Investigating the impact of non-standard positioning on the accuracy of skull tracking algorithms using dual-panel imaging systems.

OBJECTIVE

This study investigates the impact of non-standard positioning on the accuracy of 6D-skull tracking using dual-panel imaging systems. It explores whether positioning patients' heads at various angles during intracranial lesion treatment affects the accuracy of the CyberKnife 6D-skull tracking system.

MATERIALS AND METHODS

A heterogeneous density skull phantom was used to simulate various patient skull positioning angles. To accurately compare 6D-skull tracking and fiducial tracking, their center coordinates were pre-set to be identical in the treatment plan. The phantom was positioned using fiducial tracking, and the offset value recorded. The system was then switched to 6D-skull tracking to observe the corresponding offset. The difference between the two tracking methods was calculated, and a paired-sample T-test was conducted to assess statistical significance across different angles. Additionally, the gamma passing rate (criteria: 3%/3mm) was employed to quantitatively delineate dosimetric disparities attributable to positional variations.

RESULTS

Paired sample T-tests on the deviations between rotational and translational parameters of fiducial tracking and skull tracking under identical conditions revealed no statistically significant differences between the methods across all selected angles. The minimal deviations and lack of statistical significance demonstrate that both tracking methods are equivalent in skull positioning. Furthermore, the gamma passing rate analysis showed that in all tested conditions, the rates exceeded 95%, which aligns with clinical requirements. This high passing rate indicates a high degree of dosimetric accuracy and consistency between the two tracking methods, providing robust assurance of treatment precision in skull positioning.

CONCLUSION

Since fiducial tracking is not affected by patient or phantom positioning, this study compares the registration results of 6D-skull tracking with fiducial tracking under the same conditions. The results show minimal deviations and no statistically significant differences, indicating that 6D-skull tracking is not dependent on the skull's positioning angle. Furthermore, the gamma passing rate analysis was conducted to quantitatively assess the dosimetric differences arising from variations in patient positioning. Our results demonstrated that under all tested conditions, the gamma passing rates exceeded the clinically accepted threshold of 95%, confirming the clinical adequacy of both tracking methods in maintaining treatment precision. In clinical practice, patients do not need to maintain a strict supine position; the algorithm can accurately perform registration even if patients need to rotate their heads or lie prone. Clinical recommendations should prioritize patient comfort and safety without imposing overly strict requirements.