The accuracy and precision of the KIM motion monitoring system used in the multi-institutional TROG 15.01 Stereotactic Prostate Ablative Radiotherapy with KIM (SPARK) trial.

Purpose Kilovoltage intrafraction monitoring (KIM) allows for real-time image guidance for tracking tumor motion in six-degrees-of-freedom (6DoF) on a standard linear accelerator. This study assessed the geometric accuracy and precision of KIM used to guide patient treatments in the TROG 15.01 multi-institutional Stereotactic Prostate Ablative Radiotherapy with KIM trial and investigated factors affecting accuracy and precision. Methods Fractions from 44 patients with prostate cancer treated using KIM-guided SBRT were analyzed across four institutions, on two different linear accelerator models and two different beam models (6 MV and 10 MV FFF). The geometric accuracy and precision of KIM was assessed from over 33 000 images (translation) and over 9000 images (rotation) by comparing the real-time measured motion to retrospective kV/MV triangulation. Factors potentially affecting accuracy, including contrast-to-noise ratio (CNR) of kV images and incorrect marker segmentation, were also investigated. Results The geometric accuracy and precision did not depend on treatment institution, beam model or motion magnitude, but was correlated with gantry angle. The centroid geometric accuracy and precision of the KIM system for SABR prostate treatments was 0.0 +/- 0.5, 0.0 +/- 0.4 and 0.1 +/- 0.3 mm for translation, and -0.1 +/- 0.6 degrees, -0.1 +/- 1.4 degrees and -0.1 +/- 1.0 degrees for rotation in the AP, LR and SI directions respectively. Centroid geometric error exceeded 2 mm for 0.05% of this dataset. No significant relationship was found between large geometric error and CNR or marker segmentation correlation. Conclusions This study demonstrated the ability of KIM to locate the prostate with accuracy below other uncertainties in radiotherapy treatments, and the feasibility for KIM to be implemented across multiple institutions.