ObjectiveThe aim of this study is to analyze which tracking modality is more suitable for stereotactic body radiosurgery of lumbosacral spinal tumors by comparing prone and supine patient treatment setup.MethodsEighteen patients with lumbosacral spinal tumors were selected. CT simulation was performed in the supine position (fixed with a vacuum cushion) and prone position (fixed with a thermoplastic mask and prone plate), respectively. The plans in the supine and prone positions were designed using the xsight spine tracking (XST) and xsight spine prone tracking (XSPT) modalities, respectively. The dose-volume histogram (DVH) parameters, namely, V100%, D95%, Dmean, conformity index (CI), and heterogeneity index (HI) in planning target volume (PTV), as well as Dmax, D0.1cc, D1cc, and D5cc in the cauda equina and bowel were recorded. The supine plans were simulation plans and were not used for treatment, which were only used to record the alignment errors. The spinal tracking correction errors (alignment error) and correlation errors of the synchrony respiratory model in the prone position were recorded during the treatment. After treatment, the simulation plan of the supine position was implemented and the spinal tracking correction errors were recorded. The parameters of correction error and DVH parameters for the two positions were analyzed using the paired t-test to compare the difference in positioning accuracy and dose distribution. In addition, the correlation errors of the synchrony respiratory model in the prone position were analyzed to evaluate the prediction accuracy of the synchrony model.ResultsFor patient setup, the correction error of the supine position in interior/posterior was (0.18 ± 0.16) mm and the prone position was (0.31 ± 0.26) mm (P< 0.05). The correction error of the supine position in inferior/superior was (0.27 ± 0.24) mm, and the prone position was (0.5 ± 0.4) mm (P< 0.05). The average correlation errors of the synchrony model for left/right, inferior/superior, and anterior/posterior in the prone position were (0.21 ± 0.11) mm, (0.41 ± 0.38) mm, and (0.68 ± 0.42) mm, respectively. For the dose distribution, compared with prone plans, the average CI in supine plans was increased by 4.5% (P< 0.05). There was no significant difference in HI, PTV V100%, D95%, and Dmean between the prone and supine plans. Compared with supine plans, average D1cc and D5cc for the cauda equina was significantly decreased by 4.7 and 15.3% in the prone plan (P< 0.05). For the bowel, average Dmax, D0.1cc, D1cc, and D5cc were reduced by 8.0, 7.7, 5.2, and 26.6% in prone plans (P< 0.05) compared with supine plans.ConclusionCompared with the supine setup, the prone setup combined with XSPT modality for the lumbosacral spinal stereotactic body radiosurgery can spare the bowel and cauda equina of the middle and low dose irradiation, and decrease the number of beams and monitor units.