Encrypted login | home

Program Information

A Novel EPID-Based 3D Dose Reconstruction for IMRT QA

no image available
A Alhazmi

A Alhazmi1*, C Gianoli1 , S Neppl2 , S Veloza1 , J Martins1 , M Podesta3 , F Verhaegen3 , M Reiner2 , C Belka2 , K Parodi1. (1) Ludwig Maximilian University of Munich, Garching. (2)LMU Munich, Department of Radiation Oncology,Munich. (3) Maastro Clinic, Maastricht University Medical Center, Maastricht.


SU-I-GPD-T-259 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: The goal of this study is to implement and investigate a prediction free pretreatment IMRT QA methodology for in-house 3D dose distribution reconstruction based on EPID, in comparison to a commercial dosimetry device.

Methods: The raw EPID images acquired in air without a dosimetric phantom underwent offset, dead pixel, and pixel sensitivity corrections. The corrected images were then converted into dose-to-water, relying on well-established methods, producing 2D dose images. A novel back-projection algorithm was implemented to reconstruct the 3D dose distribution, by numerically modeling a virtual cylindrical water phantom. The 2D dose image at each depth of the water phantom was convolved with depth-specific scatter and attenuation kernels. The kernels were obtained relying on scatter and attenuation models for different regular and irregular fields. Thus, depth-specific parameters were defined via numerical optimization, which serve as a look-up table for future reconstruction from future EPID raw images.The 3D dose distribution of two step&shoot IMRT plans (a head&neck and a prostate case) were reconstructed and compared with the corresponding 3D dose distribution from OCTAVIUS, using the recommended global gamma evaluation with (3%, 3mm) criteria.

Results: The acceptance criteria of the gamma evaluation were satisfied with a passing rate of 99.4% and 99.9% within the phantom, for head&neck and prostate patients respectively.

Conclusion: The results suggest that the developed QA methodology can provide comparable accuracy with respect to a commercial dosimetry device (e.g. OCTAVIUS). The implementation in daily routine is compatible with clinical requirements and the set-up time efficiency is optimized thanks to the in air acquisition. Different from any methodology proposed in literature, the 3D dose distribution is obtained independently from any prediction. The qualitative and quantitative validation will be extended to VMAT and statistical evaluation (~20 patients) will be performed.

Contact Email: