Table of contents

We present an experimental study of a novel method to verify the range of proton therapy beams. Differential cross sections were measured for 15 prompt gamma-ray lines from proton-nuclear interactions with ¹²C and ¹⁶O at proton energies up to 150 MeV. These cross sections were used to model discrete prompt gamma-ray emissions along proton pencil-beams. By fitting detected prompt gamma-ray counts to these models, we simultaneously determined the beam range and the oxygen and carbon concentration of the irradiated matter. The performance of the method was assessed in two phantoms with different elemental concentrations, using a small scale prototype detector. Based on five pencil-beams with different ranges delivering 5 × 10⁸ protons and without prior knowledge of the elemental composition at the measurement point, the absolute range was determined with a standard deviation of 1.0–1.4 mm. Relative range shifts at the same dose level were detected with a standard deviation of 0.3–0.5 mm. The determined oxygen and carbon concentrations also agreed well with the actual values. These results show that quantitative prompt gamma-ray measurements enable knowledge of nuclear reaction cross sections to be used for precise proton range verification in the presence of tissue with an unknown composition.

There is increasing interest in using Cerenkov emissions for quality assurance and in vivo dosimetry in photon and electron therapy. Here, we investigate the production of Cerenkov light during proton therapy and its potential applications in proton therapy. A primary proton beam does not have sufficient energy to generate Cerenkov emissions directly, but we have demonstrated two mechanisms by which such emissions may occur indirectly: (1) a fast component from fast electrons liberated by prompt gamma (99.13%) and neutron (0.87%) emission; and (2) a slow component from the decay of radioactive positron emitters. The fast component is linear with dose and doserate but carries little spatial information; the slow component is non-linear but may be localised.

The properties of the two types of emission are explored using Monte Carlo modelling in GEANT4 with some experimental verification. We propose that Cerenkov emissions could contribute to the visual sensation reported by some patients undergoing proton therapy of the eye and we discuss the feasibility of some potential applications of Cerenkov imaging in proton therapy.

The majority of current whole-body PET scanners are based on pixelated scintillator arrays with a transverse pixel size of 4 mm. However, recent studies have shown that decreasing the pixel size to 2 mm can significantly improve image spatial resolution. In this study, the performance of Digital Photon Counter (DPC) from Philips Digital Photon Counting (PDPC) was evaluated to determine their potential for high-resolution whole-body time of flight (TOF) PET scanners. Two detector configurations were evaluated. First, the DPC3200-44-22 DPC array was coupled to a LYSO block of 15  ×  15 2  ×  2 × 22 mm³ pixels through a 1 mm thick light guide. Due to light sharing among the dies neighbour logic of the DPC was used. In a second setup the same DPC was coupled directly to a scalable 4  ×  4 LYSO matrix of 1.9  ×  1.9  ×  22 mm³ crystals with a dedicated reflector arrangement allowing for controlled light sharing patterns inside the matrix. With the first approach an average energy resolution of 14.5% and an average CRT of 376 ps were achieved. For the second configuration an average energy resolution of 11% and an average CRT of 295 ps were achieved. Our studies show that the DPC is a suitable photosensor for a high-resolution TOF-PET detector. The dedicated reflector arrangement allows one to achieve better performances than the light guide approach. The count loss, caused by dark counts, is overcome by fitting the matrix size to the size of DPC single die.

Solving the inverse problem of electrocardiography (ECG) and magnetocardiography (MCG) is often referred to as cardiac source imaging. Spatial properties of ECG and MCG as imaging systems are, however, not well known. In this modelling study, we investigate the sensitivity and point-spread function (PSF) of ECG, MCG, and combined ECG+MCG as a function of source position and orientation, globally around the ventricles: signal topographies are modelled using a realistically-shaped volume conductor model, and the inverse problem is solved using a distributed source model and linear source estimation with minimal use of prior information. The results show that the sensitivity depends not only on the modality but also on the location and orientation of the source and that the sensitivity distribution is clearly reflected in the PSF. MCG can better characterize tangential anterior sources (with respect to the heart surface), while ECG excels with normally-oriented and posterior sources. Compared to either modality used alone, the sensitivity of combined ECG+MCG is less dependent on source orientation per source location, leading to better source estimates. Thus, for maximal sensitivity and optimal source estimation, the electric and magnetic measurements should be combined.

This work analysed the feasibility of using a fast, customized Monte Carlo (MC) method to perform accurate computation of dose distributions during pre- and intraplanning of intraoperative electron radiation therapy (IOERT) procedures. The MC method that was implemented, which has been integrated into a specific innovative simulation and planning tool, is able to simulate the fate of thousands of particles per second, and it was the aim of this work to determine the level of interactivity that could be achieved. The planning workflow enabled calibration of the imaging and treatment equipment, as well as manipulation of the surgical frame and insertion of the protection shields around the organs at risk and other beam modifiers. In this way, the multidisciplinary team involved in IOERT has all the tools necessary to perform complex MC dosage simulations adapted to their equipment in an efficient and transparent way. To assess the accuracy and reliability of this MC technique, dose distributions for a monoenergetic source were compared with those obtained using a general-purpose software package used widely in medical physics applications. Once accuracy of the underlying simulator was confirmed, a clinical accelerator was modelled and experimental measurements in water were conducted. A comparison was made with the output from the simulator to identify the conditions under which accurate dose estimations could be obtained in less than 3 min, which is the threshold imposed to allow for interactive use of the tool in treatment planning. Finally, a clinically relevant scenario, namely early-stage breast cancer treatment, was simulated with pre- and intraoperative volumes to verify that it was feasible to use the MC tool intraoperatively and to adjust dose delivery based on the simulation output, without compromising accuracy. The workflow provided a satisfactory model of the treatment head and the imaging system, enabling proper configuration of the treatment planning system and providing good accuracy in the dosage simulation.

To investigate the effect of metal implants in proton radiotherapy, dose distributions of different, clinically relevant treatment plans have been measured in an anthropomorphic phantom and compared to treatment planning predictions. The anthropomorphic phantom, which is sliced into four segments in the cranio-caudal direction, is composed of tissue equivalent materials and contains a titanium implant in a vertebral body in the cervical region. GafChromic® films were laid between the different segments to measure the 2D delivered dose. Three different four-field plans have then been applied: a Single-Field-Uniform-Dose (SFUD) plan, both with and without artifact correction implemented, and an Intensity-Modulated-Proton-Therapy (IMPT) plan with the artifacts corrected. For corrections, the artifacts were manually outlined and the Hounsfield Units manually set to an average value for soft tissue. Results show a surprisingly good agreement between prescribed and delivered dose distributions when artifacts have been corrected, with > 97% and 98% of points fulfilling the gamma criterion of 3%/3 mm for both SFUD and the IMPT plans, respectively. In contrast, without artifact corrections, up to 18% of measured points fail the gamma criterion of 3%/3 mm for the SFUD plan. These measurements indicate that correcting manually for the reconstruction artifacts resulting from metal implants substantially improves the accuracy of the calculated dose distribution.

The air kerma-area product (KAP) is used for settings of diagnostic reference levels. The International Atomic Energy Agency (IAEA) recommends that doses in diagnostic radiology (including the KAP values) be estimated with an accuracy of at least ±7% (k = 2). Industry standards defined by the International Electrotechnical Commission (IEC) specify that the uncertainty of KAP meter measurements should be less than $\pm 25%$ (k = 2). Medical physicists willing to comply with the IAEA's recommendation need to apply correction factors to KAP values reported by x-ray units. The aim of this work is to present and evaluate a calibration method for built-in KAP meters on clinical x-ray units. The method is based on (i) a tandem calibration method, which uses a reference KAP meter calibrated to measure the incident radiation, (ii) measurements using an energy-independent ionization chamber to correct for the energy dependence of the reference KAP meter, and (iii) Monte Carlo simulations of the beam quality correction factors that correct for differences between beam qualities at a standard laboratory and the clinic. The method was applied to the KAP meter in a Siemens Aristos FX plus unit. It was found that values reported by the built-in KAP meter differed from the more accurate values measured by the reference KAP meter by more than 25% for high tube voltages (more than 140 kV) and heavily filtered beams (0.3 mm Cu). Associated uncertainties were too high to claim that the IEC's limit of 25% was exceeded. Nevertheless the differences were high enough to justify the need for a more accurate calibration of built-in KAP meters.

Accurate energy calibration is critical for the application of energy-resolved photon-counting detectors in spectral imaging. The aim of this study is to investigate the feasibility of energy response calibration and characterization of a photon-counting detector using x-ray fluorescence. A comprehensive Monte Carlo simulation study was performed using Geant4 Application for Tomographic Emission (GATE) to investigate the optimal technique for x-ray fluorescence calibration. Simulations were conducted using a 100 kVp tungsten-anode spectra with 2.7 mm Al filter for a single pixel cadmium telluride (CdTe) detector with 3  ×  3 mm² in detection area. The angular dependence of x-ray fluorescence and scatter background was investigated by varying the detection angle from 20° to 170° with respect to the beam direction. The effects of the detector material, shape, and size on the recorded x-ray fluorescence were investigated. The fluorescent material size effect was considered with and without the container for the fluorescent material. In order to provide validation for the simulation result, the angular dependence of x-ray fluorescence from five fluorescent materials was experimentally measured using a spectrometer. Finally, eleven of the fluorescent materials were used for energy calibration of a CZT-based photon-counting detector. The optimal detection angle was determined to be approximately at 120° with respect to the beam direction, which showed the highest fluorescence to scatter ratio (FSR) with a weak dependence on the fluorescent material size. The feasibility of x-ray fluorescence for energy calibration of photon-counting detectors in the diagnostic x-ray energy range was verified by successfully calibrating the energy response of a CZT-based photon-counting detector. The results of this study can be used as a guideline to implement the x-ray fluorescence calibration method for photon-counting detectors in a typical imaging laboratory.

In clinical ion beam therapy, protons as well as heavier ions such as carbon are used for treatment. For protons, β⁺-emitters are only induced by fragmentation reactions in the target (target fragmentation), whereas for heavy ions, they are additionally induced by fragmentations of the projectile (further referred to as autoactivation). An approach utilizing these processes for treatment verfication, by comparing measured Positron Emission Tomography (PET) data to predictions from Monte Carlo simulations, has already been clinically implemented. For an accurate simulation, it is important to consider the biological washout of β⁺-emitters due to vital functions. To date, mathematical expressions for washout have mainly been determined by using radioactive beams of ¹⁰C- and ¹¹C-ions, both β⁺-emitters, to enhance the counting statistics in the irradiated area. Still, the question of how the choice of projectile (autoactivating or non-autoactivating) influences the washout coefficients, has not been addressed.

In this context, an experiment was carried out at the Heidelberg Ion Beam Therapy Center with the purpose of directly comparing irradiation-induced biological washout coefficients in mice for protons and ¹²C-ions. To this aim, mice were irradiated in the brain region with protons and ¹²C-ions and measured after irradiation with a PET/CT scanner (Siemens Biograph mCT). After an appropriate waiting time, the mice were sacrificed, then irradiated and measured again under similar conditions. The resulting data were processed and fitted numerically to deduce the main washout parameters.

Despite the very low PET counting statistics, a consistent difference could be identified between ¹²C-ion and proton irradiated mice, with the ¹²C data being described best by a two component fit with a combined medium and slow washout fraction of 0.50 ± 0.05 and the proton mice data being described best by a one component fit with only one (slow) washout fraction of 0.73 ± 0.06.

Accurate segmentation of hip joint cartilage from magnetic resonance (MR) images offers opportunities for quantitative investigations of pathoanatomical conditions such as osteoarthritis. In this paper, we present a fully automatic scheme for the segmentation of the individual femoral and acetabular cartilage plates in the human hip joint from high-resolution 3D MR images. The developed scheme uses an improved optimal multi-object multi-surface graph search framework with an arc-weighted graph representation that incorporates prior morphological knowledge as a basis for segmentation of the individual femoral and acetabular cartilage plates despite weak or incomplete boundary interfaces. This automated scheme was validated against manual segmentations from 3D true fast imaging with steady-state precession (TrueFISP) MR examinations of the right hip joints in 52 asymptomatic volunteers. Compared with expert manual segmentations of the combined, femoral and acetabular cartilage volumes, the automatic scheme obtained mean (± standard deviation) Dice's similarity coefficients of 0.81 (± 0.03), 0.79 (± 0.03) and 0.72 (± 0.05). The corresponding mean absolute volume difference errors were 8.44% (± 6.36), 9.44% (± 7.19) and 9.05% (± 8.02). The mean absolute differences between manual and automated measures of cartilage thickness for femoral and acetabular cartilage plates were 0.13 mm (± 0.12) and 0.11 mm (± 0.11), respectively.

Hyperpolarized gas magnetic resonance imaging (MRI) generates highly detailed maps of lung ventilation and physiological function while CT provides corresponding anatomical and structural information. Fusion of such complementary images enables quantitative analysis of pulmonary structure-function. However, direct image registration of hyperpolarized gas MRI to CT is problematic, particularly in lungs whose boundaries are difficult to delineate due to ventilation heterogeneity. This study presents a novel indirect method of registering hyperpolarized gas MRI to CT utilizing ¹H-structural MR images that are acquired in the same breath-hold as the gas MRI. The feasibility of using this technique for regional quantification of ventilation of specific pulmonary structures is demonstrated for the lobes.

The direct and indirect methods of hyperpolarized gas MRI to CT image registration were compared using lung images from 15 asthma patients. Both affine and diffeomorphic image transformations were implemented. Registration accuracy was evaluated using the target registration error (TRE) of anatomical landmarks identified on ¹H MRI and CT. The Wilcoxon signed-rank test was used to test statistical significance.

For the affine transformation, the indirect method of image registration was significantly more accurate than the direct method (TRE = 14.7  ±  3.2 versus 19.6  ±  12.7 mm, p = 0.036). Using a deformable transformation, the indirect method was also more accurate than the direct method (TRE = 13.5  ±  3.3 versus 20.4  ±  12.8 mm, p = 0.006).

Accurate image registration is critical for quantification of regional lung ventilation with hyperpolarized gas MRI within the anatomy delineated by CT. Automatic deformable image registration of hyperpolarized gas MRI to CT via same breath-hold ¹H MRI is more accurate than direct registration. Potential applications include improved multi-modality image fusion, functionally weighted radiotherapy planning, and quantification of lobar ventilation in obstructive airways disease.

Multileaf collimator (MLC) tracking is a promising and clinically emerging treatment modality for radiotherapy of mobile tumours. Still, new quality assurance (QA) methods are warranted to safely introduce MLC tracking in the clinic. The purpose of this study was to create and experimentally validate a simple model for fast motion-including dose error reconstruction applicable to intrafractional QA of MLC tracking treatments of moving targets.

MLC tracking experiments were performed on a standard linear accelerator with prototype MLC tracking software guided by an electromagnetic transponder system. A three-axis motion stage reproduced eight representative tumour trajectories; four lung and four prostate. Low and high modulation 6 MV single-arc volumetric modulated arc therapy treatment plans were delivered for each trajectory with and without MLC tracking, as well as without motion for reference. Temporally resolved doses were measured during all treatments using a biplanar dosimeter. Offline, the dose delivered to each of 1069 diodes in the dosimeter was reconstructed with 500 ms temporal resolution by a motion-including pencil beam convolution algorithm developed in-house. The accuracy of the algorithm for reconstruction of dose and motion-induced dose errors throughout the tracking and non-tracking beam deliveries was quantified.

Doses were reconstructed with a mean dose difference relative to the measurements of −0.5% (5.5% standard deviation) for cumulative dose. More importantly, the root-mean-square deviation between reconstructed and measured motion-induced 3%/3 mm γ failure rates (dose error) was 2.6%. The mean computation time for each calculation of dose and dose error was 295 ms.

The motion-including dose reconstruction allows accurate temporal and spatial pinpointing of errors in absorbed dose and is adequately fast to be feasible for online use. An online implementation could allow treatment intervention in case of erroneous dose delivery in both tracking and non-tracking treatments.

In this study, we develope a novel method to directly evaluate an absorbed dose-to-water for kilovoltage-cone beam computed tomography (kV-CBCT) in image-guided radiation therapy (IGRT). Absorbed doses for the kV-CBCT systems of the Varian On-Board Imager (OBI) and the Elekta X-ray Volumetric Imager (XVI) were measured by a Farmer ionization chamber with a ⁶⁰Co calibration factor. The chamber measurements were performed at the center and four peripheral points in body-type (30 cm diameter and 51 cm length) and head-type (16 cm diameter and 33 cm length) cylindrical water phantoms. The measured ionization was converted to the absorbed dose-to-water by using a ⁶⁰Co calibration factor and a Monte Carlo (MC)-calculated beam quality conversion factor, k_Q, for ⁶⁰Co to kV-CBCT. The irradiation for OBI and XVI was performed with pelvis and head modes for the body- and the head-type phantoms, respectively. In addition, the dose distributions in the phantom for both kV-CBCT systems were calculated with MC method and were compared with measured values. The MC-calculated doses were calibrated at the center in the water phantom and compared with measured doses at four peripheral points. The measured absorbed doses at the center in the body-type phantom were 1.96 cGy for OBI and 0.83 cGy for XVI. The peripheral doses were 2.36–2.90 cGy for OBI and 0.83–1.06 cGy for XVI. The doses for XVI were lower up to approximately one-third of those for OBI. Similarly, the measured doses at the center in the head-type phantom were 0.48 cGy for OBI and 0.21 cGy for XVI. The peripheral doses were 0.26–0.66 cGy for OBI and 0.16–0.30 cGy for XVI. The calculated peripheral doses agreed within 3% in the pelvis mode and within 4% in the head mode with measured doses for both kV-CBCT systems. In addition, the absorbed dose determined in this study was approximately 4% lower than that in TG-61 but the absorbed dose by both methods was in agreement within their combined uncertainty. This method is more robust and accurate compared to the dosimetry based on a conventional air-kerma calibration factor. Therefore, it is possible to be used as a standard dosimetry protocol for kV-CBCT in IGRT.

The aim of this study is to present a modulation index (MI) for volumetric modulated arc therapy (VMAT) based on the speed and acceleration analysis of modulating-parameters such as multi-leaf collimator (MLC) movements, gantry rotation and dose-rate, comprehensively. The performance of the presented MI (MI_t) was evaluated with correlation analyses to the pre-treatment quality assurance (QA) results, differences in modulating-parameters between VMAT plans versus dynamic log files, and differences in dose-volumetric parameters between VMAT plans versus reconstructed plans using dynamic log files. For comparison, the same correlation analyses were performed for the previously suggested modulation complexity score (MCS_v), leaf travel modulation complexity score (LTMCS) and MI by Li and Xing (MI Li&Xing). In the two-tailed unpaired parameter condition, p values were acquired. The Spearman's rho (r_s) values of MI_t, MCS_v, LTMCS and MI Li&Xing to the local gamma passing rate with 2%/2 mm criterion were −0.658 (p < 0.001), 0.186 (p = 0.251), 0.312 (p = 0.05) and −0.455 (p = 0.003), respectively. The values of r_s to the modulating-parameter (MLC positions) differences were 0.917, −0.635, −0.857 and 0.795, respectively (p < 0.001). For dose-volumetric parameters, MI_t showed higher statistically significant correlations than the conventional MIs. The MI_t showed good performance for the evaluation of the modulation-degree of VMAT plans.

FoCa is an in-house modular treatment planning system, developed entirely in MATLAB, which includes forward dose calculation of proton radiotherapy plans in both active and passive modalities as well as a generic optimization suite for inverse treatment planning.

The software has a dual education and research purpose. From the educational point of view, it can be an invaluable teaching tool for educating medical physicists, showing the insights of a treatment planning system from a well-known and widely accessible software platform. From the research point of view, its current and potential uses range from the fast calculation of any physical, radiobiological or clinical quantity in a patient CT geometry, to the development of new treatment modalities not yet available in commercial treatment planning systems.

The physical models in FoCa were compared with the commissioning data from our institution and show an excellent agreement in depth dose distributions and longitudinal and transversal fluence profiles for both passive scattering and active scanning modalities. 3D dose distributions in phantom and patient geometries were compared with a commercial treatment planning system, yielding a gamma-index pass rate of above 94% (using FoCa's most accurate algorithm) for all cases considered.

Finally, the inverse treatment planning suite was used to produce the first prototype of intensity-modulated, passive-scattered proton therapy, using 13 passive scattering proton fields and multi-leaf modulation to produce a concave dose distribution on a cylindrical solid water phantom without any field-specific compensator.

Dynamic contrast-enhanced (DCE) images can be acquired at multiple time points and multiple slice locations of a tumor. Image segmentation and registration are important preprocessing steps that can improve subsequent analysis of DCE images by kinetic modeling. An automatic system for region-of-interest segmentation and registration of DCE images is presented. Tissue segmentation is performed using a combination of thresholding and morphological operations, and further refined using shape information from consecutive images. The segmented regions are subsequently registered based on a mutual information method that accounts for possible tissue movement between slices. The proposed segmentation and registration methods are applied on actual DCE CT datasets to illustrate feasibility of practical implementation in the clinic.

The purpose of this study was to investigate what pseudo-CT (pCT) strategy is sufficient for online MRI based treatment plan adaptation of a stereotactic treatment for spinal bone metastases. For this purpose, the dosimetric accuracy of five increasingly complex pCT strategies was evaluated using the planning CT data of 20 patients suffering from spinal metastases. For each pCT, a treatment plan was developed and simulated on both the pCT and the original CT data of the patient. The two resulting dose distributions were compared using gamma analysis of 2%/2 mm. In this paper, a Gamma Pass Rate (GPR) of ⩾95% within the Target Volume (TV) was considered clinically acceptable. We additionally demonstrated in this paper the automatic generation of each investigated pCT strategy with the use of dedicated MRI data complemented with pre-treatment CT data of a patient in treatment position. The dosimetric accuracy of a pCT increases when additional bulk densities are utilized for a pCT. However, the dosimetric accuracy of even the most complex 'bulk density' pCT strategy used in this study had an average GPR of only 78% within the TV. However, if information on the heterogeneous electron density distribution within the affected vertebral body was available, a clinically acceptable 99% mean GPR was observed. All pCTs could successfully be generated using the MRI data in combination with the CT data of a patient in treatment position. The results presented in this study show that a simple 'bulk density' pseudo-CT strategy is not feasible for online MRI based treatment plan adaptation for spinal bone metastases. However, a clinically acceptable result is generated if the information on the heterogeneous electron density (ED) distribution within the affected vertebral bone is available. Therefore, any pCT strategy for this tumor site should include a method which can estimate the heterogeneous ED of the affected vertebral bone.

The relative biological effectiveness (RBE) concept is commonly used in treatment planning for ion beam therapy. Whether models based on in vitro/in vivo RBE data can be used to predict human response to treatments is an open issue. In this work an alternative method, based on an effective radiobiological parameterization directly derived from clinical data, is presented. The method has been applied to the analysis of prostate cancer trials with protons and carbon ions.

Prostate cancer trials with proton and carbon ion beams reporting 5 year-local control (LC5) and grade 2 (G2) or higher genitourinary toxicity rates (TOX) were selected from literature to test the method. Treatment simulations were performed on a representative subset of patients to produce dose and linear energy transfer distribution, which were used as explicative physical variables for the radiobiological modelling. Two models were taken into consideration: the microdosimetric kinetic model (MKM) and a linear model (LM). The radiobiological parameters of the LM and MKM were obtained by coupling them with the tumor control probability and normal tissue complication probability models to fit the LC5 and TOX data through likelihood maximization. The model ranking was based on the Akaike information criterion.

Results showed large confidence intervals due to the limited variety of available treatment schedules. RBE values, such as RBE = 1.1 for protons in the treated volume, were derived as a by-product of the method, showing a consistency with current approaches. Carbon ion RBE values were also derived, showing lower values than those assumed for the original treatment planning in the target region, whereas higher values were found in the bladder. Most importantly, this work shows the possibility to infer the radiobiological parametrization for proton and carbon ion treatment directly from clinical data.

In addition to conventional short-lived radionuclides, longer-lived isotopes are becoming increasingly important to positron emission tomography (PET). The longer half-life both allows for circumvention of the in-house production of radionuclides, and expands the spectrum of physiological processes amenable to PET imaging, including processes with prohibitively slow kinetics for investigation with short-lived radiotracers. However, many of these radionuclides emit 'high-energy' positrons and gamma rays which affect the spatial resolution and quantitative accuracy of PET images.

The objective of the present work is to investigate the positron range distribution for some of these long-lived isotopes.

Based on existing Monte Carlo simulations of positron interactions in water, the probability distribution of the line of response displacement have been empirically described by means of analytic displacement functions.

Relevant distributions have been derived for the isotopes ²²Na, ⁵²Mn, ⁸⁹Zr, ⁴⁵Ti, ⁵¹Mn, ^94mTc, ^52mMn, ³⁸K, ⁶⁴Cu, ⁸⁶Y, ¹²⁴I, and ¹²⁰I. It was found that the distribution functions previously found for a series of conventional isotopes (Jødal et al2012Phys. Med. Bio.57 3931–43), were also applicable to these non-conventional isotopes, except that for ¹²⁰I, ¹²⁴I, ⁸⁹Zr, ⁵²Mn, and ⁶⁴Cu, parameters in the formulae were less well predicted by mean positron energy alone.

Both conventional and non-conventional range distributions can be described by relatively simple analytic expressions. The results will be applicable to image-reconstruction software to improve the resolution.

A simple formula to determine the human average whole-body SAR (SAR_wb) under realistic propagation conditions is proposed in the GHz region, i.e. from 1.45 GHz to 5.8 GHz. The methodology is based on simulations of ellipsoidal human body models. Only the exposure (incident power densities) and the human mass are needed to apply the formula. Diffuse scattered illumination is addressed for the first time and the possible presence of a Line-of-Sight (LOS) component is addressed as well. As validation, the formula is applied to calculate the average whole-body SAR_wb in 3D heterogeneous phantoms, i.e. the virtual family (34 year-old male, 26 year-old female, 11 year-old girl, and 6 year-old boy) and the results are compared with numerical ones—using the Finite-Difference Time-Domain (FDTD) method—at 3 GHz. For the LOS exposure, the average relative error varies from 28% to 12% (resp. 14–12%) for the vertical polarization (resp. horizontal polarization), depending on the heteregeneous phantom. Regarding the diffuse illumination, relative errors of −39.40%, −11.70%, 10.70%, and 10.60% are obtained for the 6 year-old boy, 11 year-old girl, 26 year-old female, and 34 year-old male, respectively. The proposed formula estimates well (especially for adults) the SAR_wb induced by diffuse illumination in realistic conditions. In general, the correctness of the formula improves when the human mass increases. Keeping the uncertainties of the FDTD simulations in mind, the proposed formula might be important for the dosimetry community to assess rapidly and accurately the human absorption of electromagnetic radiation caused by diffuse fields in the GHz region. Finally, we show the applicability of the proposed formula to personal dosimetry for epidemiological research.

This paper describes a new approach to detect microcalcification clusters (MCs) in digital breast tomosynthesis (DBT) via its planar projection (PPJ) image. With IRB approval, two-view (cranio-caudal and mediolateral oblique views) DBTs of human subject breasts were obtained with a GE GEN2 prototype DBT system that acquires 21 projection angles spanning 60° in 3° increments. A data set of 307 volumes (154 human subjects) was divided by case into independent training (127 with MCs) and test sets (104 with MCs and 76 free of MCs). A simultaneous algebraic reconstruction technique with multiscale bilateral filtering (MSBF) regularization was used to enhance microcalcifications and suppress noise. During the MSBF regularized reconstruction, the DBT volume was separated into high frequency (HF) and low frequency components representing microcalcifications and larger structures. At the final iteration, maximum intensity projection was applied to the regularized HF volume to generate a PPJ image that contained MCs with increased contrast-to-noise ratio (CNR) and reduced search space. High CNR objects in the PPJ image were extracted and labeled as microcalcification candidates. Convolution neural network trained to recognize the image pattern of microcalcifications was used to classify the candidates into true calcifications and tissue structures and artifacts. The remaining microcalcification candidates were grouped into MCs by dynamic conditional clustering based on adaptive CNR threshold and radial distance criteria. False positive (FP) clusters were further reduced using the number of candidates in a cluster, CNR and size of microcalcification candidates. At 85% sensitivity an FP rate of 0.71 and 0.54 was achieved for view- and case-based sensitivity, respectively, compared to 2.16 and 0.85 achieved in DBT. The improvement was significant (p-value = 0.003) by JAFROC analysis.

A parametric model for the x-ray linear attenuation coefficient is used to describe the compositional dependence of Hounsfield numbers measured by medical CT scanners. Measurements with materials of known density and composition, that span and evenly sample the compositional range of tissues, are written as linear simultaneous equations and solved for model coefficients. An algorithm is identified for this purpose. Results are expressed as atomic cross-sections in units of barn per electron divided by the attenuation coefficient for water. With the CT scanner characterised, a virtual CT scan can be simulated to predict HN for tissues based upon their known density and composition. Similar calculations using the tabulations and mixture rule deliver attenuation coefficients and mass energy absorption coefficients for mono-energetic radiation 10 keV to 20 MeV. Results are presented for measurements with a radiotherapy CT simulator, the RMI-467 phantom with tissue substitute materials, plus common polymer materials and silicon. Published measurements with earlier generations of the phantom and tissue substitutes using different CT scanners are also considered. Measured atomic cross-sections differ from expectations for mono-energetic radiation due to the use of a filtered spectrum and energy integrating detection system. The cross-sections for different CT scanners are similar, without large variations with kVp. Results are presented showing the relationship between predicted HN for tissues, electron density and photon interaction coefficients for healthy tissues and mono-energetic radiation. A strategy is suggested for accommodating strongly attenuating materials such as calculi and metallic implants.

Radiotherapy (RT) based on magnetic resonance imaging (MRI) as the only modality, so-called MRI-only RT, would remove the systematic registration error between MR and computed tomography (CT), and provide co-registered MRI for assessment of treatment response and adaptive RT. Electron densities, however, need to be assigned to the MRI images for dose calculation and patient setup based on digitally reconstructed radiographs (DRRs). Here, we investigate the geometric and dosimetric performance for a number of popular voxel-based methods to generate a so-called pseudo CT (pCT).

Five patients receiving cranial irradiation, each containing a co-registered MRI and CT scan, were included. An ultra short echo time MRI sequence for bone visualization was used. Six methods were investigated for three popular types of voxel-based approaches; (1) threshold-based segmentation, (2) Bayesian segmentation and (3) statistical regression. Each approach contained two methods. Approach 1 used bulk density assignment of MRI voxels into air, soft tissue and bone based on logical masks and the transverse relaxation time T₂ of the bone. Approach 2 used similar bulk density assignments with Bayesian statistics including or excluding additional spatial information. Approach 3 used a statistical regression correlating MRI voxels with their corresponding CT voxels. A similar photon and proton treatment plan was generated for a target positioned between the nasal cavity and the brainstem for all patients. The CT agreement with the pCT of each method was quantified and compared with the other methods geometrically and dosimetrically using both a number of reported metrics and introducing some novel metrics.

The best geometrical agreement with CT was obtained with the statistical regression methods which performed significantly better than the threshold and Bayesian segmentation methods (excluding spatial information). All methods agreed significantly better with CT than a reference water MRI comparison. The mean dosimetric deviation for photons and protons compared to the CT was about 2% and highest in the gradient dose region of the brainstem. Both the threshold based method and the statistical regression methods showed the highest dosimetrical agreement.

Generation of pCTs using statistical regression seems to be the most promising candidate for MRI-only RT of the brain. Further, the total amount of different tissues needs to be taken into account for dosimetric considerations regardless of their correct geometrical position.

MicroSPECT is one of the main functional imaging techniques used in the preclinical setting. Even though high-resolution images can be obtained with currently available systems, their sensitivity is often quite low due to the use of multi-pinhole collimation. This results in long acquisition times and hampers dynamic imaging. However, it has already been shown that this limited sensitivity can be overcome using high-resolution detectors. In this article, we therefore investigated the use of a digital photon counter (DPC) in combination with a 2 mm thick monolithic LYSO crystal for SPECT imaging. These light sensors contain arrays of avalanche photodiodes whose signals are directly digitised. The DPCs have the advantage that they are very compact, have a high intrinsic resolution, are MR compatible and allow disabling cells with a high dark count rate. In order to investigate the influence of the temperature dependent dark count rate on the detector performance, we compared it at 3 °C and 18 °C.

At 3 °C, we observed an energy resolution of 28.8% and an intrinsic spatial resolution of 0.48 mm. Furthermore, the count rate at 10% loss is 60 kcps. Next, we looked at the event loss at 18 °C caused by the higher dark count rate and found a 5% loss compared to the 3 °C measurements. At this higher temperature the energy resolution becomes 29.2% and the intrinsic spatial resolution decreases to 0.52 mm. Due to the 5% count loss, the count rate at 10% loss increases to 63 kcps. A small degradation of the detector performance is thus observed at 18 °C.

These results show the usefulness of this detector for SPECT imaging together with its excellent intrinsic spatial resolution. A drawback of the detector is its low, spatially varying energy resolution. Even though the detection efficiency and intrinsic spatial resolution are better at 3 °C, results are still acceptable at 18 °C.

Observing cavitation bubbles deep within tissue is very difficult. The development of a method for probing cavitation, irrespective of its location in tissues, would improve the efficiency and application of ultrasound in the clinic. A synchrotron x-ray imaging technique, which is capable of detecting cavitation bubbles induced in water by a sonochemistry system, is reported here; this could possibly be extended to the study of therapeutic ultrasound in tissues. The two different x-ray imaging techniques of Analyzer Based Imaging (ABI) and phase contrast imaging (PCI) were examined in order to detect ultrasound induced cavitation bubbles. Cavitation was not observed by PCI, however it was detectable with ABI. Acoustic cavitation was imaged at six different acoustic power levels and six different locations through the acoustic beam in water at a fixed power level. The results indicate the potential utility of this technique for cavitation studies in tissues, but it is time consuming. This may be improved by optimizing the imaging method.

Proton CT nowadays aims at improving hadron therapy treatment planning by mapping the stopping power of materials. In order to optimize a spatial resolution of the reconstructed images, the most likely path (MLP) of each proton can be computed. We investigated the errors in the computation of this path due to the configuration of the system, i.e. the spatial resolution of the tracking planes, their material budget, and their positioning. A method for computing the uncertainty in the estimated paths for a given system was derived. The uncertainties upon the entrance and exit of the object were propagated analytically in the path computation. This procedure was then used to evaluate the impact of each parameter, and to compare different systems. We show that the intrinsic characteristics of the system generate an uncertainty in the positions and directions of the particles propagated during the MLP computation. The spatial resolution and material budget of the trackers in particular may affect the path estimation, and thus the spatial resolution of an image.

Small-field output factor measurements are traditionally very difficult because of steep dose gradients, loss of lateral electronic equilibrium, and dose volume averaging in finitely sized detectors. Three-dimensional (3D) dosimetry is ideal for measuring small output factors and avoids many of these potential challenges of point and 2D detectors. PRESAGE 3D polymer dosimeters were used to measure the output factors for the 4 mm and 8 mm collimators of the Leksell Perfexion Gamma Knife radiosurgery treatment system. Discrepancies between the planned and measured distance between shot centers were also investigated. A Gamma Knife head frame was mounted onto an anthropomorphic head phantom. Special inserts were machined to hold 60 mm diameter, 70 mm tall cylindrical PRESAGE dosimeters. The phantom was irradiated with one 16 mm shot and either one 4 mm or one 8 mm shot, to a prescribed dose of either 3 Gy or 4 Gy to the 50% isodose line. The two shots were spaced between 30 mm and 60 mm apart and aligned along the central axis of the cylinder. The Presage dosimeters were measured using the DMOS-RPC optical CT scanning system. Five independent 4 mm output factor measurements fell within 2% of the manufacturer's Monte Carlo simulation-derived nominal value, as did two independent 8 mm output factor measurements. The measured distances between shot centers varied by ±0.8 mm with respect to the planned shot displacements. On the basis of these results, we conclude that PRESAGE dosimetry is excellently suited to quantify the difficult-to-measure Gamma Knife output factors.

The automatic patient positioning system and its alignment is critical and specified to be less than 0.35 mm for a radiosurgical treatment with the latest robotized Gamma Knife Perfexion (GKPFX). In this study, we developed a quantitative QA procedure to verify the accuracy and robustness of such a system. In particular, we applied the test to a unit that has performed >1000 procedures at our institution. For the test, a radiochromic film was first placed inside a spherical film phantom and then irradiated with a sequence of linearly placed shots of equal collimator size (e.g. 4 mm) via the Leksell Gamma Knife Perfexion system (PFX). The shots were positioned with either equal or unequal gaps of approximately 8 mm both at center and off-center positions of the patient positioning system. Two independent methods of localizing the irradiation shot center coordinates were employed to measure the gap spacing between adjacent shots. The measured distance was then compared with the initial preset values for the test. On average, the positioning uncertainty for the PFX delivery system was found to be 0.03  ±  0.2 mm (2σ). No significant difference in the positioning uncertainty was noted among measurements in the x-, y- and z-axis orientations. In conclusion, a simple, fast, and quantitative test was developed and demonstrated for routine QA of the submillimeter PFX patient positioning system. This test also enables independent verification of any patient-specific shot positioning for a critical treatment such as a tumor in the brainstem.

The Chi index described in the article 'A revision of the γ-evaluation concept for the comparison of dose distributions' by Bakai et al (Phys. Med. Biol. 2003 48 3543–53) indicates that smooth acceptance tubes, defining upper and lower limits of dose difference and distance to agreement, can be pre-defined for a given dose distribution based on the local dose gradient.

Mathematical analysis and simulations indicate that the Chi index as described by Bakai et al does not produce smooth acceptance criteria in rapidly varying dose gradients. Instead, 'horns' are generated in the acceptance tubes which lead to the production of unacceptably large acceptance criteria and the possibility of false negatives.