Dosimetric Impact of Angular Deviations in Positioning for Spinal Radiosurgery

Abstract

During image-guided radiosurgery, it is common to account only for translational shifts while leaving rotational deviations uncorrected. The aim of this study is to investigate the rotational setup deviations and estimate the dosimetric impact in spinal radiosurgery. Stereoscopic image registrations of 711 spine radiosurgery procedures (546 patients) were retrospectively reviewed and classified to cervical (C), thoracic (T), and lumbar (L) spines. The probability of rotational deviations were estimated in three orthogonal axis. In order to investigate the dosimetric impact, the original simulation CT images were rotated in the range of ±5 deg in each axis. Then, the dosimetric changes were recorded; including the coverage (V90) and minimum dose (Dmin) to the target, and the dose of 10% volume (D10) and maximum dose (Dmax) to the spinal cord. Target rotations of >5 deg and >3 deg occurred respectively in 1% and 8% of cases. The mean rotations were small (|μ|<0.6 deg) with no axis and site dependent trends. The standard deviations (σ) of different sites and axes ranged from 1.4 to 2.0 deg (mean σ=1.7 deg). The overall dosimetric changes of the target coverage(V90) and cord D10 were small. Five degree rotation resulted in a target V90 decrease of 1±1%, 2±2% and 3±1%, respectively in the lateral, vertical, and longitudinal axes. The corresponding cord D10 increased by 3±6%, 2±2%, and 6±5%. The changes in cord Dmax ranged from 92% to 138% of the original reference spinal cord. Corresponding changes in target Dmin were 59~106%. Large target rotations occurred in a small percentage of patients. The overall dosimetric impact on target coverage and cord D10 was small. However larger effects were observed on the target Dmin and cord Dmax. Special care must be taken for targets of elongated shapes involving multiple vertebrae.