FriXy-gel Dosimetry

What is a FriXy-gel dosimeter?

A FriXy-gel dosimeter is a FAX (Ferrous-Sulphate/Agarose/Xylenol) layered-gel dosimeter imaged with a portable optical system. Conformal radiotherapy can take advantage of this dosimeter: in fact it is possible to design FriXy-gel phantoms simulating the situation of interest, with good tissue-equivalence for all kinds of radiation, except thermal neutrons (in the standard composition). By properly changing gel composition, tissue-equivalence for thermal neutrons is obtained, too and interesting perspective of separating various dose contributions apperars as a promising development.
The absorbed dose can be 3D imaged with a spatial resolution better than 1x0,5x0,5 mm3.

How does a FAX dosimeter work?

As known, in ferrous sulphate solutions ionising radiation starts a chain of chemical reactions which results in the conversion of ferrous ions Fe2+ into ferric ions Fe3+. The conversion yield has shown to be proportional, till saturation, to the absorbed dose. Therefore, after ionising radiation, from the variation of some detectable physical parameter depending on the ferrous and ferric ion amounts, the absorbed dose can be indirectly determined. In conventional Fricke dosimetry, the light absorption at about 300 nm is utilised, because such an absorption, negligible before ferrous ion oxidation, results to be proportional to the ferric ion concentration, that is to the absorbed dose. Spectrophotometric analysis has proved to be very reliable. Moreover, the different paramagnetism of ferrous and ferric ions gives an interesting technique for dose measuring: in fact, Nuclear Magnetic Resonance (NMR) analysis gives the possibility of spatial determination of paramagnetic species, because of their different influence on the spin relaxation times of the hydrogen nuclei of the solution. On account of this consideration, the feasibility of measuring 3-D distributions of absorbed dose in Fricke-infused gel-phantoms by NMR imaging has been suggested (1,2). The sensitivity of such a technique is lower than that of spectophotometry, but this disadvantage is counterbalanced by the fact that, when ferrous sulphate solution is incorporated into a gel, the ferrous ion oxidation yield has resulted to be considerably higher. In previous works, we have enquired the feasibility of dose imaging by means of NMR analysis and the possibility of applying such a technique in thermal neutron fields for BNCT. The main drawback consisted in the not negligible diffusion of Fe2+ and Fe3+ ions in the phantom. This effect causes a continuum loss of spatial resolution during the time between irradiation and analysis, so that a prompt phantom imaging after exposure is necessary to achieve good spatial resolution. Very often it is difficult to have such a possibility, in particular when exposures are performed in a nuclear reactor.
Therefore, we have considered an alternative technique for gel analysis, utilising spectrophotometry. The proposed method for gel-phantom imaging is based on transmittance measurements; we have designed and constructed a very simple portable instrument for image detection, which can be quickly assembled near the irradiation facility.
A considerable enhancement of the sensitivity of optical analysis is obtained by adding to the gel components a proper metal-ion indicator, which yields absorption in the visible spectrum. We have chosen Xylenol Orange (Fluka Chemie) which induces absorption at about 585 nm (3). In Fig. 1 the difference in Optical Density between irradiated gel-samples and reference gel sample, as measured with a spectrophotometer, is shown.

Fig.1 Difference in Optical Density between irradiated gel-samples and reference gel sample. 

The difference in absorbance, at this wavelength, between irradiated and non-irradiated gels has shown to be linearly correlated to the absorbed dose. Visually, by increasing the absorbed dose, the colour of this Fricke-Xylenol-Orange infused gel (which for the sake of brevity we call FriXy-gel) changes from orange to violet.

References

1. J.C.Gore, Y.S.Kang, R.J.Schulz, Measurement of radiation dose distributions by nuclear magnetic resonance (NMR) imaging, Phys. Mad. Biol. 29, 1189-97, (1984)

2. R.J.Schulz, A.F.deGuzman, D.B.Nguyen, J.C.Gore, Dose-response curves for Fricke-infused gels as obtained by nuclear magnetic resonance. Phys. Med. Biol. 35, 1611-1622, (1990)

3. A.Appleby and A Leghrouz, Imaging of radiation dose by visible color development in ferrous-agarose-xylenol orange gels. Med. Phys. 18, 309-312, (1991)

How does the FriXy-gel image detection system work?

The analysis technique is based on transmittance imaging performed by means of a CCD camera. In order to measure transmittance, the phantom to be inspected is composed of a set of piled up gel layers. Each layer consists of a stratum of gel within two transparent polyethylene or mylar films, held by a proper frame of the desired thickness and shape (Fig.2).

Fig.2

After exposure of the whole phantom to ionising radiation, each layer is promptly imaged and from the so obtained 2D images, the 3D distribution is reconstructed by means of convenient software. The instrument for transmittance image acquisition is composed of a CCD camera, an optical filter, a light diffuser and a PC. The interference filter (585 nm) is placed between the 50 mm camera lens and the CCD detector, to match the wavelength of the absorption maximum. A schematic view of the instrument is shown in Fig. 3a and an image of the setup is shown in Fig. 3b.

Fig.3a Instrument for imaging

Fig 3b

The absorbed dose can be correlated to the difference of optical absorbance between irradiated and non-irradiated gels. The GL values can be easily converted in differences of absorbance value, or Optical Density (DOD) with simple mathematics: 

The acquired transmittance images include a stripe of transmittance standards, with different optical densities. In a first step, the Grey Level values measured on the strip are utilised to test the stability of the light source and to evaluate suitable correction factors. Moreover, with properly made software for image elaboration, the Optical Density images are obtained by means of direct dot elaboration of GL images. Finally, if some gels are exposed to known doses and analysed, then the gamma-calibration curve is obtained and transmittance images can be converted into dose images. For attaining good result reliability, the calibration procedure has to be performed with gel samples arranged in the same gel preparation, and moreover irradiation and analysis have to be carried out in an interval of time as short as reasonably possible, preferably in the same day.

Results with gamma-rays 

In Fig.4 a series of gel samples uniformly irradiatied with different gamma doses are shown. The change in color from yellow/orange (unirradiated gel) to violet (~8 Gy) is evident.

Fig.4

Fig.5 shows a zoom of the central region (15mmx15mm) of the image of a gel sample irradiated with a collimated gamma beam.

Fig.5

Software for the imaging system

Proper software has been designed and developed to process the acquired images. This software, that we name FriXyDataToolkit or more frequently, for sake of brevity, FXY software, processes several mathematical steps: the first one consists in individuating and reducing all factors, like as non-uniformity in illumination, which can alter the detected data.
Then, FXY proceeds with the proper pixel-to-pixel manipulation of the two acquired images (before and after irradiation) of each gel layer. This software provides all necessary calculation such as linear bi-dimensional smoothing, contrast enhancement, tracing of isodose curves and so on.
A peculiar feature of the FXY code is the possibility of visualising on screen the resulting data in a true 3D. This means that the user can see the volume of interest from different angles. It is possible to show on screen a particular section of the irradiated volume with a chromatic dose level representation.
The software has been written in C++ code and it is articulated in three distinct sections. The first one regards the simple manipulation of single images. The second one performs the manipulation of a stock of images as a whole project. This allows a complete data processing with calibration of gel response, calibration of all the used hardware (non-uniform light source, instability in the alimentation power, etc.). The third section provides the visualisation of results. This code performs simple 2D images and 3D images, too.
In fact, it is of main importance to design proper software performing true 3D feature, in order to have realistic spatial representation allowing rapid knowledge of dose distribution. This 3D visualisation is performed by building spatial images starting from the images of every gel layer.
It is possible to choose different image-filters and functions at every step of calculation. The results can be saved in two different kinds of files: simple 24 bit .BMP file and .FXY. The last one is a special format that registers only the essential data with the suitability to append various explanation texts on the image history. This new format has been designed to increase calculation velocity.
All sample images are converted into numerical 2D-matrices and after optimal filtering (smoothing, contrast enhancement, etc.) this software extracts iso-contour lines.
After image filtration and iso-contour calculation, the 3D matrix is composed.
The final 3D visualisation is obtained by utilising the algorithms of triangular mashes, marching squares and marching cubes, with a continuous control of the transparency of the 3D image. The control of transparency allows the user to see the inner 3D construction and to estimate some particulars otherwise not visible. It is also possible to calculate, besides 3D iso-contours, the area and volume of the regions of interest.
All 3D rendering functions have been developed utilising VTK (Visualization ToolKit) libraries and the final rendering is performed in OpenGL windows.
To achieve best clinical planning treatment, FXY code was designed with the aim of giving the capability of performing comparison between experimental data from measurement and theoretical data obtained from external calculations.
The above-cited hardware calibration is performed by a properly designed code. In order to amend with respect to not-uniformity of illumination and sample thickness, the software utilises three images: the image of the illuminator and the two images of the transmitted light from blank and from irradiated sample. After proper pixel by pixel elaboration, the corrected image is reconstructed. Moreover, with the software also the instability in time of the illuminator power can be rectified, utilising proper grey-level reference strips. 

The described software has been employed in the analysis of the results of some experimental measurements, in order to test its capability.

Fig 6

In Fig. 6 the acquired image of a gel layer irradiated with a Gamma-Knife is shown. In Fig.7 the iso-contour graph (Fig. 7a) and the bi-dimensional profile (Fig. 7b), generated with the FXY software, are reported. The effect of noise and of non-uniform illumination is evident in both images. In Fig. 8a and 8b a zoom of the central region (10mmx10mm) of two images (before and after the smoothing process operated with the FXY software) are shown.

Fig 7a

Fig 7b

Fig 8a

Fig 8b

Starting from the set of 2D matrices, the VTK structured-point file is created using C++ code. This file is utilized as input for the Tcl/Tk script, which provides the rendering of isodose surfaces, in a completely interactive manner.

To show an example of the possible 3D visualizations, in Fig. 9 the isodose surface corresponding to 50% of maximum dose is reported.

Fig 9

In order to evidence the effect of smoothing process, in Fig. 10 an example of isodose surface before (Fig. 10a ) and after (Fig. 10b) is reported.

Fig 10a

Fig 10b

All the obtained results, like those here reported, evidence that the here described method allows reliable 3D dose measurements, with millimetric spatial resolution, performed in a simple and quick technique and translated in a convenient format for prompt comparison with calculated data and with diagnostic output.

Software developement:

All FXY software code has been developed in the FriXy and TLD Laboratory as part of graduation theses.