Use of scintillating fibers for dosimetry

Размещено на http://www.allbest.ru/

Use of scintillating fibers for dosimetry

dosimeter detector radiographic fiber

Several types of dosimeters have been developed to measure the precise dose and the percentage depth dose (PDD) using detectors such as ion chambers, silicon-diode detectors, diamond detectors, liquid ion chambers and radiographic films. They are partially ineffective, considering requests for new approaches to therapy such as: small sensitive volume, dose rate independence, measurements in real-time, use of high doses, no dependence on environmental conditions, such as temperature, humidity and pressure. For this reason new dosimetric systems have been studied and proposed in recent years (Aoyama et al., 1997; Staub et al., 2004; Archambault et al., 2005; Consorti et al. 2005).

These include dosimeters based on scintillating fibers. Difficulties arising in the conventional techniques have been overcome by providing features such as: small sensitive volume, high spatial resolution, water or tissue equivalent material, linear response, independence from energy and from dose rate, independence from environmental conditions such as temperature, pressure and humidity, high resistance to radiation exposure, real-time response.

One of the first works that showed the potential of scintillating fibers to measure the depth dose for electron beams was published in 1997 (Takahiko et al., 1997). Over the past 10 years several proposals have appeared in literature with the use of scintillating fibers. Some of these have used small volumes of scintillating fibers coupled to light guides for reading the response to light (Archambault & Arsenault, 2005; Frelin & Fontbonne, 2006; Bongsoo et al., 2008; Staub, 2004). In others, detector elements have been incorporated in a water phantom equivalent in order to have a 3D representation (Fontbonne et al., 2002; Guillot et al., 2010; Lacroix et al., 2008). In all these approaches, the technique of elimination of Cerenkov light was needed. This contribution is very small and negligible compared to the scintillation light in the scintillating fibers. However the Cerenkov light becomes dominant, in the light guides, connecting the scintillating fibers to the electronic readout, when the guides are subject to radiation (Archambault et al., 2006).

The approach described in this paragraph is different because it is very similar to that adopted in X-ray diagnostics (Bartesaghi et al., 2007a, 2007b). The dose absorbed along the beam is read by scintillating fibers assembled in homogeneous planes orthogonal to the beam axis. The light produced is integrated and collected for the read out at the end of each fiber, giving the projection of the signal in one dimension. The fibers assembled in each homogeneous plane are ready at one end. The approach is similar to the reading of onedimensional Radon transformation of the dose absorbed. In this chapter we present the results of the study of the therapeutic beam of electrons (the system can be used for other particles, photons, protons etc.) with a homogeneous layer of plastic scintillating fibers.

Dosimeter characteristics

According to the clinical dosimetry, the absorbed dose is calculated in the biological tissue. For this reason, dosimeters built with equivalent tissue are more suitable. Water is the source material for phantoms taken as equivalent tissue in all electron beam dosimetry protocols. This choice comes from the fact that approximately 2/3 of the human body is made up water. The dose distribution curves in water and in tissues are very similar, because water is an excellent image of diffusion and absorption properties of the human body. However, in many cases the employment of water as a test phantom is not very useful. Thus it can be replaced by solid materials with similar physical properties like polystyrene and perspex (PMMA). In theory, water tissue-equivalent materials should have effective atomic number Zeff (electron number per gram), and a density equal to water. In clinical applications, however, a material with electron density as close as possible to water is fairly suitable (McLaughlin & Chalkey, 1965).

For these reasons the system DOSIORT (IORT Dosimeter) is conceived as a box made of a tissue-equivalent material (polystyrene) with a density = 1.05 g/cmі. Inside the box there are six sensitive layers spaced 4 mm apart and set perpendicularly to the Z direction of the incident beam, as shown in Fig 2.

Each layer is composed of a grid consisting of two bundles of 190 scintillating optical fibers having a square cross section of 0.50.5 mmІ and crossing one over another so as to define a XY layer with an area of approximately 1010 cmІ. We use 380 BCF-60 scintillating optical fibers produced by Saint Gobain Crystals.

The light emitted in the fibers is read-out by two photodiode arrays Hamamatsu S8865-128 for each bundle. The photodiode are read sequentially using the Hamamatsu driver C9118 and the signal is digitized and processed by a computer with dedicated software.

The read-out system is therefore composed of 24 arrays. A dedicated electronic system is able to acquire, process and display the reconstructed electron beam image in real time.

Fig. 2. Scintillating fiber dosimeter. On the left the characteristics of the layers are shown. On the right the photo of the dosimeter and of one double ribbon are shown

The system is thus a solid phantom having a density approaching 1 g/cmі, with sensitive layers of scintillating fibers set at fixed positions in a calorimetric configuration for the containment of electrons of energy 4-12 MeV. The prototype is able to define the physical and geometrical characteristics of the electron beam (energy, isotropy, homogeneity, etc) and to measure the parameters needed to select the energy, the intensity and the Monitor Units (MU) for the exposure: the Percentage Depth Dose (PDD); the Beam profiles; the Isodose curves; the values of dose per MU (cGy/MU).

Another important thing that must be considered is spatial resolution. The spatial resolution is a key feature for next-generation dosimeters. This feature will become more important in the next few years, in fact next-generation accelerators have a beam size up to 3 mm. DOSIORT responds fully to the demands of new approaches to radiotherapy. Its innovative detection system, made up of optical fibers which provide high spatial resolution and photodiodes with a sensitive area of 0.3Ч0.6 mmІ with pitch of 0.4 mm, gives a high spatial resolution approximately 0.5 mm.

The sequence of the most relevant steps before beginning are: electronic noise measurement to be subtracted; choice of the integration time to ensure a dynamic range large enough to have a linear response from the detector electronic; evaluation of calibration factor for each photodiodes response by exposing the fibers homogeneously to the same beam.

There are three different configurations for applying the dosimeter: using one layer (1D) at fixed depth and rotating it around the beam axis, using one double layers (2D) at different depths and using all six double layers (3D) at different depths. All the configurations are able to get the results in real time but the first system gives a more accurate measurement of the dose at fixed depth and it provides a series of measurements at different angles, the second one gives the results in depth more quickly but is less accurate in the reconstruction of the dose outside the FOV (field of view) region, the third is able to give a 3D estimation of the beam in depth into a single measure. However, all configurations require the acquisition of a number of measurements in a time that is in any case about a tenth of that needed to obtain the results with the ionization chambers or other traditional dosimeters. The choice is connected to the level of accuracy needed in the measurements.

Results using one layer of DOSIORT

The response of the detector was tested through exposure to a beam orthogonal to the layer surface using only one layer. The photon beam of 6 MV was selected with a FOV of 8x8 cmІ. The system was rotated manually around the beam axis and the projected data were collected every 5 degrees for a total of 37 positions from -87.5 to 92.5 degrees.

The projection of the dose in arbitrary units at different angles and positions along the rotating axes is shown in Fig. 3, a.

Fig. 3. a) Projected dose at 37 rotating angles around the beam axis; b) Back Projection without filter; c) Iterative reconstruction; d) Filtered Back Projection with ramp filter

The plot shows the typical behaviour of a square FOV with maximum values around the diagonal at -45 and 45 degrees.

The 37 projections were used to reconstruct the transverse profile of the dose. The fibers collect the light generated along their axis, thus each acquired profile corresponds to the projection of the beam delivered along an axis. This system is not dissimilar to image reconstruction problems in tomography, where several projections have to be composed to trace them back to the original image. For this reason the first reconstruction method chosen was the back-projection algorithm widely used in tomography. Nevertheless using a backprojection approach without a filter we obtained a poor profile (Fig.3, b) while with the introduction of a ramp filter (Fig. 3, d) a better reconstructed image can be obtained. We additionally developed a dedicated algorithm based on the principle of the tomographic iterative methods.

The iterative method uses only two orthogonal projections. The choice of the two projections is very important to determine a good result in the reconstruction. We have selected the projections at 0 and 90 degrees. The idea is to sum the data collected by each fiber along an axis with the corresponding fiber for each of the two different angles, weighing the projection contribution on the basis of its concordance with the other projections results.

The image obtained in this way is a weighted sum of the contributions of two projections. Then the difference between the sum along an axis of the reconstructed image and the acquired values is calculated. This parameter is taken as the error to minimize iterating the method. Each value difference is projected back on the reconstructed image to correct the values along the fiber and then a new difference between reconstructed and acquired data is calculated.

The reconstruction using the back projection with ramp filter reproduces quite accurately the dose profile from the maximum to about 10%. Some fluctuations are visible in the flat region. The iterative approach describes quite well the profile in the FOV used [-40 to +40 mm] where the dose varies from 100% to 50%.

Results using one double layer of DOSIORT

The energy 9 MeV was selected for the test of 1 double layer (XY) of DOSIORT. The FOV was set at 4Ч4 cmІ in order to study the energy response containing the doses absorbed.

The data were taken in different acquisition tests. In each test we changed the geometry of the setup superimposing over the previous configuration a sheet of polystyrene of 4.2 mm in water equivalent thickness. In this way we simulated a homogeneous phantom with measurements at different depths.

The double layer 2D was used for the reconstruction of the XY map of dose absorption at different depths. The XY projections collected for each acquisition, after correction for noise and calibration, were used to reconstruct the dose through the iterative method used for 1D detector and explained in (Brancaccio, 2004).

3D Results using DOSIORT

The full detector DOSIORT may be used to measure the dose in three dimensions.

The system, optimized through Monte Carlo simulation as explained in (Lamanna et al., 2009a), is able to contain the full shower produced with electrons of energy 6 to 9 MeV. Greater energies may be measured using the same technique as in the previous paragraph: by superimposing sheets of polystyrene over the detector. The thickness of the sheets must be selected to position the build-up inside the detector. The isodose curves measured in each double layer for a beam of electrons of energy 9 MeV, using a FOV of 40Ч40 mmІ without external sheets are shown in Fig. 4. The dose is reconstructed using the iterative method described previously. The curves are well described in the FOV region. Outside some artefacts connected to the reconstruction method are visible.

The central X slice of the isodoses shown in Fig. 7 are represented on the left side of Fig. 5. The dose profile in depth is well reconstructed. The comparison with the ionization chamber results is done in the right part of the same figure, using the average of the dose in a central surface of area 22 mm, corresponding to 44 scintillating fibers. The measurement through DOSIORT reproduces the PDD curve obtained through the ionization chamber.

Fig. 4. Isodose curves reconstructed through the iterative method described in the test using data taken with 6 double layers of DOSIORT

Fig. 5. Depth dose X profile on the left and PDD curve on the right: DOSIORT compared to the results of the Ionization Chamber

Размещено на Allbest.ru