Monte Carlo προσομοίωση μαστογραφικής απεικόνισης με ακτινοβολία συγχρότρου
Loading...
Date
2014-06-10
Authors
Φυτούση, Νίκη
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In the framework of this thesis, a simulation model, based on Monte Carlo techniques, was developed for the study of breast imaging using Synchrotron Radiation (SR). The basic core of the model was developed in previous dissertations of our Department for conventional mammography. The existing model was expanded to include SR physical and geometrical characteristics and test the potential of SR for further optimization in breast imaging. The SR model was validated against experimental data from SYRMEP beamline of the Elettra Synchrotron Light Facility.
The alterations of the new model comparing to the one for conventional mammography mostly concerned the geometry used (source-to-slit distance 22m), the narrow gaussian almost monoenergetic beam, and the scanning of the region of interest by a uniform movement of the phantom and the detector.
Besides the generation of an image, the model was enriched with dose parameters (incident air kerma, backscatter radiation, entrance surface air kerma-Ke), in order to evaluate SR for both image quality and dose.
The validation showed excellent results for 16-28 keV, for both image quality (subject contrast-SC) and dose (incident air kerma and corresponding number of photons) indices. In the case of SC, Pearson's correlation R was calculated 0.996, while in the case of dose validation, R was equal to 1. Regarding the backscatter, the validation was based on published data and the deviation did not exceed 2% for 20 keV.
The performance of SR was then evaluated with mathematical phantoms designed to simulate difficult imaging tasks; the first experiment concerned a 4cm-thick phantom made of adipose tissue with embedded spheres of PMMA, glandular tissue and water of increasing diameter; the second one concerned a step wedge of 0-75% glandularity in a 5cm-thick background of adipose tissue and small calcifications in each step.
The first experiment was used to compare SR energies (16-25 keV) to conventional mammographic spectra (Mo/Mo, Mo/Rh, Rh/Rh, W/Mo, W/Rh, W/Nb and W/Pd at 28 kVp) in terms of SC and CNR (Contrast-to-Noise Ratio). The results demonstrate that there is an energy range where CNR maximized (18-21 keV). The image quality indices are highly affected by the size and composition of the lesion, with PMMA showing a slightly degraded CNR compared to the same size glandular tissue inhomogeneity. SR energies between 18-21 keV demonstrate improved imaging performance compared to conventional spectra.
For the second experiment, a Figure of Merit (FOM=CNR^2/Ke), was used as an index of the overall perfomance. The energy on the detector was kept constant at 7μGy for SR energies 19-25 keV (for 20 keV, the corresponding Ke was 1.5mGy). The results showed that the best performance (highest FOM), is observed in higher energies. However, taking into consideration the contrast-detail visibility, the best performance was observed at 21-22 keV.
Synchrotron Radiation seems promising for breast imaging, since it shows better performance in cases where conventional mammography faces limitations. However, further exhaustive performance studies, in terms of resolution and dose, are necessary in order to consider it as a reliable alternative for mammographic applications.
Description
Keywords
Synchrotron radiation, Breast imaging, Monte Carlo simulation, Subject contrast, Dose, Mammography