Objective To compare the setup errors of breast cancer patients who received postmastectomy radiotherapy (PMRT) with head/neck and shoulder thermoplastic mask and integral cervicothoracic mask.

Methods A retrospective analysis was performed on 40 female breast cancer patients (median age 46 (29–68) years) who treated with PMRT in Cancer Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College from June to December 2019. The supra/infraclavicular nodal region was irradiated with intensity-modulated radiotherapy technique and chest wall ± internal mammary chain were irradiated with electrons. A total of 271 cone beam CT (CBCT) images of the supra/infraclavicular nodal region were analyzed retrospectively. One hundred and twenty eight CBCT images were from 19 patients immobilized with head/neck and shoulder thermoplastic mask, and 143 CBCT images were from 21 patients immobilized with integral cervicothoracic thermoplastic mask. The setup errors of the two groups were analyzed, and the movement amplitude ΔX, ΔY, ΔZ and three-dimensional displacement d of acromioclavicular joint were measured. According to van Herk's formula, M_PTV=2.5∑+0.7σ, the margin from clinical target volume to planning target volume (PTV) is calculated. According to whether body mass index (BMI)≥24 kg/m², the patients were divided into two groups to analyze whether there was statistical difference of the setup errors. The differences were compared by independent sample t test.

Results The translational setup errors of the supra/infraclavicular nodal region for head/neck and shoulder and integral cervicothoracic thermoplastic mask groups in the left–right, superior–inferior, and anterior–posterior directions were as follows: (1.78±2.40) mm and (2.26±1.91) mm (t=0.687, P=0.496); (2.88±2.44) mm and 2.29±1.89 mm (t=2.249, P=0.030); (2.48±1.79) mm and (1.66±2.00) mm (t=0.998, P=0.325), respectively; the rotational setup errors in the sagittal directions were (0.91±0.61)° and (0.67±0.53)°, and the difference was statistically significant (t=3.555, P=0.001). The arithmetic mean of movement amplitude of the acromioclavicular joint ( ΔX, ΔY, and ΔZ ) were 3.65, 4.61, 3.11 mm and 2.32, 1.97, 1.63 mm in the X, Y, and Z directions, respectively. The arithmetic mean of three-dimensional displacement d were 7.36 mm and 4.03 mm, respectively. The calculated PTV margin in the left–right, superior–inferior, and anterior–posterior directions with head/neck and shoulder and integral cervicothoracic mask groups were 6.1, 8.9, and 7.5 mm and 7.0, 7.0, and 5.5 mm, respectively. No significant difference was found between the BMI<24 and BMI≥24 kg/m² groups in translation and rotation setup errors in the three directions (t=−1.103 to 0.938, all P>0.05).

Conclusions The integral cervicothoracic mask system for patients who received PMRT in the supra/infraclavicular nodal region after the modified radical mastectomy provides smaller setup errors than the head and neck–shoulder thermoplastic mask system in the superior–inferior direction. This system also has remarkable body position fixation effects on the acromioclavicular region.