Potentials and limitations of guiding liver stereotactic body radiation therapy set-up on liver-implanted fiducial markers

Purpose: We investigated the potentials and limitations of guiding liver stereotactic body radiation therapy (SBRT) set-up on liver-implanted fiducial markers. Methods and Materials: Twelve patients undergoing compression-supported SBRT in a stereotactic body frame received fluoroscopy at treatment preparation and before each treatment fraction. In fluoroscopic videos we localized the markers and diaphragm tip at expiration and the spine (measurements on free-breathing and abdominal compression). Day-to-day displacements, rotations (markers only), and deformations were determined. Marker guidance was compared to conventional set-up strategies in treatment set-up simulations. Results: For compression, day-to-day motion of markers with respect to their centers of mass (COM) was σ = 0.9 mm (random error SD), ∑ = 0.4 mm (systematic error SD), and <2.1 mm (maximum). Consequently, assuming that markers were closely surrounding spherical tumors, marker COM-guided set-up would have required safety margins of ∼2 mm. Using marker COM as the gold standard, other set-up methods (using no correction, spine registration, and diaphragm tip craniocaudal registration) resulted in set-up errors of 1.4 mm < σ < 2.8 mm, 2.6 mm < ∑ < 5.1 mm, and 6.3 mm < max < 12.4 mm. Day-to-day intermarker motion of <16.7%, 2.2% median, and rotations between 3.5° and 7.2° were observed. For markers not surrounding the tumor, e.g., 5 cm between respective COMs, these changes could effect residual tumor set-up errors up to 8.4 mm, 1.1 mm median (deformations), and 3.1 mm to 6.3 mm (rotations). Compression did not systematically contribute to deformations and rotations, since similar results were observed for free-breathing. Conclusions: If markers can be implanted near and around the tumor, residual set-up errors by marker guidance are small compared to those of conventional set-up methods, allowing high-precision tumor radiation set-up. However, substantial errors may result if markers are not implanted precisely, requiring further research to obtain adequate safety margins.