52. Transmission detector with 3D dose reconstruction: compass system validation for patient specific dose verification

Abstract

Introduction Methods used for external radiotherapy plan verification do not generally provide clinically relevant results. Compass 4.0 software combined with Dolphin (IBA) transmission detector is a plan verification system which reconstructs the delivered dose distribution inside the patient’s CT-dataset. This study aimed to evaluate this system for VMAT plans verification. Methods The transmission detector is a 2D array of 1513 ionization chambers. With an independent beam modeling and a collapsed cone dose engine, the system provides two ways for treatment plans verification [ Fig. 1 ]: a dose calculation verification with an independent dose calculation; a verification of the actual dose delivered by the linear accelerator (measurement-based dose reconstruction). VMAT plans, using 6 MV photons beams, were generated in Pinnacle v9.10 (Philips) TPS. Elekta linear accelerators Synergy and Versa HD, equipped with an Agility multi leaf collimator (MLC) were used for beams delivery. System validation was realized in two steps: (i) Verification of beam modelling and dose reconstruction method: both Compass calculated dose distribution and reconstructed dose distribution were compared to radiochromic films measurements (gamma index 2D, criteria 3% local/3 mm, threshold 30%) for 11 Head and Neck (H&N) plans. (ii) Evaluation of error detection abilities for two treatment plan complexities (prostate and H&N). Two kinds of error were introduced: MLC positional uncertainties from 1 to 5 mm (one leaf bank shift, two banks shifts in same and in opposite direction) and output errors (MU increase from 1% to 5%). The system was used for 17 clinical cases (nine H&N, eight lungs including SBRT). Measurement-based dose distributions ere compared with planned dose distributions (volume of PTV receiving at least 95% of prescribed dose, V95%, mean dose to organs at risks, Dmean). Results On average, Compass calculated dose distributions compared with film measurements showed a mean gamma value ( γ mean) of 0.422 (standard-deviation σ = 0.078) and a percentage of accepted points (%accepted pts) of 95.7% ( σ = 3.7%). Reconstructed dose distributions compared with film measurements showed a γ mean of 0.380 ( σ = 0.072) and a %accepted pts of 96.8 % ( σ = 2.1%). The system could detect 1 mm leaf bank shift for all the simulated leaf banks shifts, 1% and 3% MU increase for H&N and prostate plan respectively. For clinical cases, differences between planned dose and Compass reconstructed dose were on averaged 1.5% ( σ = 3.8%) for V95% and 53 cGy ( σ = 137 cGy) for the organs at risk. Conclusions The performances of Compass software combined with Dolphin transmission detector were validated for clinical implementation: Compass calculated and reconstructed doses were in agreement with film measurements, 1 mm leaf bank shift error was detected. The dose reconstruction on patient CT-dataset gives a relevant clinical meaning to patient quality assurance results. An interesting evolution of the product would be the delivered dose reconstruction on daily CBCT, for adaptative radiotherapy purpose.