mobility, setup error analysis, radiotherapy