Program Information
Simple and Automatic Calibration Technique for Plastic Scintillation Dosimeters
C Laliberte-Houdeville*1,2,3, L Beaulieu1,2,3, L Archambault1,2,3, (1) Universite Laval, Quebec, QC, (2) Centre de recherche sur le cancer, Universite Laval, Quebec, Qc, (3) Centre de recherche du CHU de Quebec, CHU de Quebec, Quebec, Qc
Presentations
TH-CD-304-9 (Thursday, July 16, 2015) 10:00 AM - 12:00 PM Room: 304
Purpose:
To validate a new faster, simpler and potentially more accurate calibration technique for plastic scintillation dosimeters (PSD). The proposed technique calibrates the dosimeter automatically using either a specific irradiation sequence or simply by using it during treatments or QA.
Methods:
PSDesigner, a multi-purpose PSD simulation platform written in Python, is used to simulate PSDs under a 6 MV, 600MU/min calibrating sequence at dmax. The signal is generated as a linear combination of experimentally measured emission spectra. The user specifies both the beam sequence parameters (jaws positions and treatment table displacement) and the optical probe geometry (e.g. positions of the scintillating elements). As noisy measurement data are collected, a computational technique, Independent Components Analysis (ICA), evaluates the system’s eigenvectors that compose the calibration matrix. Known dose profiles are used to convert the eigenvectors units from current (from collected light by a PMT) to dose. Error on doses is evaluated as a function of total irradiation time to quantify the system’s performance and ability to calibrate itself automatically as it is being used.
Results:
The decreasing error on doses measurements as the probe is being irradiated validates the use of ICA to calibrate the system. A stable calibration is achieved in less than 30s of real-time irradiation, showing promising results in favor of future auto-calibrating PSDs. Comparison between noisy and noise-less simulations also indicates that such system is insensitive to random errors. Introduction of a system defect shows how the system can correct itself by shifting its original set of eigenvectors. The system thus exhibit auto-correction and auto-diagnostic capabilities.
Conclusion:
This work validates a new auto-calibration technique for PSDs that relies on ICA. Unlike traditional calibration techniques, this simple and straightforward method does not get more complex as the number of scintillating elements and measurement points increases.
Contact Email: