Recent Medical Physics Graduate Abstracts

We are pleased to publish a review of medical physics graduate work completed at GreekUniversities. In this issue, work presented in 1999 is reported. These abstracts were collected in part using the Bell and Howell web site. For more information on the abstracts or to order a copy of a dissertation, contact Bell & Howell Information and Learning Company (formerly UMI), 300 North Zeeb Road, Ann Arbor, MI 48106-1346 USA. Telephone (734) 761-7400; E-mail: [email protected]; Web-page: http://www.umi.com/.

Monte Carlo Study Of Photon Beams From Medical Linear Accelerators: Optimization, Benchmark And Spectra

BEAM is a general purpose EGS4 user code for simulating radiotherapy sources (Rogers <italic>et al</italic>. Med. Phys. 22, 503–524, 1995). The BEAM code is optimized by first minimizing unnecessary electron transport (a factor of 3 improvement in efficiency). The efficiency of the uniform bremsstrahlung splitting (UBS) technique is assessed and found to be 4 times more efficient. The Russian Roulette technique used in conjunction with UBS is substantially modified to make simulations additionally 2 times more efficient. Finally, a novel and robust technique, called selective bremsstrahlung splitting (SBS), is developed and shown to improve the efficiency of photon beam simulations by an additional factor of 3–4, depending on the end- point considered. The optimized BEAM code is benchmarked by comparing calculated and measured ionization distributions in water from the 10 and 20 MV photon beams of the NRCC linac. Unlike previous calculations, the incident <italic>e</italic><super> −</super> energy is known independently to 1%, the entire extra-focal radiation is simulated and <italic>e</italic><super>−</super> contamination is accounted for. Both beams use clinical jaws, whose dimensions are accurately measured, and which are set for a 10 x 10 cm<super>2</super> field at 110 cm. At both energies, the calculated and the measured values of ionization on the central-axis in the buildup region agree within 1% of maximum dose. The agreement is well within statistics elsewhere on the central-axis. Ionization profiles match within 1% of maximum dose, except at the geometrical edges of the field, where the disagreement is up to 5% of dose maximum. Causes for this discrepancy are discussed. The benchmarked BEAM code is then used to simulate beams from the major commercial medical linear accelerators. The off-axis factors are matched within statistical uncertainties, for most of the beams at the 1 σ level and for all at the 2 σ level. The calculated and measured depth-dose data agree within 1% (local dose), at about 1% (1 σ level) statistics, at all depths past depth of maximum dose for almost all beams. The calculated photon spectra and average energy distributions are compared to those published by Mohan <italic>et al</italic>. and decomposed into direct and scattered photon components.

Novel methods for event detection and signal compression are presented for the electrocardiogram (ECG). The event detection method uses a simple and fast algorithm which is highly effective. The method is based on the local and fuzzy evaluation of the size of Haar wavelet transform coefficients of the signal. The average error rate of the method was 0.68% over the entire MIT-BIH database which is of the same order as previously published results. The method achieves this level of performance without operator interference, and with globally fixed parameters which is unique in the literature. The compression method is based on active error control, and is the first method which has successfully applied a local error measure to the electrocardiogram compression problem. A resampling strategy based on the physiology of the signal is used to achieve performance improvement and computationally feasible access to singular value decomposition. The method is simultaneously capable of higher compression and higher fidelity of the reconstructed approximation than previously reported results. The average compression rates achieved by the method are 27.1:1, and 15.4:1 for the 100 and 200 series of the MIT-BIH database respectively.

Scatter Factors and Peak Scatter Factors for Cobalt-60, 6 MV, 10 MV, and 18 MV Photon Beams

The aim of external beam radiotherapy is to deliver a prescribed dose to a target volume accurately and uniformly while sparing the surrounding healthy tissue. In radiation dosimetry calculations, many functions are employed to achieve this goal, and the Peak Scatter Factor (PSF) is one of the fundamental functions used in dosimetry.

A brief background of some of the basic physics employed in external beam radiotherapy is given, illustrating some of the applications of the PSF in dosimetry. Also, the evolution of the definition of the PSF is discussed by presenting the PSF definitions quoted in several dosimetric references. In addition, concerns debated among physicists regarding the consistency of tabulated values of the PSF in dosimetric references with the definition of the PSF are presented.

A practical method for measuring the PSF for megavoltage photon beams is developed. The method is applied to C0-60, 6 MV, 10 MV, and 18 MV photon beams using water, polytyrene, and solid water phantoms. The measured PSF's are compared to tabulated PSF and Normalized Peak Scatter Factors (NPSF) published in the British Journal of Radiology (BJR), supplement 25.

Quantitative Analysis of Metabolic Breast Images from Positron Emission Mammography (PEM)

X-ray mammography cannot always distinguish between benign and malignant breast lesions. This leads to unnecessary biopsies, costs, and stress for the patient. Positron Emission Mammography (PEM) provides images of increased glucose metabolism in malignant tumours compared with healthy tissue. After injection of a radioactively-labelled glucose analog, cancerous tumours appear as bright spots on the breast image.

Quantitative analysis of PEM images consists in comparing the amount of activity absorbed in both breasts of a patient. Based on ROC analysis of 15 subjects, an asymmetry of 10% in the number of counts detected from each breast was taken as a sign of cancer. The application of the count asymmetry method seems to result in a 22% improvement of PEM accuracy (from 64% to 86%). It is particularly useful for the detection of big or diffuse tumours. Quantitative data will also provide tools for future applications of PEM technology, such a follow-up of patients after cancer therapy.

Magnetic resonance imaging (MRI) can be used to image diffusion in liquids, such as water in brain structures. Molecular diffusion can be isotropic or anisotropic, depending on the fluid's environment, and can therefore be characterized by a scalar, D, or by a tensor, D, in the respective cases. For anisotropic environments, the eigenvector of D corresponding to the largest eigenvalue indicates the preferred direction of diffusion.

This thesis describes the design and implementation of diffusion tensor imaging on a clinical MRI system. An acquisition sequence was designed and post-processing software developed to create diffusion trace images, scalar anisotropy maps, and anisotropy vector maps. A number of practical imaging problems were addressed and solved, including optimization of sequence parameters, accounting for flow effects, and dealing with eddy currents, patient motion, and ghosting. Experimental validation of the sequence was performed by calculating the trace of the diffusion tensor measured in various isotropic liquids. The results agreed very well with the quantitative values found in the literature, and the scalar anisotropy index was also found to be correct in isotropic phantoms. Anisotropy maps, showing the preferred direction of diffusion, were generated in human brain in vivo. These showed the expected white matter tracts in the corpus callosum.

Coronary arteries obstructed by atherosclerosis can be cleared by a balloon angioplasty accompanied by a permanent scaffolding implant ("stent"). There is, however, a 25-45% occurrence rate of excessive thickening of the treated vessel wall a few weeks post-angioplasty, leading to a re-obstruction of the vessel ("restenosis"), and recurrent symptoms. Endovascular radiotherapy, a potential preventive treatment against restenosis, can be delivered by a stent impregnated with radionuclides. The present thesis examines the local dosimetry of radioactive stents. A computer algorithm, DOSECOP, was developed based on the dose point kernel theory, to calculate the dose distribution created by a ³²P-implanted stent in the surrounding vessel wall. The dosimetric significance of four parameters is also assessed, namely: 1) mechanical distortions of the stent, 2) the exact location of radionuclides on the stent, 3) the self-attenuation of the stainless steel stent, and 4) the width of the stent's struts in the calculation models.

The aim of stereotactic radiosurgery is to deliver a high and uniform radiation dose to the target volume and a minimized dose to the surrounding healthy tissue. Various linac-based radiosurgical techniques are used clinically: multiple non-coplanar converging arcs, dynamic arc rotation, and conical rotation. The techniques differ in their beam distribution over the patient's head.

A study of the beam distribution characteristics for the clinical linac-based radiosurgical techniques is presented. Two spiral linac-based radiosurgical techniques are developed: the uniform dose-rate spiral irradiation and the dose-rate-weighted spiral irradiation. Both exhibit the same spiraling beam entry trace over the patient's head; however, they differ in their beam distribution along the spiral. The dose-rate-weighted spiral irradiation provides a uniform beam distribution over the 2p solid angle available in radiosurgery.

The currently existing techniques and the spiral techniques are then compared using the cumulative dose-volume histogram (CDVH) tools available with the McGill Treatment Planning System (MPS). The dose-rate-weighted spiral technique leads to lower dose inhomogeneities within the target volume and better dose conformity within the target. Moreover, it also encompasses smaller volumes of tissue at all isodose levels with larger differences at low isodose levels. A conclusion is reached that the dose-rate-weighted spiral irradiation technique offers interesting advantages over the currently used clinical linac-based techniques.

Measurements of modulation transfer function (MTF), noise power spectrum (NPS), and detective quantum efficiency (DQE) for metal/film portal detectors are reported for the Cobalt-60 and 10 MV spectra. The detectors consist of a double-emulsion portal film secured between plates of aluminum, copper, brass or lead with thicknesses from 0 to 4.81 mm. The study of MTF, NPS, and DQE shows that both photons and secondary electrons produced within the front-plate and backscattered electrons from the back-plate affect metal/film portal imaging. Study of DQE indicates that the best portal detectors are those without back-plates, and with high density front-plates with thicknesses less than the maximum electron range.

This MFT data was modeled with the logit analysis. It is shown that the parameters resulting from the logit analysis depend on the mass thickness and the atomic number of the metal plates.

Metal/amorphous selenium (a-Se) electrostatic-based detectors have been developed for porta imaging. The detectors consist of a-Se photoconductive layers of varied thicknesses deposited on plates of varying thicknesses of aluminum, copper, and stainless steel. The metal-plates of the detectors face the incident 6 MV and Co-60 photon spectra during imaging. The sensitivity of the a-Se detectors to dose, electric field across the a-Se layer, plate type, and a-Se thickness is studied. A model showing a cubic relationship between the a-Se latent surface voltage and dose is derived and experimentally verified. A contrast-detail phantom is used to study the image quality and contrast-resolution characteristics of the metal/a-Se detectors. The metal/a-Se detectors produce better quality contrast-detailed images at a considerably lower dose than that offered by the other commercial available portal systems, mainly due to the low inherent noise of the novel detectors.

A semi-automatic technique for the direct set-up alignment of radiosurgical circular fields from an isocentric linac to treatment room laser cross-hairs is described. Because film or a-Se is each sensitive to laser and ionizing radiation, they are used to acquire images of the positioning lasers superimposed directly onto the radiosurgical circular fields. An algorithm extracts the co-ordinates of the centre of the collimator image and of the intersection of the laser cross-hair image and subsequently determines the deviation, to within a precision of ~ 0.04 mm. The technique is also used to perform quality assurance on a Clinac-18 linac and shows a (0.53 ± 0.05) mm wobble from the nominal isocenter of the linac.

The present generation of medical linear accelerators is computer controlled providing great precision in dose delivery in addition to other options, such as conformal radiotherapy, which involves intensity modulated fields. These fields are produced with multileaf collimators (MLCs) capable of delivering a radiation beam with a pre-determined modulated intensity using the dynamic capabilities of the MLC leaves. Two dynamic beam delivery methods are currently used: the step and shoot method and the continuous motion method. The first consists of several static subfields with the motion of the leaves occurring without the presence of radiation, while in the other the leaves may move with the beam on.

The work presented here intends to prove that dynamically enabled linear accelerators can be used with confidence by verifying the accuracy and the stability of motions of the movable axes for the two dynamic beam delivery methods. Prior to clinical use, the integrity of the entire beam delivery system must be tested and the dosimetry related to the MLC must be examined. The purpose of this thesis is to develop, analyze and perform tests for the commissioning of the dynamic beam delivery capabilities of a medical linear accelerator and to catalogue these tests to facilitate their implementation in a routine quality assurance program.

This project is to develop methods to retrospectively determine the size of alpha-emitting particles that have been collected on personal air samplers. The alpha radiation from such particles produces a cluster of tracks on the surface of an etched nuclear track detector, CR-39. The number of tracks in a cluster, as well the diameter of the cluster, are dependent on several factors: the diameter of the hot particle, the distance between the particle and the CR-39 surface, the composition of the particle, and the alpha-particle energy.

The dependence of the alpha-emitting particle size and the number of registered tracks were revealed, and produced predictions of the track density distribution observed on the CR-39 plastic. There is a good fit between the simulation of track density observed on the CR-39 and the tracks arising from uranium oxide and plutonium oxide particles.

Beta and Electron Dose Imaging Using a Microspectrophotometer System and Radiochromic Film.

A dose imaging technique of measuring the distribution of beta particles and high energy electrons, using a microspectrophotometer system and radiochromic film, has been investigated. The investigation entailed the commissioning of the microspectro-photometer and the development of a newly proposed method in film calibration. It was tested successfully by comparing the measured dose, determined using the calibration method, of films irradiated by a ³²P source, and the expected dose, determined by beta counting based on the equilibrium geometry with dose loss correction to the film substrate (20%). This method was also demonstrated to extend the usable dose range of the film.

The dose imaging technique was applied to two experiments. The first experiment examines the beta depth dose distribution in Lucite, irradiated by a ³²P volume source. The experimental results were compared to Monte Carlo simulation, performed using the ITS code. The experimental distributions were found to agree closely with the Monte Carlo values up to a depth of 3 mm, but were significantly higher beyond.

The second experiment examines the electron dose distribution in Lucite, irradiated by a narrow 10 MeV electron beam. The lateral experimental dose distributions were found to agree well, except at the tail end, with Monte Carlo values up to about 20 mm depth. Elsewhere, the experimental dose was found to be higher. The disagreement with Monte Carlo simulations at low doses in both experiments perhaps indicates a real discrepancy; further experimental investigations are required. The lateral spread of the beam was also investigated by determining the linear scattering power. The experimental values of the linear scattering power were found to be smaller than those quoted in the ICRU report 35 by a factor of 1.8 or less.

Electroencephalographic Evidence for Auditory Cortical Plasticity in Humans Trained on a Frequency Discrimination Task

Animal studies have shown that the tonotopic organization of the auditory cortex is not statically fixed, but can be remodeled by experience. The purpose of this study was to investigate whether or not frequency discrimination training can induce changes in the cortical representation of a selected frequency in humans. Six human subjects were trained for approximately 3 weeks to detect a change in pitch between two tones (40 Hz amplitude modulated) using a standard frequency of 2040 Hz. Each subject was tested on his/her discriminative ability before and after training using three different standards (2040 Hz, 1840 Hz, and 2240 Hz). EEG data were recorded both before and after training and changes in transient and steady-state responses were investigated. Behaviourally, every subject improved at the discrimination task using the trained frequency. However, only three subjects demonstrated transfer to both untrained frequencies. In the EEG data, the P2-N1 amplitude increased in five of the six subjects and the Ni latency decreased in all six for the 2040 Hz set. These two findings were statistically significant (p<0.05) for the group. There were no statistically significant findings for the side frequencies. The change in the 40 Hz steady-state response was also not significant, increasing in three subjects and decreasing in the other three. These findings indicate that changes are expressed in the secondary auditory cortex. These findings may also be applicable to the treatment of tinnitus.

A Diffusion Theory Model of Spatially Resolved Fluorescence from Depth Dependent Fluorophore Concentrations.

Photodynamic therapy (PDT) currently utilizes drug and light doses which are primarily based on clinical experience. This can lead to a dose which is not sufficient to destroy the entire tumor, or alternatively, it can lead to the undesirable destruction of healthy tissue around the treatment area. PDT of topically applied photosensitizers is one focus of this research. This concerns the diffusion of an externally applied drug into the tissue, as well as its subsequent destruction during the irradiation procedure.

This work involves the non-invasive measurement of the inherent fluorescence of the photosensitizer, allowing the determination of the concentration and distribution of drug within the tissue, and thus optimizing this treatment. To do this, one must be able to describe the propagation of light within the tissue. Consequently, a photon diffusion model has been developed to calculate the steady-state spatially resolved fluorescence from a pencil beam excitation in a depth dependent medium. The validity of this model was then verified by comparison with Monte Carlo simulations and measurements made on phantoms with optical properties similar to those of human tissue. Theoretical conditions were then explored, and potential uses of the model were demonstrated.

Beta-Gradient Isochrons Using Electron Paramagnetic Resonance: Towards a New Dating Method in Archaeology

Electron paramagnetic resonance, or electron spin resonance (ESR), can be used to date mammal teeth found in archaeological sites. This thesis aims to improve the accuracy of the ESR dating technique by focusing on the beta-component of the dose rate. The theoretical distribution in the enamel of the beta dose from uranium, thorium, and potassium sources was determined using Monte Carlo techniques. This method provides the most comprehensive estimates of archaeological beta doses by incorporating the complete beta spectrum, internal conversion and Auger electrons, bremsstrahlung, and effects owing to radon loss and the moisture content of the sediment. It was found that current methods for calculating beta dose rates using One-Group theory and the ROSY software are between 50/0 and 180/0 higher than the Monte Carlo results. In addition, the Monte Carlo results suggest a strong theoretical gradient in the average dose rate to successive layers of enamel.

Based on these results, a new ESR dating method was developed using beta-gradient isochrons and experimental work was conducted to assess the feasibility of this new method. The standard dating method is limited by the model of uranit2im uptake by the tooth during burial. In the proposed method, an isochron, whose slope is equal the age of the tooth, is constructed by plotting the total dose to different slices of uranium-free enamel as a function of the beta dose rate. Only the outer portion of uranium-free enamel is sampled, thereby isolating the constant dose provided by the surrounding sediment from any internal dose. Protocols for the removal and analysis of thin layers of enamel were developed. Using a low-speed saw fitted with a diamond-edged blade, a total of twenty-five layers were removed from four teeth. Isochrons were constructed for the four teeth, and reasonable ages were obtained for the two teeth with accurate sediment information. Therefore, if the sediment directly in contact with the enamel is collected, the beta-gradient isochron method is a feasible dating technique for mammal teeth.

The two-dimensional dose distribution of two ¹²⁵I brachytherapy seed sources recorded in GafChromic^TM dosimetry media has been measured using a scanning spectral microphotometer. The two-dimensional dose rate distribution, as well as the vertical and horizontal dose rate profiles of the two seeds, referred to as the Isogen prototype PR01 and PR03 seed sources, has been presented. The Isogen PR01 seed source has been characterized using the dosimetry calculation formalism recommended by the American Association of Physicists in Medicine (AAPM). The radial dose function, calculated for the PR01 seed source, has been found to correspond with that predicted theoretically.

X-ray fluorescence (XRF) has been demonstrated to be a useful technique for measuring trace quantities of heavy metals in various tissues within the body. This thesis investigates a means of improving the measurement of lead in bone, as well as increasing the existing sensitivity of measuring kidney mercury content. The XRF measurement of uranium is also explored.

This work assesses the feasibility of a normalisation method for the 57Co/900 system, in relating detected signal to the lead content of the sample. The feasibility of normalisation has been shown, which reduces subject dose and improves system transportability, as well as removes subjectivity, by eliminating the need for acquiring planar x-ray images of the measurement site.

In the measurement of renal mercury concentrations, a gain in sensitivity by increasing the x-ray tube operating voltage of the current system is investigated. It was found that 250 kV, rather than 175 kV, and a titanium rather than uranium filter, results in a 2.5 ñ 0.2 times gain in sensitivity. This potential improvement could have profound clinical implications for the accuracy of occupational monitoring, and for assessing whether there is a quantitative relationship between biological fluid levels and mercury content in this critical organ.

The XRF measurement of bone uranium content is also explored. Both source-excited and polarised systems have been developed, however, the sensitivity is currently beyond that which is useful for occupational monitoring of exposure to this toxin. The particular case of measuring uranium in survivors of "Friendly Fire" incidents (from Operation Desert Storm) is investigated, and the first detectable quantity of uranium has been observed in a member of this cohort, with the XRF system designed and built during the course of this work.

Modern treatment planning systems are able to conform more adequately to tumors, but the large volume of data that needs to be processed in 3-dimensional conformal radiotherapy renders the precise reporting and analysis of doses actually delivered to irradiated organs and volumes of interest very difficult. A new method of summarizing and reporting non-uniform dose distributions, better known as the EUD method, is described in this work. The EUD concept assumes that any two dose distributions are equivalent if they cause the same radiobiological effect. In this thesis, the EUD concept is applied to normal tissues, as they are rarely avoided during irradiation. Since normal tissues, unlike tumors, vary from one another in their architecture and therefore behave differently to radiation, only the most fundamental normal tissue architectures, serial and parallel, are considered in this work. Extensions of the idealized EUD concept to include nonuniform density of clonogens, dose per fraction effects, absolute volumes, intra-patient inhomogeneity, and inhomogeneity of patient population are also presented in this work. The application of the basic EUD concept for tumors, and serial and parallel architecture organs is demonstrated with a number of simple dose-volume histograms and relevant clinical dose distributions.

Biological Responses Of Tumour Cells To Freezing Using A Novel Cryosurgical Model System

A novel cryosurgical model system is developed that allows physical and thermal measurements to be correlated with biological outcomes to better understand cellular responses to freezing and ice-ball dynamics. A realistic model system is designed, biological assessment techniques validated and biological freeze-thaw experiments performed and analysed. Cell viability is evaluated using fluorescent stain Syto/EB and metabolic assay alamarBlue after subjecting cells to a single or double freeze-thaw cycle, or freeze cycling. The effects of cooling rate, end temperature, number of cycles, and temperature cycling are investigated, resulting in the double freeze-thaw protocol being the most effective at killing cells. At regions closest to the probe, increased cooling rates and repeated freezes are most important in cell death. In the periphery where cells survive, complete thawing and increased cycle numbers are necessary. Since increases above two freeze-thaw cycles lead to severe complications, adjuvants to increase cell death are necessary in addition to cryosurgery.

"Experimental Determination of Relative Outputs of Sr-90 Ophthalmic Applicators and the Anisotropy Function of the Model 6711 I-125 Seed"

One of the oldest treatment modalities in radiotherapy, brachytherapy, has recently benefited from more accurate dosimetry protocols to aid in precise treatment planning. The main purpose of the thesis was to study the dosimetry of two brachytherapy sources: Sr-90 in ophthalmic applicators and I-125 in seed form for interstitial implants. For Sr-90, particular attention was paid to using a combination of radiochromic films and a document scanner to record the relative surface dose rates from ophthalmic applicators. For the model 6711 I-125 seed, the dose rate around the source was measured using LiF TLD detectors to examine the anisotropy in the dose distribution for varying angles and at different radial distances up to 10 cm. The latter experiments were conducted using two different techniques: one incorporating a pre-read anneal and the other based on glow curve analysis. The results were compared to published Monte Carlo calculations and TG43 recommended values.

The Risk Of Breast Cancer From Hormone Replacement Therapy Combined With Mammographic Radiation Exposure

Previous studies of the relative risk of breast cancer from hormone replacement therapy (HRT) have taken into account possible confounding by age, reproductive history, breast history and type of menopause. In addition to these factors, this study accounted for possible confounding by mammographic radiation exposure and tested for interaction between mammographic radiation exposure and HRT use. The study included women age 60 years and older who had sufficient radiological indications for a breast biopsy. The likelihood of a positive diagnosis of breast cancer associated with current HRT use and mammographic radiation exposure was estimated using a case-control design and involved 2,110 postmenopausal women who had received breast biopsies at a breast imaging clinic between 1989 and 1998. Women who reported current hormone replacement therapy in the study population were found to be less likely to be diagnosed with a malignant breast cancer than women who reported no current HRT use. Also, women who had received mammographic radiation exposure one year or more from the date of diagnosis were less likely to be diagnosed with a malignant breast cancer than women who had not received any past mammographic exposure. However, no interaction was found between mammographic exposure and current HRT use affecting the likelihood of a woman in the study population receiving a diagnosis of breast cancer.

Characterization Of Small High Energy Photon Beams In Homogeneous And Heterogeneous Media

This thesis advances the study of small high energetic photon fields in radiotherapy. Small photon field irradiation is aimed at delivering a uniform dose to a well defined target while minimizing the dose to the surrounding normal tissue. The dosimetry of small x-ray fields is complicated by two factors: the relationship between detector size and field dimensions and the lack of equilibrium in lateral charged particles. Additionally, a longitudinal charged particle disequilibrium is present when materials with different atomic composition and density than water are introduced in a water-like phantom. Small radiation dosimeters such as diamond, diode, film and a mini-ion chamber have a better spatial resolution to detect the steep dose fall-off at the edge of small photon fields than the large Markus chamber. The line spread function (LSF) of the film densitometer can be estimated by simple measurement of a slit image. Deconvolution of the measured beam profile from a linear accelerator (linac) with the LSF of a detector yields an estimate of the true inherent beam profile of the linac. Conversely, the LSF of any detector can be estimated by deconvolution from measured data once the inherent profile is known. Similarly, a blurring function representing the finite source size effect of the head of the linac which is missing in a Monte Carlo simulation can be obtained. Because the deconvolution process is highly sensitive to noise, the Total Least Squares (TLS) approach offers a reasonable means to overcome this problem. To deal with inhomogeneous media, the density scaling theorem has been modified to incorporate the effect of a change in atomic number of a material. This modified scaling found an application in the convolution-superposition dose model and provided better agreement with the Monte Carlo generated data. The idea of electronic disequilibrium has been taken into account in our simple depth dose model. A prototype second order differential equation allowed energy to be carried away, analogous to the notion of electron range, and hence we were able to simulate a build-up region for the depth dose curve as well as inhomogeneities.

X-ray Computed Tomography for Performing Polymer Gel Dosimetry: A Feasibility Study

Radiation therapy treatment of cancer is increasingly concerned with delivering dose distributions that conform to the tumour volume. For verification of treatment planning computer dose calculations, these conformal therapies demand an accurate, sensitive, high resolution three dimensional (3D) dosimeter. Polymer gels are novel, inherently 3D, tissue equivalent radiation dosimeters. Traditionally, dose distributions recorded in polymer gel are read out using magnetic resonance imaging (MRI). Presented here is a feasibility study on a new 3D dosimetry technique that uses x-ray computed tomography (CT) to read dosimetric information from polymer gels. The technique exploits a gel density change that occurs in response to ionizing radiation. This study has three main goals: 1) to develop a protocol for producing quality CT polymer gel images; 2) to evaluate the nature and reproducibility of the dosimeter's CT number (NCT)-dose response; and 3) to compare this technique with MRI polymer gel dosimetry. A quantitative discussion of the density changes occurring in the gel in response to ionization radiation is also provided. Experiments are conducted using a PAG (polyacrylamide and gelatin) gel dosimeter irradiated with four intersecting 10 MV photon beams. The NCT -dose response is found to be linear and reproducible over the range of 200 to 1000cGy. At room temperature the response is (8.7 ( 0.3)x10-3 NCT/cGy resulting in a limited dose resolution, ~ 100 cGy. Gel temperature during imaging is determined to have only a small effect, 0.3%/(C, on the dose response. Spatial resolution is 0.5 mm in the image plane and 10mm-1mm (depending on noise requirements) in the third dimension. Despite the low dose resolution, preliminary results indicate this technique provides accurate localization of high dose regions and, given the availability and speed of CT imaging, has the potential to be a valuable and practical tool for radiation therapy clinics.

Detection Of Soft Tissue Abnormalities In Mammographic Images For Early Diagnosis Of Breast Cancer

Treatment of breast cancer is currently effective only if it is detected at an early stage. X-ray mammography is the most effective method for early detection, however, mammographic images are complex. Researchers have been utilizing image processing and image analysis techniques to assist radiologists in their difficult task of detecting tumors in mammographic images. To aid radiologists in earlier detection of breast cancer, a retrospective study of mammograms was conducted. In this pioneer study, screening mammograms taken prior to the detection of a malignant mass were analyzed. The aim is to determine if there exists any signs of cancer development in the screening mammograms prior to the detection of a mass by the radiologist. For 58 biopsy proven breast cancer patients who were diagnosed by identifying a malignant mass in their mammograms, 224 previous screening mammograms were collected. These mammograms were reviewed by an expert radiologist and three regions were marked on each of the two mammographic projections of each case: (1) the region which corresponds to the site in which the malignant mass subsequently developed, (2) a similar normal region on the same mammogram, and (3) the normal region on the previous screening mammogram of the opposite breast which corresponds to region 1. Sixty-two texture and photometric image features were calculated for all the marked areas. A stepwise discriminant analysis found that six of these features best distinguish between the normal and abnormal regions. The best linear classification function resulted in 72% average classification. At its current stage, the system can be used by a radiologist to examine suspicious patterns in a mammogram. The regions which are flagged by the system have a 72% chance of developing a malignant mass by the time of the next screening. Therefore, further evaluation of these patients (e.g., a screening examination sooner than the usual one year interval) can result in earlier detection of breast cancer. A novel segmentation algorithm for mammogram partitioning based on fuzzy sets theory was also devised. This algorithm considers the fact that malignant masses and parenchymal patterns have unclear and fuzzy boundaries in a mammogram. It also takes into account the effects of neighboring pixels for this segmentation. This algorithm was evaluated in combination with a texture feature extraction step for detection of malignant masses in mammograms. The mass detection scheme resulted in 94.3% true-positive detection rate and 0.24 false-positives per image on a set of 35 mammograms.

Thermal coagulation therapy aims at treating tumors with high temperatures in the range 60-900C. A critical requirement for effective therapy is that the heat pattern from heating devices must be matched to the tumor's contours as accurately as possible. For this matter, new devices that generate heat patterns suited to treat tumors of irregular shapes are being developed. These devices, however, need to be calibrated and tested prior to their use in therapy. This thesis describes the development of a phantom material designed to characterize heating devices in terms of the spatial extent of thermal response they produce. The material records thermal patterns from heating devices by coagulating at high temperatures and can be imaged using magnetic resonance imaging (MRI) to depict these patterns after a heat treatment is delivered to the phantom. Experiments were conducted that characterize its MRI properties with temperature and demonstrate its ability to record thermal patterns from microwave heating devices. Calibrations of this sort can be used for quality assurance, comparison of devices, or for designing and testing new devices.

Monitoring Changes in Aortic Diameter by One-Dimensional (1-D) Magnetic Resonance Imaging (MRI) of Perpendicular Diameters

The thesis introduces a method, diameter sanning, for monitoring the cross-sectional size of blood vessels rapidly. Diameter scanning creates a one-dimensional (1-D) profile of a strip along the diameter of a vessel using magnetic resonance imaging (MRI). The strips in diameter scanning are chosen to be much wider than pixels in a typical two-dimensional (2D) image to increase the signal-to-noise ratio. A second strip perpendicular to the first and imaged sequentially allows the detection, estimation, and correction of errors in diameter measurements resulting from a difference between the prescribed centre and the true vessel centre. In both phantom and volunteer studies, the diameter measurements derived from 1-D profiles agree with measurements derived from 2-D images. Diameter scanning is nearly an order of magnitude faster than 2-D MRI and has the potential for real-time implementation.

Tissue autofluorescence is investigated as a possible method for detection of early cancer during endoscopy of the gastrointestinal (GI) tract. In the present work, in vivo fluorescence point-spectroscopy of colonic tissues, with blue light excitation, showed significant differences in autofluorescence intensity and spectral line shape between normal, preneoplastic and neoplastic lesions. These data formed the basis for developing imaging algorithms currently used in a prototype real-time light-induced fluorescence endoscope (LIFE), undergoing multicenter clinical trials. Clinical results of fluorescence imaging and spectroscopy are described. The mechanisms by which LIFE imaging and spectroscopy differentiate normal tissues from early and late stage colonic cancers are not well understood.

In this thesis, human colonic tissues were analyzed by confocal fluorescence microscopy, microspectrofluorimetry, histological and immunohistological staining, and transmission electron microscopy. Together with studies of the attenuation of light in gastrointestinal tissues, the measurements were used to develop a quantitative model of the autofluorescence process in vivo. Excitation and emission matrices (EEMs) of tissues and potential fluorophores were used to identify fluorophores contributing to this autofluorescence. The cellular and extracellular origins and changes in fluorophore distributions in normal, preneoplastic and neoplastic tissues are described. Modeling showed that LIFE-GI imaging and spectroscopy results in vivo may be explained by a combination of many factors, including the effect of different illumination/detection geometries, changes in tissue architecture, changes in tissue optical properties, and altered fluorophore distributions, as normal colon progresses toward neoplasia.

Furthermore, a more detailed study of mucosal tissue autofluorescence from normal colon, hyperplastic polyps, and adenomatous polyps is presented. In these tissues, significant differences in mucosal fluorophore composition were found, including the discovery of a possible endogenous autofluorescent biomarker specific to colonic dysplasia. Finally, a summary of results and conclusions are presented, and future studies are outlined.

Low magnitude loading on the body has become an important issue with the occurrence of injuries that have been attributed to repetitive loading of tissue at magnitudes of forces and motions that are below the known maximum strength of the tissues in question. The general purpose of this thesis was to explore the in vivo low level static and dynamic loading on the spine and determine if these loading parameters could generate injuries when similar low magnitude parameters were tested in vitro. In vivo activities (walking, sitting, standing, and back extensor exercises) were examined to quantify the magnitude of low back joint loads, motions, and muscular activations levels required. In vitro highly repetitive loading was performed at modest flexion/extension moments or angular rotations combined with low magnitude axial compressive forces. Walking was found to be a highly dynamic/cyclic activity with moderate spine loads and small lumbar spine angular motions. The back extensor exercises produced a range of joint loads from low loads for exercises such as single leg extension to loads exceeding spinal compression limits for contraindicated trunk extension exercises. Sitting and standing both resulted in low magnitude joint forces and muscular activity levels. Standing exhibited very small and static ranges of spine postures whereas sitting resulted in a range of postures from 30% to 80% of the lumbar spines flexion range of motion. The in vitro highly repetitive testing of porcine cervical spine motion segments at low magnitude compressive loads and modest flexion/extension motions and moments resulted in intervertebral disc herniations. The angular stiffness of specimens increased throughout the testing cycle and increased magnitudes of axial compressive loads resulted in increased probability and severity of disc damage. This research documents the magnitude of lumbar spine joint forces, spinal motions, and muscular activation levels during common in vivo activities and demonstrates that low level repetitive loading scenarios can result in spinal injuries.

In-Vivo Short Echo Hydrogen Spectroscopy: Precise Quantification And Application To Mental Illness (Cerebral Metabolites)

Clinical investigation into the causes of mental illness would benefit from the precise determination of in-vivo cerebral metabolite levels. Short echo 1H magnetic resonance spectroscopy may be used to measure in-vivo levels of twelve metabolites. However, precise quantification of these spectra remains problematic at 1.5 Tesla (T) due to low signal to noise ratio, low resolution, and a variable spectral baseline from uncharacterized macromolecule resonances. The object of this work was to increase the measurement precision associated with in-vivo short echo 1H spectroscopy. New software was developed for the non-interactive quantification of spectra in the time domain. Using this software, an approach to spectral quantification was developed at 1.5 T that optimized measurement precision. This optimized technique was applied in clinical studies of schizophrenia and obsessive-compulsive disorder at 1.5 T. The effect of field strength on quantification precision was also determined using data acquired at 1.5 T and 4.0 T. The modeling of macromolecule resonances, the fitting of resonances adjacent to the spectral region of interest, and the elimination of time domain filtering, were all components of a quantification strategy that lead to increased in-vivo measurement precision. The acquisition of data at 4.0 T also increased measurement precision due to the increased signal to noise ratio and chemical shift dispersion compared to 1.5 T. The quantification strategy developed was used to obtain reliable measures of N-acetylaspartate, glutamate, glutamine, phosphocreatine and creatine, and choline containing compounds from 4.5 cm 3 voxels at 1.5 T and from 1.5 cm3 voxels at 4.0 T. The reduction of voxel volume to 1.5 cm3 at 4.0 T significantly reduced partial volume sampling. Clinical studies at 1.5 T using this technique found increased levels of glutamine in the left medial prefrontal cortex of never treated schizophrenic patients compared to controls, and decreased levels of N-acetylaspartate in the left corpus striatum of obsessive compulsive patients compared to controls. These findings illustrate that short echo 1H magnetic resonance spectroscopy is a valuable technique for clinical research. Future improvements in measurement precision, especially at higher field strengths, will enable the detection of subtle metabolite level differences in small patient groups

Coronary Circulatory Reserve In Normotensive Hyperdynamic Sepsis (Myocardial, Oxygen Demand)

Sepsis is a syndrome describing a malignant inflammatory response to an infection leading to an alteration of the cardiovascular system including myocardial depression. loss of vasomotor control, and microcirculatory dysfunction. Most studies concluded that myocardial ischemia is not the cause of myocardial depression in sepsis since coronary blood flow is elevated in this syndrome. However, these studies did not take into account the important feature of the coronary circulation which has to adjust quickly to changes in myocardial O₂ demand. This feature is supported by the coronary circulatory reserve consisting of a blood flow and an O₂ extraction (O ₂E) reserve. In the first experiment, a sham group and sheep rendered septic by cecal ligation and perforation underwent a hypoxia trial to challenge and exhaust the coronary circulation. The conclusion of this study was that sepsis did not completely abolish but significantly depressed both components of coronary circulatory reserve. As it is suspected that sepsis left-shifts the O₂ dissociation curve (ODC) and. thereby, depresses hemoglobin O₂ unloading, a retrospective reconstruction of ODCs was undertaken and their effect on the systemic and coronary circulation was analyzed. The p50 was significant lower in septic sheep than in sham sheep. The left-shift of the ODC in CLP sheep was associated with a depression in maximum systemic O₂E while this was not seen in the coronary circulation. This is most likely due to the convergence of different ODCs at lower pO₂s where O₂E happens in the coronary circulation. To further analyze the effects of the blood's O₂ carrying capacity on the coronary circulatory reserve. a study was designed to compare septic sheep with two different hemoglobin levels (70 g/l by isovolemic hemodilution, 120 g/l by blood transfusion) during the hypoxia trial. Blood transfusion was not able to restore coronary circulatory reserve. However. hemodilution imposed significant alterations on coronary blood flow including a maldistribution of intramyocardial blood flow suggesting that maintaining a high hemoglobin in sepsis is beneficial in terms of the coronary circulation

An algorithm was developed in order to reduce operator dependence in ultrasound-guided breast biopsy, by automatically locating the needle in the ultrasound image, and displaying its location on the image for the user. Ultrasound images of a typical breast biopsy needle inserted in a tissue-mimicking agar were obtained to test the algorithm. The resulting images were examined by a group of observers who recorded the values of the angle, intercept and tip coordinates of the needle in the image, and inter- and intra-observer variability studies were performed on the results. The results of the algorithm segmentation were compared to the values recorded by the observers, and physical measurements recorded at the time the images were acquired. Acceptable limits on the deviation between the algorithm and observer results were established. The algorithm success rate was as high as 67% when the insertion distance of the needle was greater than 10 mm.

Computed Rotational Angiography: Use Of C-Arm-Mounted XRII For 3D Imaging Of Intracranial Vessels During Neuro-Interventional Procedures

Recent advances in endovascular therapy for the treatment of cerebrovascular diseases have been made possible mainly through the development of interventional devices and of new techniques for their deployment. Generally, image guidance during procedures such as the treatment of aneurysms using platinum coils is provided by two-dimensional projection images—real-time x-ray fluoroscopy and digital subtraction angiography (DSA). However, quantitative three-dimensional (3D) images could provide benefits for diagnosis, treatment planning and therapy by providing additional information to the interventionalist. We have developed a system capable of providing 3D volume reconstructions during interventional procedures, while preserving the real-time low-dose capability and flexible patient access of the standard DSA system, and called it Computed Rotational Angiography (CRA). We produced reconstructions using a clinical C-arm mounted X-ray image intensifier (XRII), acquiring 130 projections during an intra-arterial injection of contrast (gantry rotation covered 200° in 4.5 s). Projections were first corrected for angle-dependent distortion in the XRII to within ±0.1 pixels (global bi-polynomial warp algorithm) and for non-idealities in gantry motion to within ±0.12 pixels (simple shift algorithm). Three-dimensional volume images were then reconstructed from the corrected projections using a modified conebeam backprojection algorithm. Vessel signal-difference-to-noise ratio (SNR) in these volumes is limited by artifact from view aliasing, time-varying opacification of vessels and the presence of dense platinum coils. We investigated the relative magnitudes of these artifact sources using a rigid in vitro flow-through model of the intracranial vasculature (vessel diameters known to within ±0.15 mm; vessel flow within 16% on average of flow specified in the literature) and using an in vivo porcine model including a surgically-created side-wall aneurysm. The SNR was measured as a function of contrast injection rate, blood-flow rate, contrast dilution, and selective injection site. An injection rate of 3 ml/s provided SNR 25 for all combinations of parameters investigated. Vessel diameters were within ±6% of the specified phantom values. Platinum coils could be segmented from contrast-filled aneurysms. We have shown that CRA volumes provide sufficient image quality to characterize vessel, aneurysm, coil and residual neck geometries during interventional procedures

Design And Performance Of A Quadrature Elliptic Birdcage Resonator For Magnetic Resonance Imaging

Functional magnetic resonance imaging (fMRI) of the human brain exploits the blood-oxygenation-level-dependent (BOLD) contrast of magnetic resonance images to identify areas within the cortex that respond to a presented stimulus or task. At submillimeter spatial resolution, fMRI becomes quite difficult, even at a magnetic field strength of 4 Tesla, since the signal-to-noise ratio (SNR) within image voxels is significantly reduced. This makes reliable detection of a BOLD response a forbidding task. The object of this thesis was to provide methods to (a) improve the signal-to-noise ratio for high spatial resolution functional magnetic resonance images and to (b) appropriately prescribe image orientations and locations to optimize the BOLD response within selected regions of interest. These methods included the construction and implementation of a radio frequency (RF) surface coil to improve SNR and RF homogeneity throughout the image and a functional scout imaging technique that uses receiver phase cycling to create functional maps as the scanner collects the image data. Using the newly constructed coil and a modified echo planar imaging sequence and image reconstruction scheme, fMRI studies at submillimeter spatial resolution were performed to identify the ocular dominance columns within the human primary visual cortex (V1) of healthy individuals with normal or corrected-to-normal vision. Area V1 was identified using a fMRI experiment that demonstrated a correlation between the magnitude of the BOLD response within VI and the stimulus contrast. This correlation was not detected in higher order visual areas. Functional MRI studies of V1 of individuals who have unilateral amblyopia were performed to identify the neuronal correlate of the reduction of contrast sensitivity of the amblyopic eye. The pooled fMRI response to monocular stimulation of the amblyopic eye was significantly reduced, and reflected the decrease in contrast sensitivity. High resolution fMRI revealed that the ocular dominance columns of the amblyopic eye were significantly reduced in size for individuals who developed amblyopia during infancy. The results of this work demonstrate a fMRI technique that can reliably resolve functional units of the cortex on a submillimeter scale. In addition, this technique can be used to investigate brain plasticity at the cortical columnar level resulting from developmental visual disorders or trauma.

A new imaging technique called coherent-scatter computed tomography (CSCT) is being developed in the Imaging Research Laboratories at the John P. Robarts Research Institute to measure calcified tissues, such as bone-mineral content distribution (BMCD) for diagnosing osteoporosis. A proof-of-concept CSCT imaging system has been developed previously. However, the image quality and required acquisition time were insufficient. Thus, development of a practical CSCT system was necessary. This thesis describes the first practical CSCT system based on diagnostic x-ray equipment. The following tasks were fulfilled: (1) developing a new real-time (30 frames per second, 640 x 480 pixels per frame and 8-bit per pixel) image acquisition system based on a Silicon Graphics $O/sb2$ workstation; (2) minimising noise in the new image acquisition system; (3) developing an x-ray generator controller; (4) developing low-noise digital x-ray intensity sensors; and (5) developing specimen rotate and translate motions that operate in synchrony with the video acquisition. During testing of the image acquisition system, noise-power aliasing inherent to the current interline video system was discovered. A theoretical description of the image noise-power spectrum (NPS) for interline and non-interline video formats was conducted using a linear system approach. It was also the first time that the concept of cyclo-stationary random process was applied to medical imaging. (Abstract shortened by UMI.)

The aneurysm wall is composed of layers of collagen fibres. Wall strength is related to both the strength of the fibres and their orientation. When the aneurysm enlarges, the amount and organization of the collagen fibres change, potentially leading to areas where rupture can occur. Intact aneurysms were obtained at autopsy, fixed at physiological pressure, sectioned and stained with picrosirius red. Using a polarized light microscope, I determined the directional organization of the collagen fibres across the layers of the walls of four aneurysms. The level of retardation and thus mechanical strength of the fibres was also measured. By combining both sets of data I predicted tensile strength as a function of direction on the aneurysmal wall. Comparing the weakest to the strongest direction, the breaking strength varies by a factor of up to 2x, implying a significant degree of mechanical anisotropy

Flexor tendon injuries are common and frequently lead to disability. This thesis is composed of four studies that investigate a new surgical technique, the Savage technique of flexor tendon repair. In the first study, the in-vitro static strength of this technique was found to be superior to conventional repairs. In the second study, the cyclic fatigue strength of the Savage repair was found to be sufficient to withstand the forces of active motion, without repair rupture or gap formation. In the third study, the gliding properties of the Savage repair were compared to the optimal conventional technique, and found to be equivalent. In the fourth study, an active motion rehabilitation protocol was simulated, and it was found that tendon load and work of flexion were minimized with the wrist in a position of extension. Based upon these studies, active motion of flexor tendon injuries repaired using the Savage technique offers promise as a clinical tool.

Send comments regarding this page to .

GreekMEDICAL PHYSICS THESES 1999