Clinical Implementation of Multi-Target Winston-Lutz Using the QUASARTM Penta-Guide Phantom
Abstract
Purpose
Winston-Lutz (WL) testing is routinely used to assess coincidence between a linac’s mechanical and radiation isocenter. Although dedicated commercial phantoms exist for off-axis target verification, they require additional resources and training. The QUASARTM Penta-Guide (Modus Medical Devices, Canada) is a widely used phantom containing hollow spheres at isocenter and several off-axis positions. We developed and clinically validated a multi-target WL (MTWL) workflow using the Penta-Guide.
Methods
Standard Penta-Guide setup for single-target WL (STWL) was adopted, with MLC fields and analysis modified for multi-target assessment. With the phantom positioned at isocenter, nine EPID images were acquired using a 10FFF beam at multiple gantry, collimator, and couch angles. Two 3×3cm² MLC field-in-field apertures were delivered per image to localize both spheres: one at isocenter, and one off-axis with an offset of 2, 4, and 6cm in the LEFT, UP, and IN directions (radial distance of 7.48cm) relative to isocenter. Pylinac was modified to analyze low-contrast spheres using a Hough transform. The off-axis target was analyzed with additional processing due to beam unflatness and phantom edges. Localization accuracy was evaluated using controlled couch shifts across multiple Varian Edge linacs. Workflow time was also assessed.
Results
Controlled-shift testing demonstrated isocenter localization accuracy of 0.05±0.06mm, 0.09±0.10mm, and 0.05±0.06mm, and off-axis accuracy of 0.36±0.25mm, 0.21±0.27mm, and 0.18±0.20mm for LEFT, UP, and IN shifts, respectively. Across 21 routine QA sessions, maximum 2D error was 0.63±0.10 mm at isocenter and 0.81±0.14mm off-axis. Setup time was unchanged, while delivery time increased by 243±4s due to additional collimator and MLC motion. Mean analysis time was 1.38±0.10s for STWL and 9.16±0.31s for MTWL.
Conclusion
A practical MTWL workflow was implemented using a clinically ubiquitous phantom. Localization accuracy for both targets was within 1mm, with minimal additional workflow burden, enabling clinical adoption without purchasing or constructing an additional MTWL phantom.