Iranian Journal of Medical Physics
Volume:3 Issue: 13, 2007

Diffusion Tensor Imaging (DTI) is a non- invasive tool for estimating the diffusion of water molecules in white matter tracts. In this method, a second order tensor is assigned to each voxel. The principal eigenvector of each tensor determines the diffusion direction in the voxel. White matter fiber tracking algorithms reconstruct three dimensional pathways using the vector field of DTI dataset. Fast Marching algorithm is a useful method for fiber tractography in which the co-linearity of principal eigenvectors determines the speed of diffusion. In this research, the speed function has been modified. The proposed speed function has an adaptive fractional anisotropy weighted factor which can set the speed according to brain environment (isotropic and non-isotropic). By this modification, the algorithm can extract the fiber voxels with high accuracy. This method has high performance in fiber crossing detection. The ability of the modified algorithm in the presence of different noise levels has been assessed and it is shown that the proposed algorithm is robust in the presence of noise.Discussion and In the modified algorithm, the eigenvalues have been considered in the adaptive weighted factor. This factor changes the speed value adaptively according to brain environment. By this modification, the diffusion front can be entered into the crossing regions because FA threshold is not considered in this algorithm. Eliminating the FA threshold leads to increasing the time of tractography by a factor of almost two and half. Additionally, the algorithm becomes robust in comparison to noise because for high FA value the effect of noise is reduced.

Brachytherapy sources of 103Pd with low photon energies are being used for interstitial implants in prostate tumors. Dosimetric parameters of these sources such as air kerma strength, dose rate constant, geometry function, radial dose function and anisotropy function should be calculated and measured accurately according to the recommendations of American Association of Physicists in Medicine (AAPM) addressed in TG-43U1 protocol. The first type of 103Pd brachytherapy seeds have been manufactured at Agricultural, Medical and Industrial Research Institute (AMIRI)- Atomic Energy Organization of Iran. Therefore, it is necessary to determine the TG-43U1 dosimetric parameters before clinical implementation.In this work, the radial dose function, anisotropy function and anisotropy constant for the first 103Pd seed manufactured at AMIRS have been determined using TLD-100 dosimeters. The TLD having dimensions of 3.1 ×3.1×0.9 mm3 were calibrated using 60Co radiation field. The TLD measurements were preformed in 30×30×30 cm3 Perspex phantoms in order to provide full scattering conditions. In this work, two phantoms were designed. The first phantom was used for the experimental determination of radial dose function and the second one was used for the experimental determination of anisotropy function. Dose distribution around the source was measured by TLD-100 dosimeters using RMKL KFKI reader. The radial dose function, anisotropy function and anisotropy constant of the first AMIRI 103Pdwereseed determined according to TG-43U1 recommendations. The obtained result was in good agreement with the results published by Meigooni et al for Best double-walled model 103Pd source.Discussion and The spatial dose distribution around the source and dosimetric parameters determined in this work showed that in dosimetry point of view, the new model 103Pd source manufactured at AMIRI is suitable for use in brachytherapy of prostate cancer.

This study is used to compare the bone mineral density (BMD) of lumbar spine (LS) and femur neck (FN) for premenopausal women. The BMD measurements using dual energy x-ray absorptiometry technique (DXA) have been performed at Isfahan Osteoporosis Diagnosis Center since March 2002. Among the referred subjects 185 premenopausal women who had no known history of disease or taking any medication that affects BMD were selected. The long-term reproducibility (coefficient of variation, CV) of the DXA scanner for BMD measurements during the study period was assessed, using the phantom provided by the manufacturer. A Norland XR46 system was used for the investigations. The mean BMD for the FN and LS were found to be 0.859 ± 0.136 and 1.012 ± 0.161, respectively. The long-term BMD for LS and FN had a coefficient of variation of 1.0 and 1.2%, respectively. Discussion and Although a lower BMT T-score has been reported for LS compared to FN for women, the result obtained in this study shows a significantly lower BMT mean T-score (t = -9.02, p<0.0001) for FN. The BMD T-score of -0.551 ± 0.99 SD and -1.09 ± 1.17 SD was found for LS and FN, respectively. This finding is inconsistent with the previously reported results, which may be due to the physiological and life-style factors. Therefore, further research is required to determine the reason(s).

In high energy x-ray radiation therapy, the photons emerging from the target of linac are contaminated by secondary electrons which are referred to as "contaminant electrons". These electrons raise the skin dose when absorbed, therefore, it is desirable to try to remove them. There are different methods including electron filters, He bag and magnetic field which can be used to reduce the contaminant electrons. In this study, the effect of a powerful magnetic field alone and together with a He bag on the x-ray of a Neptun 10 PC linac has been investigated. A magnet holder was designed and constructed to accommodate twenty pieces of NdFeB magnets which produce a very intensive magnetic field (in the order of 0.3 T) in the field center. The percent depth dose value (Ds) at the surface of the phantom was determined in a scanditronix RFA-300 water phantom using a PTW Pin-point chamber. Additionally, a He bag was placed in the air column between the head of the unit and the phantom surface to study the effect of the air column on electron contamination. The obtained measurements showed that the magnetic field causes a reduction in the Ds in 10×10, 20×20 and 25×25 cm2 field sizes by 10.7, 18.3 and 13.3%, respectively. Also the simultaneous application of the magnetic field and the He bag could cause much more reduction in Ds than magnetic field alone.Discussion and Based on other studies, the magnetic field used in this work was intense enough to eliminate almost all of the contaminant electrons emerging from the head of the linac. The share of the electrons produced by the interaction of the photons in the air column between the head and the phantom surface in Ds was only 2-3%, while but for the electrons arising from the head of the linac it was many times higher. To optimize the reduction of the contaminant electrons in the field sizes of 10×10, 20×20 and 25×25 cm2 magnetic field and He bag should be applied. Since the contaminant electrons arising from air column have a minor effect, applying the magnetic field alone is enough to reduce the contaminant electrons.

In diagnostic radiology, the presence of scattered photons effectively decreases the image ‎quality. The introduction of grid has great impact on the improvement of image contrast and as a result on the image quality. A detailed ‎knowledge of scattered radiation distribution is necessary to optimize grid design.‎ In this work, the Monte Carlo Method (MCNP4C code) was used for the calculation of scattered radiation distribution and scatter to primary ratio (SPR). ‎So the variation in the phantom thickness, field size, tube voltage and aluminium filter are considered. Thereafter, the contribution of grid parameters such as strip density, grid ratio, ‎interspace material and lead-to-interspace ratio, height of strip, thickness of strip and thickness of interspace on scatter rejection was investigated by ‎the calculation of bucky factor and contrast improvement factor in 108 linear ‎parallel grids under standard ‎conditions proposed by International Electrotechnical Commission. Full simulation of photon interactions has been considered for the ‎evaluation of aluminium grids and cotton fiber as ‎interspace material. The simulation of x-ray spectra and photon transport in materials was ‎validated by comparing it to the measured data. An increase of 15% in SPR for tube voltage ‎between 40 and 140 kV was obtained in a standard phantom whereas an increase of 99% ‎was observed for phantom thicknesses between 3 and 40 cm at 120 kV. It has been shown that in grids with similar strip density, both contrast improvement factor and bucky factor increase with ‎increasing grid ratio. The same behavior was observed for grids with similar grid ratio ‎when strip density is decreased. Discussion and The grids with cotton fiber show better performance than aluminium interspace grids with the same grid parameters especially in ‎low strip density and high grid ratio. Likewise, the grids with high strip density and high ‎grid ratio perform better than low strip density ones.

In this work, the simulation of gamma-camera system is performed using MCNP code for the comprehensive evaluation of various imaging parameters in nuclear medicine. By employing this code, the simulation of various sources having different energy, different gantry geometries, phantom, collimators with different capabilities and various detectors is possible. In addition to the mentioned capa bilities, studying the effects of reconstruction algorithm, scattering media, collimator wall and edge absorption is also possible. In this paper, the ability of the developed code for the simulation of gamma-camera system is evaluated and the obtained result was compared to the experimental data of a SPECT system with LEGP collimator. In this simulation, the effects of source distance from collimator, scattering media and the combination of these two effects on FWHM and FWTM of line spread function (LSF) is considered and the simulation results are compared to the results obtained experimentally. The obtained simulation and experimental values of FWHM and FWTM were in good agreement while increasing the source depth in the scattering media. For the simulation, increasing the distance between the collimator and source in air (2 cm) causes an increase of 10.9 and 10.5% for FWHM and FWTM, respectively. The experimental values are 10.5 and 11.4%. In the next step, the increasing values of FWHM and FWTM are 18.2 and 23.6% in simulation and 19.93 and 23.34% in scatter media for simulation and practical modes, respectively. By removing the distance effect the simulated value of FWHM and FWTM changes by 1.9 and 9.8%, respectively. The obtained experimental value of FWHM and FWTM changes by 1.08 and 9.12%, respectively.In all cases, a good correlation between the slope of simulation and experimental results were observed.Discussion and In all cases, there is a good correlation between the simulation results and the experimental data. Considering the capability of the developed code for simulating gamma-camera system it is now possible to evaluate the quantitative and qualitative effects of many parameters such as inter-crystal scattering, scattering inside collimator, beam absorption by the edges of collimator septa on system operation and acquired images which otherwise obtaining them are experimentally impossible.

One of the causes of error and perturbation in isodose curves in conventional radiotherapy treatment planning is the existence of tissues having either very high density (bone, prosthesis) or very low density (lung, air cavities). Nowadays, the use of CT images to solve this problem is growing, and so is the need for calibration curves that convert the CT-number of the tissue to electron density for dose calculation. The conventional method to obtain this curve is purely measurement-based, in which a phantom containing various materials of known electron densities is imaged. Alternatively, a more fundamental method of stoichiometry has been used in this work. For the stoichiometric method, initially a cylindrical polyethylene phantom was built. The phantom consists of inserts of high-purity aluminum, PVC, polyethylene, water and cork to model hard bone, skeleton, fat, muscle and lung tissues, respectively. CT imaging was then performed at 120 kVp using a spiral CT scanner (GE model NXI). A system of simultaneous equations was solved to get the appropriate CT-number to electron-density conversion for each tissue type using the CT-numbers from the phantom images, the physical and radiological data of the materials. A conversion curve showing the variation of CT-number with relative electron density was also plotted. The result of the stoichiometric conversion was then compared to that from other methods. The system of simultaneous equations yielded the factors, ,. The experimental and the computed CT-numbers were compared. The highest uncertainty was estimated to be approximately 5.6% for a relatively high-density material such as aluminum and 4% for polyethylene. The curves representing the electron density based on CT-number start diverging at CT-number equal to zero and above.Discussion and Comparing the results obtained from the experimental and computational methods suggest an acceptable level of accuracy for the computational (stoichiometric) conversion. The uncertainty in the electron density obtained is greater for materials of higher electron density.

A genetic algorithm (GA) in conjunction with neural network was proposed to be used to differentiate between breast lesions based on the mass and microcalcification findings. These findings were encoded as features for a genetic algorithm for feature selection and classified with a three-layered neural network to predict the outcome of biopsy. The system was established through the optimization of the classification performance of neural network which was used as evaluation function. A database containing records of 119 patients (49 malignant and 70 benign cases) each of which consisted of 12 parameters was used. The artificial neural network (ANN) was trained using 79 cases containing masses and microcalcifications previously diagnosed by surgical biopsy and tested with 40 cases. The performance of hybrid model was then compared to that of the experienced radiologist in terms of sensitivity, specificity, accuracy and ANN was used to determine receiver operating characteristic curve analysis. The proposed approach was able to find an appropriate feature subset (5-7 parameters out of 12 parameters) and can approximately show the relative efficacy of each feature in classification of breast lesions. The optimized subset of features helped neural classifier to achieve better results in the classification of breast lesions. The obtained results showed that the neural network with optimized features yielded a higher diagnostic accuracy (80%), sensitivity (75%) and specificity (83%) compared to that of the radiologist (68%), (%83) and (61%). Similarly the output of the neural network with optimized features (hybrid model) outperformed the complete model by improving the accuracy, sensitivity and specificity from 70, 56 and 79% to 80, 75 and 83%, respectively. Discussion and The benefits of applying the GA as a preprocessor include improvement in the generalization ability of ANN and reducing the size of calculations through simplifying the ANN structure. Using such hybrid method demonstrated that the performance of the ANN improved by definition the importance of each evaluated features. This can lead the radiologist to a more accurate diagnosis in the clinical applications using such hybrid model as a computerized second opinion.