Assessing Gamma Index Sensitivity to Selected Linear Accelerator Mechanical Errors Using Various Dosimetric Systems

Background: Volumetric modulated arc therapy (VMAT) for stereotactic body radiotherapy (SBRT) involves highly modulated dose distributions and steep dose gradients, increasing the importance of patient-specific quality assurance (QA). Gamma index analysis is widely used for comparing measured and treatment planning system (TPS)-calculated dose distributions; however, its sensitivity to small systematic mechanical errors remains controversial, particularly across different detector systems and normalization methods. This study aimed to investigate the sensitivity of various gamma index criteria to intentional linear accelerator-based mechanical errors, including multileaf collimator (MLC) positional deviations and collimator rotation errors, using four dosimetric systems. Additionally, global and local gamma analyses were compared to evaluate their ability to detect clinically relevant geometric deviations in SBRT VMAT. Methods: Twenty-five lung SBRT VMAT plans were retrospectively selected. For each patient, six additional plans containing systematic MLC aperture widening errors (0.25, 0.5, 1, and 2 mm) and collimator rotation errors (0.5° and 2°) were generated, resulting in 175 treatment plans. Each plan was delivered to five different patient-specific QA dosimetric systems, resulting in a total of 875 QA measurements. Patient-specific QA measurements were performed using EPID-based EPIQA, ArcCHECK (SNCPATIENT/3DVH), COMPASS (Matrixx Evolution), and Octavius 4D systems. Gamma analyses were conducted using 3%/3 mm, 3%/2 mm, 2%/2 mm, 2%/1 mm, 1%/2 mm, and 1%/1 mm criteria with a 10% dose threshold. Pearson correlation analysis assessed relationships between global and local gamma results. Linear regression slopes quantified sensitivity to MLC aperture widening errors. Receiver operating characteristic (ROC) curves and area under the curve (AUC) values were calculated to evaluate error detectability. Results: Gamma sensitivity varied considerably depending on the characteristics of the dosimetric system (including detector resolution, reconstruction algorithm, and evaluation method) as well as the selected gamma criteria. The 2%/1 mm criterion consistently demonstrated high sensitivity across all systems, providing strong discrimination of ≥0.5 mm MLC errors and near-perfect AUC values for ≥1 mm deviations. The 1%/1 mm criterion yielded the steepest regression slopes but produced unstable passing rates, including in error-free plans, particularly in lower-resolution systems. Conventional 3%/3 mm and 3%/2 mm criteria frequently failed to detect clinically relevant submillimeter MLC deviations. EPID-based EPIQA showed the highest overall sensitivity and AUC performance for small errors, whereas ArcCHECK and Octavius 4D exhibited reduced responsiveness to subtle deviations. Conclusions: Gamma index performance in SBRT VMAT QA is strongly dependent on detector characteristics and gamma configuration. Among the evaluated criteria, 2%/1 mm provided the most balanced combination of sensitivity and clinical stability. Widely used 3%/3 mm thresholds were insufficient for detecting submillimeter geometric deviations. Detector-specific and technique-specific gamma protocols should be implemented to ensure reliable identification of clinically relevant mechanical errors in high-precision radiotherapy.