Site-specific performance and complexity-dependent risk assessment of patient-specific quality assurance using a 2D array detector in CyberKnife radiosurgery.

To evaluate site-specific performance and complexity-dependent risk in patient-specific quality assurance for robotic radiosurgery using a two-dimensional array detector. A retrospective analysis was performed on 137 CyberKnife stereotactic treatment plans. PSQA was conducted using the PTW 1600 SRS 2D array detector. Gamma analysis was evaluated at multiple dose-distance criteria (3%/2 mm, 2%/2 mm, and 1%/1 mm). Delivery complexity was quantified using monitor units and robotic node count and combined into a Composite Complexity Index. Outlier analysis was performed using a clinically relevant action threshold of a gamma pass rate < 90% at 2%/2 mm. Across all sites, gamma pass rates exceeded 95% at 3%/2 mm, confirming robust baseline delivery accuracy. Progressive tightening of gamma criteria revealed pronounced site dependence, with lung and liver SBRT plans exhibiting the most significant degradation at 2%/2 mm and 1%/1 mm, respectively. Complexity effects were minimal at 3%/2 mm but became significant at 2%/2 mm and 1%/1 mm (p < 0.01). Outlier analysis demonstrated low risk for brain and prostate treatments, intermediate risk for spine SBRT with complexity-dependent escalation, and consistently high outlier frequency for lung and liver SBRT independent of complexity. Array-based PSQA for robotic radiosurgery demonstrates high reliability under conventional gamma criteria while revealing strong site- and complexity-dependent behavior under stringent conditions. Anatomical site was the dominant determinant of QA variability and outlier risk. These findings support a risk-adaptive, site- and complexity-aware approach to PSQA interpretation rather than reliance on uniformly stringent acceptance thresholds.