Table of contents

Radiation dose in x-ray computed tomography (CT) has become a topic of high interest due to the increasing numbers of CT examinations performed worldwide. This review aims to present an overview of current concepts for both scanner output metrics and for patient dosimetry and will comment on their strengths and weaknesses. Controversial issues such as the appropriateness of the CT dose index (CTDI) are discussed in detail. A review of approaches to patient dose assessment presently in practice, of the dose levels encountered and options for further dose optimization are also given and discussed. Patient dose assessment remains a topic for further improvement and for international consensus. All approaches presently in use are based on Monte Carlo (MC) simulations. Estimates for effective dose are established, but they are crude and not patient-specific; organ dose estimates are rarely available. Patient- and organ-specific dose estimates can be provided with adequate accuracy and independent of CTDI phantom measurements by fast MC simulations. Such information, in particular on 3D dose distributions, is important and helpful in optimization efforts. Dose optimization has been performed very successfully in recent years and even resulted in applications with effective dose values of below 1 mSv. In general, a trend towards lower dose values based on technical innovations has to be acknowledged. Effective dose values are down to clearly below 10 mSv on average, and there are a number of applications such as cardiac and pediatric CT which are performed routinely below 1 mSv on modern equipment.

Non-invasive evaluation of the Achilles tendon elastic properties may enhance diagnosis of tendon injury and the assessment of recovery treatments. Shear wave elastography has shown to be a powerful tool to estimate tissue mechanical properties. However, its applicability to quantitatively evaluate tendon stiffness is limited by the understanding of the physics on the shear wave propagation in such a complex medium. First, tendon tissue is transverse isotropic. Second, tendons are characterized by a marked stiffness in the 400 to 1300 kPa range (i.e. fast shear waves). Hence, the shear wavelengths are greater than the tendon thickness leading to guided wave propagation. Thus, to better understand shear wave propagation in tendons and consequently to properly estimate its mechanical properties, a dispersion analysis is required. In this study, shear wave velocity dispersion was measured in vivo in ten Achilles tendons parallel and perpendicular to the tendon fibre orientation. By modelling the tendon as a transverse isotropic viscoelastic plate immersed in fluid it was possible to fully describe the experimental data (deviation<1.4%). We show that parallel to fibres the shear wave velocity dispersion is not influenced by viscosity, while it is perpendicularly to fibres. Elasticity (found to be in the range from 473 to 1537 kPa) and viscosity (found to be in the range from 1.7 to 4 Pa.s) values were retrieved from the model in good agreement with reported results.

As dose optimization for high dose rate brachytherapy becomes more complex, it becomes increasingly important to have a means of verifying that optimization results are reasonable. A method is presented for using a simple optimization as quality assurance for the more complex optimization algorithms typically found in commercial brachytherapy treatment planning systems. Quality assurance tests may be performed during commissioning, at regular intervals, and/or on a patient specific basis. A simple optimization method is provided that optimizes conformal target coverage using an exact, variance-based, algebraic approach. Metrics such as dose volume histogram, conformality index, and total reference air kerma agree closely between simple and complex optimizations for breast, cervix, prostate, and planar applicators. The simple optimization is shown to be a sensitive measure for identifying failures in a commercial treatment planning system that are possibly due to operator error or weaknesses in planning system optimization algorithms. Results from the simple optimization are surprisingly similar to the results from a more complex, commercial optimization for several clinical applications. This suggests that there are only modest gains to be made from making brachytherapy optimization more complex. The improvements expected from sophisticated linear optimizations, such as PARETO methods, will largely be in making systems more user friendly and efficient, rather than in finding dramatically better source strength distributions.

A factorized system matrix utilizing an image domain resolution model is attractive in fully 3D time-of-flight PET image reconstruction using list-mode data. In this paper, we study a factored model based on sparse matrix factorization that is comprised primarily of a simplified geometrical projection matrix and an image blurring matrix. Beside the commonly-used Siddon's ray-tracer, we propose another more simplified geometrical projector based on the Bresenham's ray-tracer which further reduces the computational cost. We discuss in general how to obtain an image blurring matrix associated with a geometrical projector, and provide theoretical analysis that can be used to inspect the efficiency in model factorization. In simulation studies, we investigate the performance of the proposed sparse factorization model in terms of spatial resolution, noise properties and computational cost. The quantitative results reveal that the factorization model can be as efficient as a non-factored model, while its computational cost can be much lower. In addition we conduct Monte Carlo simulations to identify the conditions under which the image resolution model can become more efficient in terms of image contrast recovery. We verify our observations using the provided theoretical analysis. The result offers a general guide to achieve the optimal reconstruction performance based on a sparse factorization model with an image domain resolution model.

The aim of this study is the evaluation of on-the-fly volume of intersection computation for system's geometry modelling in 3D PET image reconstruction. For this purpose we propose a simple geometrical model in which the cubic image voxels on the given Cartesian grid are approximated with spheres and the rectangular tubes of response (ToRs) are approximated with cylinders. The model was integrated into a fully 3D list-mode PET reconstruction for performance evaluation. In our model the volume of intersection between a voxel and the ToR is only a function of the impact parameter (the distance between voxel centre to ToR axis) but is independent of the relative orientation of voxel and ToR. This substantially reduces the computational complexity of the system matrix calculation. Based on phantom measurements it was determined that adjusting the diameters of the spherical voxel size and the ToR in such a way that the actual voxel and ToR volumes are conserved leads to the best compromise between high spatial resolution, low noise, and suppression of Gibbs artefacts in the reconstructed images. Phantom as well as clinical datasets from two different PET systems (Siemens ECAT HR⁺ and Philips Ingenuity-TF PET/MR) were processed using the developed and the respective vendor-provided (line of intersection related) reconstruction algorithms. A comparison of the reconstructed images demonstrated very good performance of the new approach. The evaluation showed the respective vendor-provided reconstruction algorithms to possess 34–41% lower resolution compared to the developed one while exhibiting comparable noise levels. Contrary to explicit point spread function modelling our model has a simple straight-forward implementation and it should be easy to integrate into existing reconstruction software, making it competitive to other existing resolution recovery techniques.

Four dimensional cone beam computed tomography (4DCBCT) images suffer from angular under sampling and bunching of projections due to a lack of feedback between the respiratory signal and the acquisition system. To address this problem, respiratory motion guided 4DCBCT (RMG-4DCBCT) regulates the gantry velocity and projection time interval, in response to the patient's respiratory signal, with the aim of acquiring evenly spaced projections in a number of phase or displacement bins during the respiratory cycle. Our previous study of RMG-4DCBCT was limited to sinusoidal breathing traces. Here we expand on that work to provide a practical algorithm for the case of real patient breathing data. We give a complete description of RMG-4DCBCT including full details on how to implement the algorithms to determine when to move the gantry and when to acquire projections in response to the patient's respiratory signal. We simulate a realistic working RMG-4DCBCT system using 112 breathing traces from 24 lung cancer patients. Acquisition used phase-based binning and parameter settings typically used on commercial 4DCBCT systems (4 min acquisition time, 1200 projections across 10 respiratory bins), with the acceleration and velocity constraints of current generation linear accelerators. We quantified streaking artefacts and image noise for conventional and RMG-4DCBCT methods by reconstructing projection data selected from an oversampled set of Catphan phantom projections. RMG-4DCBCT allows us to optimally trade-off image quality, acquisition time and image dose. For example, for the same image quality and acquisition time as conventional 4DCBCT approximately half the imaging dose is needed. Alternatively, for the same imaging dose, the image quality as measured by the signal to noise ratio, is improved by 63% on average. C-arm cone beam computed tomography systems, with an acceleration up to 200°/s², a velocity up to 100°/s and the acquisition of 80 projections per second, allow the image acquisition time to be reduced to below 60 s. We have made considerable progress towards realizing a system to reduce projection clustering in conventional 4DCBCT imaging and hence reduce the imaging dose to the patient.

This study reports on the relative precision, relative error, and dose differences observed when using a new full-image calibration technique in NIPAM-based x-ray CT polymer gel dosimetry. The effects of calibration parameters (e.g. gradient thresholding, dose bin size, calibration fit function, and spatial remeshing) on subsequent errors in calibrated gel images are reported. It is found that gradient thresholding, dose bin size, and fit function all play a primary role in affecting errors in calibrated images. Spatial remeshing induces minimal reductions or increases in errors in calibrated images. This study also reports on a full error propagation throughout the CT gel image pre-processing and calibration procedure thus giving, for the first time, a realistic view of the errors incurred in calibrated CT polymer gel dosimetry. While the work is based on CT polymer gel dosimetry, the formalism is valid for and easily extended to MRI or optical CT dosimetry protocols. Hence, the procedures developed within the work are generally applicable to calibration of polymer gel dosimeters.

Cone-beam computed tomography (CBCT) is an important online imaging modality for image guided radiotherapy. But suboptimal image quality and the lack of a real-time stereoscopic imaging function limit its implementation in advanced treatment techniques, such as online adaptive and 4D radiotherapy. Tetrahedron beam computed tomography (TBCT) is a novel online imaging modality designed to improve on the image quality provided by CBCT. TBCT geometry is flexible, and multiple detector and source arrays can be used for different applications. In this paper, we describe a novel dual source–dual detector TBCT system that is specially designed for LINAC radiation treatment machines. The imaging system is positioned in-line with the MV beam and is composed of two linear array x-ray sources mounted aside the electrical portal imaging device and two linear arrays of x-ray detectors mounted below the machine head. The detector and x-ray source arrays are orthogonal to each other, and each pair of source and detector arrays forms a tetrahedral volume. Four planer images can be obtained from different view angles at each gantry position at a frame rate as high as 20 frames per second. The overlapped regions provide a stereoscopic field of view of approximately 10–15 cm. With a half gantry rotation, a volumetric CT image can be reconstructed having a 45 cm field of view. Due to the scatter rejecting design of the TBCT geometry, the system can potentially produce high quality 2D and 3D images with less radiation exposure. The design of the dual source–dual detector system is described, and preliminary results of studies performed on numerical phantoms and simulated patient data are presented.

It was recently shown that high magnetic fields evoke nystagmus in human subjects with functioning vestibular systems. The proposed mechanism involves interaction between ionic currents in the endolymph of the vestibular labyrinth and the static magnetic field. This results in a Lorentz force that causes endolymph flow to deflect the cupulae of the semi-circular canals to evoke a vestibular–ocular reflex (VOR). This should be analogous to stimulation by angular acceleration or caloric irrigation. We made measurements of nystagmus slow-phase velocities in healthy adults experiencing variable magnetic field profiles of up to 7 T while supine on a bed that could be moved smoothly into the bore of an MRI machine. The horizontal slow-phase velocity data were reliably modelled by a linear transfer function incorporating a low-pass term and a high-pass adaptation term. The adaptation time constant was estimated at 39.3 s from long exposure trials. When constrained to this value, the low-pass time constant was estimated at 13.6 ± 3.6 s (to 95% confidence) from both short and long exposure trials. This confidence interval overlaps with values obtained previously using angular acceleration and caloric stimulation. Hence it is compatible with endolymph flow causing a cupular deflection and therefore supports the hypothesis that the Lorentz force is a likely transduction mechanism of the magnetic field-evoked VOR.

We designed and constructed an in vivo dosimetry system using plastic scintillation detectors (PSDs) to monitor dose to the rectal wall in patients undergoing intensity-modulated radiation therapy for prostate cancer. Five patients were enrolled in an Institutional Review Board-approved protocol for twice weekly in vivo dose monitoring with our system, resulting in a total of 142 in vivo dose measurements. PSDs were attached to the surface of endorectal balloons used for prostate immobilization to place the PSDs in contact with the rectal wall. Absorbed dose was measured in real time and the total measured dose was compared with the dose calculated by the treatment planning system on the daily computed tomographic image dataset. The mean difference between measured and calculated doses for the entire patient population was −0.4% (standard deviation 2.8%). The mean difference between daily measured and calculated doses for each patient ranged from −3.3% to 3.3% (standard deviation ranged from 5.6% to 7.1% for four patients and was 14.0% for the last, for whom optimal positioning of the detector was difficult owing to the patient's large size). Patients tolerated the detectors well and the treatment workflow was not compromised. Overall, PSDs performed well as in vivo dosimeters, providing excellent accuracy, real-time measurement and reusability.

The LabPET is an avalanche photodiode (APD) based digital PET scanner with quasi-individual detector read-out and highly parallel electronic architecture for high-performance in vivo molecular imaging of small animals. The scanner is based on LYSO and LGSO scintillation crystals (2×2×12/14 mm³), assembled side-by-side in phoswich pairs read out by an APD. High spatial resolution is achieved through the individual and independent read-out of an individual APD detector for recording impinging annihilation photons. The LabPET exists in three versions, LabPET4 (3.75 cm axial length), LabPET8 (7.5 cm axial length) and LabPET12 (11.4 cm axial length). This paper focuses on the systematic characterization of the three LabPET versions using two different energy window settings to implement a high-efficiency mode (250–650 keV) and a high-resolution mode (350–650 keV) in the most suitable operating conditions. Prior to measurements, a global timing alignment of the scanners and optimization of the APD operating bias have been carried out. Characteristics such as spatial resolution, absolute sensitivity, count rate performance and image quality have been thoroughly investigated following the NEMA NU 4-2008 protocol. Phantom and small animal images were acquired to assess the scanners' suitability for the most demanding imaging tasks in preclinical biomedical research. The three systems achieve the same radial FBP spatial resolution at 5 mm from the field-of-view center: 1.65/3.40 mm (FWHM/FWTM) for an energy threshold of 250 keV and 1.51/2.97 mm for an energy threshold of 350 keV. The absolute sensitivity for an energy window of 250–650 keV is 1.4%/2.6%/4.3% for LabPET4/8/12, respectively. The best count rate performance peaking at 362 kcps is achieved by the LabPET12 with an energy window of 250–650 keV and a mouse phantom (2.5 cm diameter) at an activity of 2.4 MBq ml⁻¹. With the same phantom, the scatter fraction for all scanners is about 17% for an energy threshold of 250 keV and 10% for an energy threshold of 350 keV. The results obtained with two energy window settings confirm the relevance of high-efficiency and high-resolution operating modes to take full advantage of the imaging capabilities of the LabPET scanners for molecular imaging applications.

In SPECT, the collimator is a crucial element in controlling image quality. We take a task performance approach to collimator performance evaluation in which an ideal observer is applied to the raw camera data without regard to the subsequent reconstruction stage. The clinical context of our collimator study is one of searching for and detecting neuroendocrine tumor metastases in the liver as seen in In-111 Octreotide SPECT. Our task involves detection and localization of a signal and thus differs from the conventionally used detection-only task. The scalar task performance metric is ALROC, the area under the localization receiver operating characteristic curve. Since In-111 emits photons at both 171 and 245 keV, the higher energy emissions can contribute significant septal scatter and penetration. Our collimator evaluations address a question previously considered by Mähler et al (2012 IEEE Trans. Nucl. Sci.59 47–53) who used a different methodology: does allowing a limited amount of septal scatter and penetration yield improved task performance? We used simulation methods to evaluate five parallel-hole collimators. The collimators had roughly equal geometric sensitivity and resolution but a range of contributions from septal effects leading to variations in total sensitivity and resolution. We found that the best performance was obtained with a collimator that allowed a moderate amount of septal scatter and penetration.

The aim of this study is to investigate the impact of respiratory motion correction and spatial resolution on lesion detectability in PET as a function of lesion size and tracer uptake. Real respiratory signals describing different breathing types are combined with a motion model formed from real dynamic MR data to simulate multiple dynamic PET datasets acquired from a continuously moving subject. Lung and liver lesions were simulated with diameters ranging from 6 to 12 mm and lesion to background ratio ranging from 3:1 to 6:1. Projection data for 6 and 3 mm PET scanner resolution were generated using analytic simulations and reconstructed without and with motion correction. Motion correction was achieved using motion compensated image reconstruction. The detectability performance was quantified by a receiver operating characteristic (ROC) analysis obtained using a channelized Hotelling observer and the area under the ROC curve (AUC) was calculated as the figure of merit. The results indicate that respiratory motion limits the detectability of lung and liver lesions, depending on the variation of the breathing cycle length and amplitude. Patients with large quiescent periods had a greater AUC than patients with regular breathing cycles and patients with long-term variability in respiratory cycle or higher motion amplitude. In addition, small (less than 10 mm diameter) or low contrast (3:1) lesions showed the greatest improvement in AUC as a result of applying motion correction. In particular, after applying motion correction the AUC is improved by up to 42% with current PET resolution (i.e. 6 mm) and up to 51% for higher PET resolution (i.e. 3 mm). Finally, the benefit of increasing the scanner resolution is small unless motion correction is applied. This investigation indicates high impact of respiratory motion correction on lesion detectability in PET and highlights the importance of motion correction in order to benefit from the increased resolution of future PET scanners.

The Albira Trimodal pre-clinical scanner comprises PET, SPECT and CT sub-systems and thus provides a range of pre-clinical imaging options. The PET component consists of three rings of single-crystal LYSO detectors with axial/transverse fields-of-view (FOVs) of 148/80 mm. The SPECT component has two opposing CsI detectors (100 × 100 mm²) with single-pinhole (SPH) or multi(9)-pinhole (MPH) collimators; the detectors rotate in 6° increments and their spacing can be adjusted to provide different FOVs (25 to 120 mm). The CT sub-system provides 'low' (200 µA, 35 kVp) or 'high' (400 µA, 45 kVp) power x-rays onto a flat-panel CsI detector. This study examines the performance characteristics and quantitative accuracy of the PET and SPECT components. Using the NEMA NU 4-2008 specifications (²²Na point source), the PET spatial resolution is 1.5 + 0.1 mm on axis and sensitivity 6.3% (axial centre) and 4.6% (central 70 mm). The usable activity range is ≤ 10 MBq (¹⁸F) over which good linearity (within 5%) is obtained for a uniform cylinder spanning the axial FOV; increasing deviation from linearity with activity is, however, observed for the NEMA (mouse) line source phantom. Image uniformity axially is within 5%. Spatial resolution (SPH/MPH) for the minimum SPECT FOV used for mouse imaging (50 mm) is 1.5/1.7 mm and point source sensitivity 69/750 cps MBq^–1. Axial uniformity of SPECT images (%CV of regions-of-interest counts along the axis) is mostly within 8% although there is a range of 30–40% for the largest FOV. The variation is significantly smaller within the central 40 mm. Instances of count rate nonlinearity (PET) and axial non-uniformity (SPECT) were found to be reproducible and thus amenable to empirical correction.

Previous methods to estimate the inherent accuracy of deformable image registration (DIR) have typically been performed relative to a known ground truth, such as tracking of anatomic landmarks or known deformations in a physical or virtual phantom. In this study, we propose a new approach to estimate the spatial geometric uncertainty of DIR using statistical sampling techniques that can be applied to the resulting deformation vector fields (DVFs) for a given registration. The proposed DIR performance metric, the distance discordance metric (DDM), is based on the variability in the distance between corresponding voxels from different images, which are co-registered to the same voxel at location (X) in an arbitrarily chosen 'reference' image. The DDM value, at location (X) in the reference image, represents the mean dispersion between voxels, when these images are registered to other images in the image set. The method requires at least four registered images to estimate the uncertainty of the DIRs, both for inter- and intra-patient DIR. To validate the proposed method, we generated an image set by deforming a software phantom with known DVFs. The registration error was computed at each voxel in the 'reference' phantom and then compared to DDM, inverse consistency error (ICE), and transitivity error (TE) over the entire phantom. The DDM showed a higher Pearson correlation (R_p) with the actual error (R_p ranged from 0.6 to 0.9) in comparison with ICE and TE (R_p ranged from 0.2 to 0.8). In the resulting spatial DDM map, regions with distinct intensity gradients had a lower discordance and therefore, less variability relative to regions with uniform intensity. Subsequently, we applied DDM for intra-patient DIR in an image set of ten longitudinal computed tomography (CT) scans of one prostate cancer patient and for inter-patient DIR in an image set of ten planning CT scans of different head and neck cancer patients. For both intra- and inter-patient DIR, the spatial DDM map showed large variation over the volume of interest (the pelvis for the prostate patient and the head for the head and neck patients). The highest discordance was observed in the soft tissues, such as the brain, bladder, and rectum, due to higher variability in the registration. The smallest DDM values were observed in the bony structures in the pelvis and the base of the skull. The proposed metric, DDM, provides a quantitative tool to evaluate the performance of DIR when a set of images is available. Therefore, DDM can be used to estimate and visualize the uncertainty of intra- and/or inter-patient DIR based on the variability of the registration rather than the absolute registration error.

Glioblastoma differ from many other tumors in the sense that they grow infiltratively into the brain tissue instead of forming a solid tumor mass with a defined boundary. Only the part of the tumor with high tumor cell density can be localized through imaging directly. In contrast, brain tissue infiltrated by tumor cells at low density appears normal on current imaging modalities. In current clinical practice, a uniform margin, typically two centimeters, is applied to account for microscopic spread of disease that is not directly assessable through imaging. The current treatment planning procedure can potentially be improved by accounting for the anisotropy of tumor growth, which arises from different factors: anatomical barriers such as the falx cerebri represent boundaries for migrating tumor cells. In addition, tumor cells primarily spread in white matter and infiltrate gray matter at lower rate. We investigate the use of a phenomenological tumor growth model for treatment planning. The model is based on the Fisher–Kolmogorov equation, which formalizes these growth characteristics and estimates the spatial distribution of tumor cells in normal appearing regions of the brain. The target volume for radiotherapy planning can be defined as an isoline of the simulated tumor cell density. This paper analyzes the model with respect to implications for target volume definition and identifies its most critical components. A retrospective study involving ten glioblastoma patients treated at our institution has been performed. To illustrate the main findings of the study, a detailed case study is presented for a glioblastoma located close to the falx. In this situation, the falx represents a boundary for migrating tumor cells, whereas the corpus callosum provides a route for the tumor to spread to the contralateral hemisphere. We further discuss the sensitivity of the model with respect to the input parameters. Correct segmentation of the brain appears to be the most crucial model input. We conclude that the tumor growth model provides a method to account for anisotropic growth patterns of glioma, and may therefore provide a tool to make target delineation more objective and automated.

Gliomas differ from many other tumors as they grow infiltratively into the brain parenchyma rather than forming a solid tumor mass with a well-defined boundary. Tumor cells can be found several centimeters away from the central tumor mass that is visible using current imaging techniques. The infiltrative growth characteristics of gliomas question the concept of a radiotherapy target volume that is irradiated to a homogeneous dose—the standard in current clinical practice. We discuss the use of the Fisher–Kolmogorov glioma growth model in radiotherapy treatment planning. The phenomenological tumor growth model assumes that tumor cells proliferate locally and migrate into neighboring brain tissue, which is mathematically described via a partial differential equation for the spatio-temporal evolution of the tumor cell density. In this model, the tumor cell density drops approximately exponentially with distance from the visible gross tumor volume, which is quantified by the infiltration length, a parameter describing the distance at which the tumor cell density drops by a factor of e. This paper discusses the implications for the prescribed dose distribution in the periphery of the tumor. In the context of the exponential cell kill model, an exponential fall-off of the cell density suggests a linear fall-off of the prescription dose with distance. We introduce the dose fall-off rate, which quantifies the steepness of the prescription dose fall-off in units of Gy mm⁻¹. It is shown that the dose fall-off rate is given by the inverse of the product of radiosensitivity and infiltration length. For an infiltration length of 3 mm and a surviving fraction of 50% at 2 Gy, this suggests a dose fall-off of approximately 1 Gy mm⁻¹. The concept is illustrated for two glioblastoma patients by optimizing intensity-modulated radiotherapy plans. The dose fall-off rate concept reflects the idea that infiltrating gliomas lack a defined boundary and are characterized by a continuous fall-off of the density of infiltrating tumor cells. The approach can potentially be used to individualize the prescribed dose distribution if better methods to estimate radiosensitivity and infiltration length on a patient by patient basis become available.

The full text of this article is available in the PDF provided.

The full text of this article is available in the PDF provided.