Comparison of two different EPID-based solutions performing pretreatment quality assurance: 2D portal dosimetry versus 3D forward projection method

Purpose: The aim of this paper is to characterize two different EPID-based solutions for pre-treatment VMAT quality assurance, the 2D portal dosimetry and the 3D projection technique. Their ability to catch the main critical delivery errors was studied. Methods: Measurements were performed with a linac accelerator equipped with EPID aSi1000, Portal Dose Image Prediction (PDIP), and PerFRACTION softwares. Their performances were studied simulating perturbations of a reference plan through systematic variations in dose values and micromultileaf collimator position. The performance of PDIP, based on 2D forward method, was evaluated calculating gamma passing rate (% GP) between no-error and error-simulated measurements. The impact of errors with PerFRACTION, based on 3D projection technique, was analyzed by calculating the difference between reference and perturbed DVH (%Delta D). Subsequently pre-treatment verification with PerFRACTION was done for 27 patients of different pathologies. Results: The sensitivity of PerFRACTION was slightly higher than sensitivity of PDIP, reaching a maximum of 0.9. Specificity was 1 for PerFRACTION and 0.6 for PDIP. The analysis of patients' DVHs indicated that the mean %Delta D was (1.2 +/- 1.9)% for D2%, (0.6 +/- 1.7)% for D95% and (- 0.0 +/- 1.2)% for Dmean of PTV. Regarding OARs, we observed important discrepancies on DVH but that the higher dose variations were in low dose area (< 10 Gy). Conclusions: This study supports the introduction of the new 3D forward projection method for pretreatment QA raising the claim that the visualization of the delivered dose distribution on patient anatomy has major advantages over traditional portal dosimetry QA systems.