PARALLELISATION OF THE MODEL-BASED ITERATIVE RECONSTRUCTION ALGORITHM DIRA.

New paradigms for parallel programming have been devised to simplify software development on multi-core processors and many-core graphical processing units (GPU). Despite their obvious benefits, the parallelisation of existing computer programs is not an easy task. In this work, the use of the Open Multiprocessing (OpenMP) and Open Computing Language (OpenCL) frameworks is considered for the parallelisation of the model-based iterative reconstruction algorithm DIRA with the aim to significantly shorten the code's execution time. Selected routines were parallelised using OpenMP and OpenCL libraries; some routines were converted from MATLAB to C and optimised. Parallelisation of the code with the OpenMP was easy and resulted in an overall speedup of 15 on a 16-core computer. Parallelisation with OpenCL was more difficult owing to differences between the central processing unit and GPU architectures. The resulting speedup was substantially lower than the theoretical peak performance of the GPU; the cause was explained.

AUTOMATIC SEGMENTATION OF PELVIS FOR BRACHYTHERAPY OF PROSTATE.

Advanced model-based iterative reconstruction algorithms in quantitative computed tomography (CT) perform automatic segmentation of tissues to estimate material properties of the imaged object. Compared with conventional methods, these algorithms may improve quality of reconstructed images and accuracy of radiation treatment planning. Automatic segmentation of tissues is, however, a difficult task. The aim of this work was to develop and evaluate an algorithm that automatically segments tissues in CT images of the male pelvis. The newly developed algorithm (MK2014) combines histogram matching, thresholding, region growing, deformable model and atlas-based registration techniques for the segmentation of bones, adipose tissue, prostate and muscles in CT images. Visual inspection of segmented images showed that the algorithm performed well for the five analysed images. The tissues were identified and outlined with accuracy sufficient for the dual-energy iterative reconstruction algorithm whose aim is to improve the accuracy of radiation treatment planning in brachytherapy of the prostate.

Breast dosimetry using high-resolution voxel phantoms.

A computer model of X-ray mammography has been developed, which uses quasi-realistic high-resolution voxel phantoms to simulate the breast. The phantoms have 400 microm voxels and simulate the three-dimensional distributions of adipose and fibro-glandular tissues, Cooper's ligaments, ducts and skin and allow the estimation of dose to individual tissues. Calculations of the incident air kerma to mean glandular dose conversion factor, g, were made using a Mo/Mo spectrum at 28 kV for eight phantoms in the thickness range 40-80 mm and of varying glandularity. The values differed from standard tabulations used for breast dosimetry by up to 43%, because of the different spatial distribution of glandular tissue within the breast. To study this further, additional voxel phantoms were constructed, which gave variations of between 9 and 59% compared with standard values. For accurate breast dosimetry, it is therefore very important to take the distribution of glandular tissues into account.

Calculation of the properties of digital mammograms using a computer simulation.

A Monte Carlo computer model of mammography has been developed to study and optimise the performance of digital mammographic systems. The program uses high-resolution voxel phantoms to model the breast, which simulate the adipose and fibroglandular tissues, Cooper's ligaments, ducts and skin in three dimensions. The model calculates the dose to each tissue, and also the quantities such as energy imparted to image pixels, noise per image pixel and scatter-to-primary (S/P) ratios. It allows studies of the dependence of image properties on breast structure and on position within the image. The program has been calibrated by calculating and measuring the pixel values and noise for a digital mammographic system. The thicknesses of two components of this system were unknown, and were adjusted to obtain a good agreement between measurement and calculation. The utility of the program is demonstrated with the calculations of the variation of the S/P ratio with and without a grid, and of the image contrast across the image of a 50-mm-thick breast phantom.

Monte Carlo simulation of a mammographic test phantom.

A test phantom, including a wide range of mammographic tissue equivalent materials and test details, was imaged on a digital mammographic system. In order to quantify the effect of scatter on the contrast obtained for the test details, calculations of the scatter-to-primary ratio (S/P) have been made using a Monte Carlo simulation of the digital mammographic imaging chain, grid and test phantom. The results show that the S/P values corresponding to the imaging conditions used were in the range 0.084-0.126. Calculated and measured pixel values in different regions of the image were compared as a validation of the model and showed excellent agreement. The results indicate the potential of Monte Carlo methods in the image quality-patient dose process optimisation, especially in the assessment of imaging conditions not available on standard mammographic units.

RADIUS--closing the circle on the assessment of imaging performance.

The RADIUS (Radiological Imaging Unification Strategy) project addresses the assessment of image quality in terms of both physical and clinically relevant measures. The aim is to unify our understanding of both types of measure as well as the numerous underlying factors that play a key role in the assessments of imaging performance. In this way it is expected to provide a solid basis for the improvement in radiological safety management, where not only radiation risks are considered but also diagnostic risks of incorrect clinical outcomes (i.e. false positive/false negative). The project has applied a variety of relevant experimental and theoretical methods to this problem, which is generic to medical imaging as a whole. Digital radiography of the chest and the breast has been employed as the clinical imaging domain vehicles for the study. The project addressed the problem from the following directions: role and relevance of pathology, human observer studies including receiver operating characteristics, image quality criteria analysis, structural noise analysis, physical measurements on clinical images, physical measurements on imaging system, modelling of imaging system, modelling of visual processes, modelling of doses delivered and IT-based scientific support strategies. This paper presents an overview of the main outcomes from this project and highlights how the research outcomes actually apply to the real world. In particular, attention will be focused on new and original findings and methods and techniques that have been developed within the framework of the project. The relevance of the project's outcomes to future European research will also be presented.

Schemes for the optimization of chest radiography using a computer model of the patient and x-ray imaging system.

A computer program has been developed to model chest radiography. It incorporates a voxel phantom of an adult and includes antiscatter grid, radiographic screen, and film. Image quality is quantified by calculating the contrast (deltaOD) and the ideal observer signal-to-noise ratio (SNR(I)) for a number of relevant anatomical details at various positions in the anatomy. Detector noise and system unsharpness are modeled and their influence on image quality is considered. A measure of useful dynamic range is computed and defined as the fraction of the image that is reproduced at an optical density such that the film gradient exceeds a preset value. The effective dose is used as a measure of the radiation risk for the patient. A novel approach to patient dose and image quality optimization has been developed and implemented. It is based on a reference system acknowledged to yield acceptable image quality in a clinical trial. Two optimizations schemes have been studied, the first including the contrast of vessels as measure of image quality and the second scheme using also the signal-to-noise ratio of calcifications. Both schemes make use of our measure of useful dynamic range as a key quantity. A large variety of imaging conditions was simulated by varying the tube voltage, antiscatter device, screen-film system, and maximum optical density in the computed image. It was found that the optical density is crucial in screen-film chest radiography. Significant dose savings (30%-50%) can be accomplished without sacrificing image quality by using low-atomic-number grids with a low grid ratio or an air gap and more sensitive screen-film system. Dose-efficient configurations proposed by the model agree well with the example of good radiographic technique suggested by the European Commission.

Energy imparted to the patient in diagnostic radiology: calculation of conversion factors for determining the energy imparted from measurements of the air collision kerma integrated over beam area.

The energy imparted to the patient in diagnostic radiology, related to radiation risk in examinations of the trunk and head, can be deduced from a measurement of the air collision kerma (or exposure) of the incident primary photons integrated over beam area by using a thin, flat ionisation chamber covering the entire roentgen beam. Factors for converting the integral of the air collision kerma to energy imparted to the patient have been calculated using a Monte Carlo method. The patient is simulated by laterally infinite water slabs with thicknesses from 100-300 mm. Calculations are performed for monoenergetic photons (5-300 keV) and energy spectra commonly used in diagnostic radiology (40-130 kV acceleration potential differences and values of the half-value thickness of air collision kerma in aluminium from 0.9 to 9.9 mm). Correction factors which take into account the additional escape of scattered photons from the sides of a laterally finite water slab as a function of field size and focal distance are also given.

A study and optimization of lumbar spine X-ray imaging systems.

A Monte Carlo program has been developed that incorporates a voxel phantom of an adult patient in a model of the complete X-ray imaging system, including the anti-scatter grid and screen-film receptor. This allows the realistic estimation of patient dose and the corresponding image (optical density map) for a wide range of equipment configurations. This paper focuses on the application of the program to lumbar spine anteroposterior and lateral screen-film examinations. The program has been applied to study the variation of physical image quality measures and effective dose for changing system parameters such as tube voltage, grid design and screen-film system speed. These variations form the basis for optimization of these system parameters. In our approach to optimization, the best systems are those that can match (or come close to) the calculated image quality measure of systems preferred in a recent European clinical trial, but with lower patient dose. The largest dose savings found were 21% for a 400 speed class system with a grid having a strip density of 40 cm(-1) and a grid ratio of 16. A further dose saving of 13% was possible when a 600 speed class system was employed. The best systems found from the optimization correspond to those recommended by the European Commission guidelines on image quality criteria for diagnostic radiographic images.

Demonstration of correlations between clinical and physical image quality measures in chest and lumbar spine screen-film radiography.

The ability to predict clinical image quality from physical measures is useful for optimization in diagnostic radiology. In this work, clinical and physical assessments of image quality are compared and correlations between the two are derived. Clinical assessment has been made by a group of expert radiologists who evaluated fulfillment of the European image criteria for chest and lumbar spine radiography using two scoring methods: image criteria score (ICS) and visual grading analysis score (VGAS). Physical image quality measures were calculated using a Monte Carlo simulation model of the complete imaging system. This model includes a voxelized male anatomy and was used to calculate contrast and signal-to-noise ratio of various important anatomical details and measures of dynamic range. Correlations between the physical image quality measures on the one hand and the ICS and VGAS on the other were sought. 16 chest and 4 lumbar spine imaging system configurations were compared in frontal projection. A statistically significant correlation with clinical image quality was found in chest posteroanterior radiography for the contrast of blood vessels in the retrocardiac area and a measure of useful dynamic range. In lumbar spine anteroposterior radiography, a similar significant correlation with clinical image quality was found between the contrast and signal-to-noise ratio of the trabecular structures in the L1-L5 vertebrae. The significant correlation shows that clinical image quality can, at least in some cases, be predicted from appropriate measures of physical image quality.

Influence of the characteristic curve on the clinical image quality of lumbar spine and chest radiographs.

The "European Guidelines on Quality Criteria for Diagnostic Radiographic Images" do not address the choice of the film characteristic (H&D) curve, which is an important parameter for the description of a radiographic screen-film system. The image contrast of clinical lumbar spine and chest radiographs was altered by digital image processing techniques, simulating images with different H&D curves, both steeper and flatter than the original. The manipulated images were printed on film for evaluation. Seven experienced radiologists evaluated the clinical image quality by analysing the fulfilment of the European Image Criteria (ICS) and by visual grading analysis (VGA) of in total 224 lumbar spine and 360 chest images. A parallel study of the effect of the H&D curve has also been made using a theoretical model. The contrast (DeltaOD) of relevant anatomical details was calculated, using a Monte Carlo simulation-model of the complete imaging system including a 3D voxel phantom of a patient. Correlations between the calculated contrast and the radiologists' assessment by VGA were sought. The results of the radiologists' assessment show that the quality in selected regions of lumbar spine and chest images can be significantly improved by the use of films with a steeper H&D curve compared with the standard latitude film. Significant (p<0.05) correlations were found between the VGA results and the calculations of the contrast of transverse processes and trabecular details in the lumbar spine vertebrae, and with the contrast of blood vessels in the retrocardiac area of the chest.

Absorbed dose in AgBr in direct film for photon energies ( < 150 keV): relation to optical density. Theoretical calculation and experimental evaluation.

In the radiological process it is necessary to develop tools so as to explore how X-rays can be used in the most effective way. Evaluation of models to derive measures of image quality and risk-related parameters is one possibility of getting such a tool. Modelling the image receptor, an important part of the imaging chain, is then required. The aim of this work was to find convenient and accurate ways of describing the blackening of direct dental films by X-rays. Since the beginning of the 20th century, the relation between optical density and photon interactions in the silver bromide in X-ray films has been investigated by many authors. The first attempts used simple quantum theories with no consideration of underlying physical interaction processes. The theories were gradually made more realistic by the introduction of dosimetric concepts and cavity theory. A review of cavity theories for calculating the mean absorbed dose in the AgBr grains of the film emulsion is given in this work. The cavity theories of GREENING (15) and SPIERS-CHARLTON (37) were selected for calculating the mean absorbed dose in the AgBr grains relative to the air collision kerma (Kc,air) of the incident photons of Ultra-speed and Ektaspeed (intraoral) films using up-to-date values of interaction coefficients. GREENING'S theory is a multi-grain theory and the results depend on the relative amounts of silver bromide and gelatine in the emulsion layer. In the single grain theory of SPIERS-CHARLTON, the shape and size of the silver bromide grain are important. Calculations of absorbed dose in the silver bromide were compared with measurements of optical densities in Ultra-speed and Ektaspeed films for a broad range (25-145 kV) of X-ray energy. The calculated absorbed dose values were appropriately averaged over the complete photon energy spectrum, which was determined experimentally using a Compton spectrometer. For the whole range of tube potentials used, the measured optical densities of the films were found to be proportional to the mean absorbed dose in the AgBr grains calculated according to GREENING'S theory. They were also found to be proportional to the collision kerma in silver bromide (Kc,AgBr) indicating proportionality between Kc,AgBr and the mean absorbed dose in silver bromide. While GREENING'S theory shows that the quotient of the mean absorbed dose in silver bromide and Kc,AgBr varies with photon energy, this is not apparent when averaged over the broad (diagnostic) X-ray energy spectra used here. Alternatively, proportionality between Kc,AgBr and the mean absorbed dose in silver bromide can be interpreted as resulting from a combination of the SPIERS-CHARLTON theory, valid at low photon energies ( < 30 keV) and GREENING'S theory, which is strictly valid at energies above 50 keV. This study shows that the blackening of non-screen films can be related directly to the energy absorbed in the AgBr grains of the emulsion layer and that, for the purpose of modelling the imaging chain in intraoral radiography, film response can be represented by Kc,AgBr (at the position of the film) independent of photon energy. The importance of taking the complete X-ray energy spectrum into full account in deriving Kc,AgBr is clearly demonstrated, showing that the concept of effective energy must be used with care.