Characterization of EPID software for VMAT transit dosimetry

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiotherapy Quality Assurance

Background:

• Transit dosimetry using electronic portal imaging devices (EPID) offers a method for verifying radiotherapy dose distributions.
• Commercial software like Dosimetry Check (DC) aims to reconstruct 3D dose distributions from EPID images.

Purpose of the Study:

• To evaluate the suitability and accuracy of Dosimetry Check (DC) software for volumetric modulated arc therapy (VMAT) transit dosimetry.
• To assess DC's sensitivity in detecting VMAT delivery, set-up errors, and anatomical variations.
• To investigate the influence of the dose computation algorithm on DC accuracy, particularly in inhomogeneous tissues.

Main Methods:

• Compared VMAT dose distributions reconstructed by DC with those calculated by a treatment planning system (TPS) using volumetric gamma analysis (3%/3 mm).
• Assessed DC accuracy against PTW OCTAVIUS®4D (OCT) in homogeneous phantom conditions.
• Introduced deliberate delivery and set-up errors, as well as anatomical variations, to evaluate DC's sensitivity.
• Examined the impact of density inhomogeneity on the dose computation algorithm within DC.

Main Results:

• DC showed agreement with TPS dose maps comparable to OCTAVIUS®4D in homogeneous phantoms (DC: 94.2% ± 3.4% vs. OCT: 95.6% ± 2.5%).
• Introduction of errors significantly reduced the gamma agreement index, with 38/56 cases exceeding the detection threshold.
• The DC dose computation algorithm demonstrated accuracy across anatomical sites, with a noted reduction in algorithm dependence for lung cases using an 'aqua vivo' approach.

Conclusions:

• Dosimetry Check (DC) software accurately reproduces VMAT 3D dose distributions in phantom studies.
• DC is sensitive to errors arising from delivery inaccuracies and patient-specific anatomical variations, making it a valuable tool for VMAT quality assurance.
• The software's performance is comparable to established QA methods, with potential for improved accuracy in complex cases through specific algorithmic approaches.