To compare organ specific cancer incidence risks for standard and complex external beam radiotherapy including cone beam CT verification following breast conservation surgery for early breast cancer.
Doses from breast radiotherapy and kilovoltage cone beam CT (CBCT) exposures were obtained from thermoluminescent dosimeter (TLD) measurements in an anthropomorphic phantom in which the positions of radiosensitive organs were delineated. Five treatment deliveries were investigated : (i) conventional tangential field whole breast radiotherapy (WBRT), (ii) non-coplanar conformal delivery applicable to accelerated partial beast irradiation (APBI), (iii) two-volume simultaneous integrated boost (SIB) treatment, (iv) forward planned three-volume SIB, (v) inverse-planned three volume SIB. Conformal and intensity modulated radiotherapy (IMRT) methods were used to plan the complex treatments. Techniques spanned the range from simple methods appropriate for patient cohorts with a low local cancer recurrence risk to complex plans relevant to cohorts with high recurrence risk. Delineated organs at risk included brain, salivary glands, thyroid, contra-lateral breast, left and right lung, oesophagus, stomach, liver, colon and bladder. Biological Effects of Ionising Radiation (BEIR) VII cancer incidence models were applied to the measured mean organ doses to determine Lifetime Attributable Risk (LAR) for ages at exposure from 35 to 80 years according to radiotherapy techniques, and included dose from the CBCT imaging.
All LAR decreased with age at exposure and were lowest for brain, thyroid, liver and bladder (< 0.1%). There was little dependence of LAR on radiotherapy technique for these organs and for colon and stomach. LAR values for the lungs for the three SIB techniques were two to three times those from WBRT and APBI. Uncertainties in the LAR models outweigh any differences in lung LAR between the SIB methods. Constraints in the planning of the SIB methods ensured that contra-lateral breast doses and LAR were comparable to WBRT, despite their added complexity. The smaller irradiated volume of the ABPI plan contributed to a halving of LAR for contralateral breast compared with the other plan types. Daily image guided radiotherapy (IGRT) for a left breast protocol using kilovoltage CBCT contributed <10% to LAR for the majority of organs, and did not exceed 22% of total organ dose.
Phantom measurements and calculations of LAR from the BEIR VII models predict that complex breast radiotherapy techniques do not increase the theoretical risk of second cancer incidence for organs distant from the treated breast, or the contralateral breast where appropriate plan constraints are applied. Complex SIB treatments are predicted to increase the risk of second cancer incidence in the lungs compared to standard whole breast radiotherapy ; this is outweighed by the threefold reduction in 5 year local recurrence risk for patients of high risk of recurrence, and young age, from the use of radiotherapy. APBI may have a favourable impact on risk of second cancer in the contra-lateral breast and lung for older patients at low risk of recurrence. Intensive use of IGRT increased the estimated values of LAR but these are dominated by the effect of the dose from the radiotherapy, and any increase in LAR from IGRT is much lower than the models’ uncertainties.
second cancer incidence; breast radiotherapy; image-guided radiotherapy; partial breast
The benefit of computer-assisted navigation depends on the registration process, at which patient features are correlated to some preoperative imagery. The operator-induced uncertainty in localizing patient features – the User Localization Error (ULE) - is unknown and most likely dominating the application accuracy. This initial feasibility study aims at providing first data for ULE with a research navigation system.
Active optical navigation was done in CT-images of a plastic skull, an anatomic specimen (both with implanted fiducials) and a volunteer with anatomical landmarks exclusively. Each object was registered ten times with 3, 5, 7, and 9 registration points. Measurements were taken at 10 (anatomic specimen and volunteer) and 11 targets (plastic skull). The active NDI Polaris system was used under ideal working conditions (tracking accuracy 0.23 mm root mean square, RMS; probe tip calibration was 0.18 mm RMS. Variances of tracking along the principal directions were measured as 0.18 mm2, 0.32 mm2, and 0.42 mm2. ULE was calculated from predicted application accuracy with isotropic and anisotropic models and from experimental variances, respectively.
The ULE was determined from the variances as 0.45 mm (plastic skull), 0.60 mm (anatomic specimen), and 4.96 mm (volunteer). The predicted application accuracy did not yield consistent values for the ULE.
Quantitative data of application accuracy could be tested against prediction models with iso- and anisotropic noise models and revealed some discrepancies. This could potentially be due to the facts that navigation and one prediction model wrongly assume isotropic noise (tracking is anisotropic), while the anisotropic noise prediction model assumes an anisotropic registration strategy (registration is isotropic in typical navigation systems). The ULE data are presumably the first quantitative values for the precision of localizing anatomical landmarks and implanted fiducials. Submillimetric localization is possible for implanted screws; anatomic landmarks are not suitable for high-precision clinical navigation.
application accuracy; navigation; human localization error; registration
We studied the performance of a dual-panel positron emission tomography (PET) camera dedicated to breast cancer imaging using Monte Carlo simulation. The PET camera under development has two 10 × 15 cm2 plates that are constructed from arrays of 1 × 1 × 3 mm3 LSO crystals coupled to novel ultra-thin (<200 μm) silicon position-sensitive avalanche photodiodes (PSAPD). In this design the photodetectors are configured “edge-on” with respect to incoming photons which encounter a minimum of 2 cm thick of LSO with directly measured photon interaction depth. Simulations predict that this camera will have 10–15% photon sensitivity, for an 8–4 cm panel separation. Detector measurements show ~1 mm3 intrinsic spatial resolution, <12% energy resolution, and ~2 ns coincidence time resolution. By performing simulated dual-panel PET studies using a phantom comprising active breast, heart, and torso tissue, count performance was studied as a function of coincident time and energy windows. We also studied visualization of hot spheres of 2.5–4.0 mm diameter and various locations within the simulated breast tissue for 1 × 1 × 3 mm3, 2 × 2 × 10 mm3, 3 × 3 × 30 mm3, and 4 × 4 × 20 mm3 LSO crystal resolutions and different panel separations. Images were reconstructed by focal plane tomography with attenuation and normalization corrections applied. Simulation results indicate that with an activity concentration ratio of tumor:breast:heart:torso of 10:1:10:1 and 30 s of acquisition time, only the dual-plate PET camera comprising 1 × 1 × 3 mm3 crystals could resolve 2.5 mm diameter spheres with an average peak-to-valley ratio of 1.3.
This study aims to demonstrate, using human cadavers the feasibility of energy-based adaptive focusing of ultrasonic waves using Magnetic Resonance Acoustic Radiation Force Imaging (MR-ARFI) in the framework of non-invasive transcranial High Intensity Focused Ultrasound (HIFU) therapy.
Energy-based adaptive focusing techniques were recently proposed in order to achieve aberration correction. We evaluate this method on a clinical brain HIFU system composed of 512 ultrasonic elements positioned inside a full body 1.5 T clinical Magnetic Resonance (MR) imaging system. Cadaver heads were mounted onto a clinical Leksell stereotactic frame. The ultrasonic wave intensity at the chosen location was indirectly estimated by the MR system measuring the local tissue displacement induced by the acoustic radiation force of the ultrasound (US) beams. For aberration correction, a set of spatially encoded ultrasonic waves was transmitted from the ultrasonic array and the resulting local displacements were estimated with the MR-ARFI sequence for each emitted beam. A non-iterative inversion process was then performed in order to estimate the spatial phase aberrations induced by the cadaver skull. The procedure was first evaluated and optimized in a calf brain using a numerical aberrator mimicking human skull aberrations. The full method was then demonstrated using a fresh human cadaver head.
The corrected beam resulting from the direct inversion process was found to focus at the targeted location with an acoustic intensity 2.2 times higher than the conventional non corrected beam. In addition, this corrected beam was found to give an acoustic intensity 1.5 times higher than the focusing pattern obtained with an aberration correction using transcranial acoustic simulation based on X-ray computed tomography (CT) scans.
The proposed technique achieved near optimal focusing in an intact human head for the first time. These findings confirm the strong potential of energy-based adaptive focusing of transcranial ultrasonic beams for clinical applications.
MRI; adaptive focusing; MR-ARFI; ultrasound transcranial therapy; HIFU
We have developed an analytic solution for spatially resolved diffuse reflectance within the δ-P1 approximation to the radiative transport equation for a semi-infinite homogeneous turbid medium. We evaluate the performance of this solution by comparing its predictions with those provided by Monte Carlo simulations and the standard diffusion approximation. We demonstrate that the δ-P1 approximation provides accurate estimates for spatially resolved diffuse reflectance in both low and high scattering media. We also develop a multi-stage nonlinear optimization algorithm in which the radiative transport estimates provided by the δ-P1 approximation are used to recover the optical absorption (μa), reduced scattering (
μs′), and single-scattering asymmetry coefficients (g1) of liquid and solid phantoms from experimental measurements of spatially resolved diffuse reflectance. Specifically, the δ-P1 approximation can be used to recover μa,
μs′, and g1 with errors within ±22%, ±18%, and ±17%, respectively, for both intralipid-based and siloxane-based tissue phantoms. These phantoms span the optical property range
4<(μs′/μa)<117. Using these same measurements, application of the standard diffusion approximation resulted in the recovery of μa and
μs′ with errors of ±29% and ±25%, respectively. Collectively, these results demonstrate that the δ-P1 approximation provides accurate radiative transport estimates that can be used to determine accurately the optical properties of biological tissues, particularly in spectral regions where tissue may display moderate/low ratios of reduced scattering to absorption (
Partial volume effects (PVE) are consequences of the limited spatial resolution in emission tomography leading to under-estimation of uptake in tissues of size similar to the point spread function (PSF) of the scanner as well as activity spillover between adjacent structures. Among PVE correction methodologies, a voxel-wise mutual multi-resolution analysis (MMA) was recently introduced. MMA is based on the extraction and transformation of high resolution details from an anatomical image (MR/CT) and their subsequent incorporation into a low resolution PET image using wavelet decompositions. Although this method allows creating PVE corrected images, it is based on a 2D global correlation model which may introduce artefacts in regions where no significant correlation exists between anatomical and functional details.
A new model was designed to overcome these two issues (2D only and global correlation) using a 3D wavelet decomposition process combined with a local analysis. The algorithm was evaluated on synthetic, simulated and patient images, and its performance was compared to the original approach as well as the geometric transfer matrix (GTM) method.
Quantitative performance was similar to the 2D global model and GTM in correlated cases. In cases where mismatches between anatomical and functional information were present the new model outperformed the 2D global approach, avoiding artefacts and significantly improving quality of the corrected images and their quantitative accuracy.
A new 3D local model was proposed for a voxel-wise PVE correction based on the original mutual multi-resolution analysis approach. Its evaluation demonstrated an improved and more robust qualitative and quantitative accuracy compared to the original MMA methodology, particularly in the absence of full correlation between anatomical and functional information.
Algorithms; Artifacts; Brain; radionuclide imaging; Humans; Imaging, Three-Dimensional; methods; Linear Models; Positron-Emission Tomography; methods; Reproducibility of Results; Whole Body Imaging; emission tomography; partial volume effects; resolution and intensity recovery; wavelet transform; multi-modality
Current respiratory motion monitoring devices used for motion synchronization in medical imaging and radiotherapy provide either 1D respiratory signals over a specific region or 3D information based on few external or internal markers. On the other hand, newer technology may offer the potential to monitor the entire patient external surface in real time. The main objective of this study was to assess the motion correlation between such an external patient surface and internal anatomical landmarks motion.
Four dimensional Computed Tomography (4D CT) volumes for ten patients were used in this study. Anatomical landmarks were manually selected in the thoracic region across the 4D CT datasets by two experts. The landmarks included normal structures as well as the tumour location. In addition, a distance map representing the entire external patient surface, which corresponds to surfaces acquired by a Time of Flight (ToF) camera or similar devices, was created by segmenting the skin of all 4D CT volumes using a thresholding algorithm. Finally, the correlation between the internal landmarks and external surface motion was evaluated for different regions (placement and size) throughout a patient’s surface.
Significant variability was observed in the motion of the different parts of the external patient surface. The larger motion magnitude was consistently measured in the central regions of the abdominal and the thoracic areas for the different patient datasets considered. The highest correlation coefficients were observed between the motion of these external surface areas and internal landmarks such as the diaphragm and mediastinum structures as well as the tumour location landmarks (0.8 ± 0.18 and 0.72 ± 0.12 for the abdominal and the thoracic regions respectively). Worse correlation was observed when one considered landmarks not significantly influenced by respiratory motion such as the apex and the sternum.
Discussion and conclusions
There were large differences in the motion correlation observed considering different regions of interest placed over a patients’ external surface and internal anatomical landmarks. The positioning of current devices used for respiratory motion synchronization may reduce such correlation by averaging the motion over correlated and poorly correlated external regions. The potential of capturing in real-time the motion of the complete external patient surface as well as choosing the area of the surface that correlates best with the internal motion should allow reducing such variability and associated errors in both respiratory motion synchronization and subsequent motion modeling processes.
Fiducial Markers; Four-Dimensional Computed Tomography; standards; Humans; Movement; Respiration; respiratory motion; patient external surface; 4D CT
Pressure ulcers are areas of soft tissue breakdown induced by a sustained mechanical stress that damages the skin and underlying tissues. They represent a considerable burden to the society in terms of health care and cost. Yet, techniques for prevention and detection of pressure ulcers still remain very limited. In this article, the authors investigated the potential of ultrasound elastography for pressure ulcer early detection. Elastography is an imaging technique providing local information on biological tissue mechanical properties. It is relevant for pressure ulcer detection as this pathology is associated with a gradual stiffening of damaged tissues, beginning in the deeper tissues and progressing toward the skin surface.
A 2D ultrasound elastography method was proposed and its ability in terms of pressure ulcer detection was validated through numerical simulations and physical acquisitions on pressure ulcer mimicking phantoms. In vivo experiments on a rat model are also reported. A maintained pressure was applied on the animal thigh, with a view to generate a pressure ulcer, and ultrasound data were acquired and processed before and after application of this pressure.
Numerical simulations demonstrated that a pressure ulcer can theoretically be detected at a very early stage with ultrasound elastography. Even when the ulcer region was characterized by a low stiffening (ratio of 1.8 relative to normal tissues), the corresponding elastogram clearly underlined the pathological area. This observation was confirmed by the results obtained on a physical phantom mimicking a pressure ulcer at an early stage. Computed elastograms showed strain differences between areas mimicking healthy and pathological tissues. Results corresponding to in vivo experiments revealed a difference in the way tissues behaved before and after the pressure was applied on the animal thigh, which strongly suggests the presence of a pathological area.
Experiments demonstrated that ultrasound elastography is a promising technique for pressure ulcer detection, especially at an early stage of the pathology, when the disease is still visually undetectable. In the absence of any gold standard method, this is also a first step toward the development of a quantitative technique.
pressure ulcer; ultrasound; elastography; medical imaging
In this article, the authors propose a new gold standard data set for the validation of two-dimensional/three-dimensional (2D/3D) and 3D/3D image registration algorithms.
A gold standard data set was produced using a fresh cadaver pig head with attached fiducial markers. The authors used several imaging modalities common in diagnostic imaging or radiotherapy, which include 64-slice computed tomography (CT), magnetic resonance imaging using Tl, T2, and proton density sequences, and cone beam CT imaging data. Radiographic data were acquired using kilovoltage and megavoltage imaging techniques. The image information reflects both anatomy and reliable fiducial marker information and improves over existing data sets by the level of anatomical detail, image data quality, and soft-tissue content. The markers on the 3D and 2D image data were segmented using analyze 10.0 (AnalyzeDirect, Inc., Kansas City, KN) and an in-house software.
The projection distance errors and the expected target registration errors over all the image data sets were found to be less than 2.71 and 1.88 mm, respectively.
The gold standard data set, obtained with state-of-the-art imaging technology, has the potential to improve the validation of 2D/3D and 3D/3D registration algorithms for image guided therapy.
gold standard; image registration; image guidance; radiotherapy
A new gold standard data set for validation of 2D/3D registration based on a porcine cadaver head with attached fiducial markers was presented in the first part of this article. The advantage of this new phantom is the large amount of soft tissue, which simulates realistic conditions for registration. This article tests the performance of intensity- and gradient-based algorithms for 2D/3D registration using the new phantom data set.
Intensity-based methods with four merit functions, namely, cross correlation, rank correlation, correlation ratio, and mutual information (MI), and two gradient-based algorithms, the backprojection gradient-based (BGB) registration method and the reconstruction gradient-based (RGB) registration method, were compared. Four volumes consisting of CBCT with two fields of view, 64 slice multidetector CT, and magnetic resonance-T1 weighted images were registered to a pair of kV x-ray images and a pair of MV images. A standardized evaluation methodology was employed. Targets were evenly spread over the volumes and 250 starting positions of the 3D volumes with initial displacements of up to 25 mm from the gold standard position were calculated. After the registration, the displacement from the gold standard was retrieved and the root mean square (RMS), mean, and standard deviation mean target registration errors (mTREs) over 250 registrations were derived. Additionally, the following merit properties were computed: Accuracy, capture range, number of minima, risk of nonconvergence, and distinctiveness of optimum for better comparison of the robustness of each merit.
Among the merit functions used for the intensity-based method, MI reached the best accuracy with an RMS mTRE down to 1.30 mm. Furthermore, it was the only merit function that could accurately register the CT to the kV x rays with the presence of tissue deformation. As for the gradient-based methods, BGB and RGB methods achieved subvoxel accuracy (RMS mTRE down to 0.56 and 0.70 mm, respectively). Overall, gradient-based similarity measures were found to be substantially more accurate than intensity-based methods and could cope with soft tissue deformation and enabled also accurate registrations of the MR-T1 volume to the kV x-ray image.
In this article, the authors demonstrate the usefulness of a new phantom image data set for the evaluation of 2D/3D registration methods, which featured soft tissue deformation. The author’s evaluation shows that gradient-based methods are more accurate than intensity-based methods, especially when soft tissue deformation is present. However, the current nonoptimized implementations make them prohibitively slow for practical applications. On the other hand, the speed of the intensity-based method renders these more suitable for clinical use, while the accuracy is still competitive.
2D/3D registration; image-guided therapy; gradient; intensity; rendering; ray-casting
In our previous study we used the linear-quadratic model [J. Nucl. Med. 35, 1861 (1994)] to confirm our initial finding, based on the time-dose-fractionation model [J. Nucl. Med. 34, 1801 (1993)], that longer-lived radionuclides (e.g., 32P, 91Y) can offer a substantial therapeutic advantage over the shorter-lived radionuclides presently used in radioimmunotherapy (e.g., 90Y). The original calculations using the linear-quadratic (LQ) model did not account for proliferation of the tumor and critical bone marrow tissues. It has been suggested that inclusion of a proliferation term in the LQ model can have a substantial impact on the biologically effective dose (BED). With this in mind, we have reexamined the therapeutic efficacy of longer versus short-lived radionuclides using the LQ model replete with proliferation terms for tumor and bone marrow. Relative advantage factors (RAF), which quantify the overall therapeutic advantage of a long-lived compared to short-lived radionuclide, were calculated accordingly. While the extrapolated initial dose rate required to achieve a given BED can be affected by the inclusion of proliferation terms for both the tumor and marrow, the relative advantage factors for the longer-lived radionuclides were not significantly affected. Longer-lived radionuclides such as 114mIn and 91Y are about three times more therapeutically effective than the shorter-lived 90Y which is currently used in RIT. In other words, for a given therapeutic effect in the tumor, a longer-lived radionuclide can result in a lower deleterious effect to the bone marrow than a short-lived radionuclide. Given that bone marrow is generally considered to be the dose-limiting organ, these results have important implications for radioimmunotherapy.
linear-quadratic model; radioimmunotherapy; dosimetry; proliferation
We are developing new techniques to improve the accuracy of computerized microcalcification detection by using the joint two-view information on craniocaudal (CC) and mediolateral-oblique (MLO) views. After cluster candidates were detected using a single-view detection technique, candidates on CC and MLO views were paired using their radial distances from the nipple. Candidate pairs were classified with a similarity classifier that used the joint information from both views. Each cluster candidate was also characterized by its single-view features. The outputs of the similarity classifier and the single-view classifier were fused and the cluster candidate was classified as a true microcalcification cluster or a false-positive (FP) using the fused two-view information. A data set of 116 pairs of mammograms containing microcalcification clusters and 203 pairs of normal images from the University of South Florida (USF) public database was used for training the two-view detection algorithm. The trained method was tested on an independent test set of 167 pairs of mammograms, which contained 71 normal pairs and 96 pairs with microcalcification clusters collected at the University of Michigan (UM). The similarity classifier had a very low FP rate for the test set at low and medium levels of sensitivity. However, the highest mammogram-based sensitivity that could be reached by the similarity classifier was 69%. The single-view classifier had a higher FP rate compared to the similarity classifier, but it could reach a maximum mammogram-based sensitivity of 93%. The fusion method combined the scores of these two classifiers so that the number of FPs was substantially reduced at relatively low and medium sensitivities, and a relatively high maximum sensitivity was maintained. For the malignant microcalcification clusters, at a mammogram-based sensitivity of 80%, the FP rates were 0.18 and 0.35 with the two-view fusion and single-view detection methods, respectively. When the training and test sets were switched, a similar improvement was obtained, except that both the fusion and single-view detection methods had superior test performances on the USF data set than those on the UM data set. Our results indicate that correspondence of cluster candidates on two different views provides valuable additional information for distinguishing FPs from true microcalcification clusters.
computer-aided diagnosis; microcalcification clusters; segmentation
The purpose of this work was to calculate radiation dose and its organ
distribution in a realistic mouse phantom from micro-computed tomography
(microCT) imaging protocols. CT dose was calculated using GATE and a voxelized,
realistic phantom. The x-ray photon energy spectra used in simulations were
precalculated with GATE and validated against previously published data. The
number of photons required per simulated experiments was determined by direct
exposure measurements. Simulated experiments were performed for three types of
beams and two types of mouse beds. Dose-volume histograms and dose percentiles
were calculated for each organ. For a typical microCT screening examination with
a reconstruction voxel size of 200 μm, the average
whole body dose varied from 80 mGy (at 80 kVp) to 160 mGy (at 50 kVp), showing a
strong dependence on beam hardness. The average dose to the bone marrow is close
to the soft tissue average. However, due to dose nonuniformity and higher
radiation sensitivity, 5% of the marrow would receive an effective dose
about four times higher than the average. If CT is performed longitudinally, a
significant radiation dose can be given. The total absorbed radiation dose is a
function of milliamperes-second, beam hardness, and desired image quality
(resolution, noise and contrast). To reduce dose, it would be advisable to use
the hardest beam possible while maintaining an acceptable contrast in the
The purpose of this study was to calculate internal absorbed dose distribution in mice from preclinical small animal PET imaging procedures with fluorine-18 labeled compounds (18FDG, 18FLT, and fluoride ion). The GATE Monte Carlo software and a realistic, voxel-based mouse phantom that included a subcutaneous tumor were used to perform simulations. Discretized time-activity curves obtained from dynamic in vivo studies with each of the compounds were used to set the activity concentration in the simulations. For 18FDG, a realistic range of uptake ratios was considered for the heart and tumor. For each simulated time frame, the biodistribution of the radionuclide in the phantom was considered constant, and a sufficient number of decays were simulated to achieve low statistical uncertainty. Absorbed dose, which was scaled to take into account radioactive decay, integration with time, and changes in biological distribution was reported in mGy per MBq of administered activity for several organs and uptake scenarios. The mean absorbed dose ranged from a few mGy/MBq to hundreds of mGy/MBq. Major organs receive an absorbed dose in a range for which biological effects have been reported. The effects on a given investigation are hard to predict; however, investigators should be aware of potential perturbations especially when the studied organ receives high absorbed dose and when longitudinal imaging protocols are considered.
small animal PET; dosimetry; GATE; Flourine-18
In dosimetry-based treatment planning protocols, patients with rapid clearance of the radiopharmaceutical require a larger amount of initial activity than those with slow clearance to match the absorbed dose to the critical organ. As a result, the dose-rate to the critical organ is higher in patients with rapid clearance and may cause unexpected toxicity compared to patients with slow clearance. In order to account for the biological impact of different dose-rates, radiobiological modeling is beginning to be applied to the analysis of radionuclide therapy patient data. To date, the formalism used for these analyses is based on kinetics derived from activity in a single organ, the target. This does not include the influence of other source organs to the dose and dose-rate to the target organ. As a result, only self-dose irradiation in the target organ contributes to the dose-rate. In this work, the biological effective dose (BED) formalism has been extended to include the effect of multiple source organ contributions to the net dose-rate in a target organ. The generalized BED derivation has been based on the Medical Internal Radionuclide Dose Committee (MIRD) schema assuming multiple source organs following exponential effective clearance of the radionuclide. A BED-based approach to determine the largest safe dose to critical organs has also been developed. The extended BED formalism is applied to red marrow dosimetry, as well as kidney dosimetry considering the cortex and the medulla separately, since both those organs are commonly dose limiting in radionuclide therapy. The analysis shows that because the red marrow is an early responding tissue (high α/β), it is less susceptible to unexpected toxicity arising from rapid clearance of high levels of administered activity in the marrow or in the remainder of the body. In kidney dosimetry, the study demonstrates a complex interplay between clearance of activity in the cortex and the medulla, as well as the initial activity ratio and the S value ratio between the two. In some scenarios, projected BED based on both the cortex and the medulla is a more appropriate constraint on the administered activity than the BED based on the cortex only. Furthermore, different fractionated regimens were considered to reduce renal toxicity. The MIRD-based BED formalism is expected to be useful for patient-specific adjustments of activity and to facilitate the investigation of dose-toxicity correlations with respect to dose-rate and tissue repair mechanism.
dosimetry; radionuclide therapy; biological effective dose
Hyperpolarized (HP) 3He MRI is an emerging tool in the diagnosis and evaluation of pulmonary diseases involving bronchoconstriction, such as asthma. Previously, airway diameters from dynamic HP 3He MR images of the lung were assessed manually and subjectively, and were thus prone to uncertainties associated with human error and partial volume effects. A model-based algorithm capable of fully utilizing pixel intensity profile information and attaining subpixel resolution has been developed to measure surrogate airway diameters from HP 3He MR static projection images of plastic tubes. This goal was achieved by fitting ideal pixel intensity profiles for various diameter (6.4 to 19.1 mm) circular tubes to actual pixel intensity data. A phantom was constructed from plastic tubes of various diameters connected in series and filled with water mixed with contrast agent. Projection MR images were then taken of the phantom. The favorable performance of the model-based algorithm compared to manual assessment demonstrates the viability of our approach. The manual and algorithm approaches yielded diameter measurements that generally stayed within 1× the pixel dimension. However, inconsistency of the manual approach can be observed from the larger standard deviations of its measured values. The method was then extended to HP 3He MRI, producing encouraging results at tube diameters characteristic of airways beyond the second generation, thereby justifying their application to lung airway imaging and measurement. Potential obstacles when measuring airway diameters using this method are discussed.
Megavoltage cone-beam computed tomography (MY CBCT) is a highly promising technique for providing volumetric patient position information in the radiation treatment room. Such information has the potential to greatly assist in registering the patient to the planned treatment position, helping to ensure accurate delivery of the high energy therapy beam to the tumor volume while sparing the surrounding normal tissues. Presently, CBCT systems using conventional MV active matrix flat-panel imagers (AMFPIs), which are commonly used in portal imaging, require a relatively large amount of dose to create images that are clinically useful. This is due to the fact that the phosphor screen detector employed in conventional MV AMFPIs utilizes only ~2% of the incident radiation (for a 6 MV x-ray spectrum). Fortunately, thick, segmented scintillating detectors can overcome this limitation, and the first prototype imager has demonstrated highly promising performance for projection imaging at low doses. It is therefore of definite interest to examine the potential performance of such thick, segmented scintillating detectors for MV CBCT. In this study, Monte Carlo simulations of radiation energy deposition were used to examine reconstructed images of cylindrical CT contrast phantoms, embedded with tissue-equivalent objects. The phantoms were scanned at 6 MV using segmented detectors having various design parameters (i.e., detector thickness, as well as scintillator and septal wall materials). Due to constraints imposed by the nature of this study, the size of the phantoms was limited to ~6 cm. For such phantoms, the simulation results suggest that a 40 mm thick, segmented CsI detector with low density septal walls can delineate electron density differences of ~2.3% and 1.3% at doses of 1.54 and 3.08 cGy, respectively. In addition, it was found that segmented detectors with greater thickness, higher density scintillator material, or lower density septal walls exhibit higher contrast-to-noise performance. Finally, the performance of various segmented detectors obtained at a relatively low dose (1.54 cGy) was compared to that of a phosphor screen similar to that employed in conventional MV AMFPIs. This comparison indicates that, for a phosphor screen to achieve the same contrast-to-noise performance as the segmented detectors, ~18 to 59 times more dose is required, depending on the configuration of the segmented detectors.
active matrix flat-panel imager; segmented scintillating detectors; megavoltage cone-beam CT; soft tissue visualization; contrast-to-noise ratio
The purpose of this study is to build and test a support vector machine (SVM) model to predict for the occurrence of lung radiation-induced Grade 2+ pneumonitis. SVM is a sophisticated statistical technique capable of separating the two categories of patients (with/without pneumonitis) using a boundary defined by a complex hypersurface. Despite the complexity, the SVM boundary is only minimally influenced by outliers that are difficult to separate. By contrast, the simple hyperplane boundary computed by the more commonly used and related linear discriminant analysis method is heavily influenced by outliers. Two SVM models were built using data from 219 patients with lung cancer treated using radiotherapy (34 diagnosed with pneumonitis). One model (SVMall) selected input features from all dose and non-dose factors. For comparison, the other model (SVMdose) selected input features only from lung dose-volume factors. Model predictive ability was evaluated using ten-fold cross-validation and receiver operating characteristics (ROC) analysis. For the model SVMall, the area under the cross-validated ROC curve was 0.76 (sensitivity/specificity =74%/75%). Compared to the corresponding SVMdose area of 0.71 (sensitivity/specificity =68%/68%), the predictive ability of SVMall was improved, indicating that non-dose features are important contributors to separating patients with and without pneumonitis. Among the input features selected by model SVMall, the two with highest importance for predicting lung pneumonitis were: (a) generalized equivalent uniform doses close to the mean lung dose, and (b) chemotherapy prior to radiotherapy. The model SVMall is publicly available via internet access.
support vector machines; radiation pneumonitis; modeling; prediction
A theoretical model is described for application in ultrasonic tissue characterization using a calibrated 2-D spectrum analysis method. This model relates 2-D spectra computed from ultrasonic backscatter signals to intrinsic physical properties of tissue microstructures, e.g., size, shape, and acoustic impedance. The model is applicable to most clinical diagnostic ultrasound systems. Two experiments employing two types of tissue architectures, spherical and cylindrical scatterers, are conducted using ultrasound with center frequencies of 10 and 40 MHz, respectively. Measurements of a tissue-mimicking phantom with an internal suspension of microscopic glass beads are used to validate the theoretical model. Results from in vitro muscle fibers are presented to further elucidate the utility of 2-D spectrum analysis in ultrasonic tissue characterization.
ultrasonic tissue characterization (UTC); tissue model; 2-D power spectrum; spectral parameter
Susceptibility effects are a very efficient source of contrast in Magnetic Resonance Imaging. However, detection is hampered by the fact the induced contrast is negative. In this work, the SIgnal Response MApping to dephaser (SIRMA) method is proposed to map susceptibility gradient to improve visualisation.
In conventional gradient echo acquisitions, the echo formation of susceptibility affected spins is shifted in k-space, the shift being proportional to the susceptibility gradient. Susceptibility gradients map can be produced by measuring this induced shifts. The SIRMA method measures these shifts from a series of dephased images collected with additional incremental dephasers. These additional dephasers correspond either to a slice refocusing gradient offset or to a reconstruction window off-centering. The signal intensity profile as a function of the additional dephaser, was determined on a pixel-by-pixel basis from the ensemble of dephased images. Susceptibility affected voxels presented a signal response profile maximum shifted compared to non-affected voxels ones. Shift magnitude and sign were measured for each pixel to determine susceptibility gradients and produce susceptibility gradient map.
In vitro experiments demonstrated the ability of the method to map gradient inhomogeneities induced by a cylinder. Quantization accuracy was evaluated comparing SIRMA images and simulations performed on the well-characterized air filled cylinder model. Performances of the SIRMA method, evaluated in vitro on cylinders filled with various superparamagnetic iron oxide SPIO concentrations, showed limited influence of acquisition parameters. Robustness of the method was then assessed in vivo after an infusion of SPIO-loaded nanocapsules into the rat brain using a convection-enhanced drug delivery approach. The region of massive susceptibility gradient induced by the SPIO loaded nanocapsules was clearly delineated on SIRMA maps and images were compared to T2* weighted images, Susceptibility Gradient Map and histological Perl’s staining slice. The potential for quantitative evaluation of SPIO distribution volume was demonstrated.
The proposed method is a promising technique for a wide range of applications especially in molecular or cellular imaging with respect to its quantitative nature and its computational simplicity.
Algorithms; Animals; Brain; anatomy & histology; Female; Image Enhancement; methods; Image Interpretation, Computer-Assisted; methods; Magnetic Resonance Imaging; methods; Rats; Rats, Sprague-Dawley; Reproducibility of Results; Sensitivity and Specificity; MRI; SPIO; susceptibility; positive imaging; molecular imaging
Correlation of information from multiple-view mammograms (e.g., MLO and CC views, bilateral views, or current and prior mammograms) can improve the performance of breast cancer diagnosis by radiologists or by computer. The nipple is a reliable and stable landmark on mammograms for the registration of multiple mammograms. However, accurate identification of nipple location on mammograms is challenging because of the variations in image quality and in the nipple projections, resulting in some nipples being nearly invisible on the mammograms. In this study, we developed a computerized method to automatically identify the nipple location on digitized mammograms. First, the breast boundary was obtained using a gradient-based boundary tracking algorithm, and then the gray level profiles along the inside and outside of the boundary were identified. A geometric convergence analysis was used to limit the nipple search to a region of the breast boundary. A two-stage nipple detection method was developed to identify the nipple location using the gray level information around the nipple, the geometric characteristics of nipple shapes, and the texture features of glandular tissue or ducts which converge toward the nipple. At the first stage, a rule-based method was designed to identify the nipple location by detecting significant changes of intensity along the gray level profiles inside and outside the breast boundary and the changes in the boundary direction. At the second stage, a texture orientation-field analysis was developed to estimate the nipple location based on the convergence of the texture pattern of glandular tissue or ducts towards the nipple. The nipple location was finally determined from the detected nipple candidates by a rule-based confidence analysis. In this study, 377 and 367 randomly selected digitized mammograms were used for training and testing the nipple detection algorithm, respectively. Two experienced radiologists identified the nipple locations which were used as the gold standard. In the training data set, 301 nipples were positively identified and were referred to as visible nipples. Seventy six nipples could not be positively identified and were referred to as invisible nipples. The radiologists provided their estimation of the nipple locations in the latter group for comparison with the computer estimates. The computerized method could detect 89.37% (269/301) of the visible nipples and 69.74% (53/76) of the invisible nipples within 1 cm of the gold standard. In the test data set, 298 and 69 of the nipples were classified as visible and invisible, respectively. 92.28% (275/298) of the visible nipples and 53.62% (37/69) of the invisible nipples were identified within 1 cm of the gold standard. The results demonstrate that the nipple locations on digitized mammograms can be accurately detected if they are visible and can be reasonably estimated if they are invisible. Automated nipple detection will be an important step towards multiple image analysis for CAD.
computer-aided detection; mammography; nipple detection; texture orientation field analysis
Mammography is the only technique currently used for detecting microcalcification (MC) clusters, an early indicator of breast cancer. However, mammographic images superimpose a three-dimensional compressed breast image onto two-dimensional projection views, resulting in overlapped anatomical breast structures that may obscure the detection and visualization of MCs. One possible solution to this problem is the use of cone beam computed tomography (CBCT) with a flat-panel (FP) digital detector. Although feasibility studies of CBCT techniques for breast imaging have yielded promising results, they have not shown how radiation dose and x-ray tube voltage affect the accuracy with which MCs are detected by CBCT experimentally. We therefore conducted a phantom study using FP-based CBCT system with various mean glandular doses and kVp values. An experimental CBCT scanner was constructed with a data-acquisition rate of 7.5 frames/s. 10.5- and 14.5cm-diameter breast phantoms made of gelatin were used to simulate uncompressed breasts consisting of 100% glandular tissue. Eight different MC sizes of calcium carbonate grains, ranging from 180–200 µm to 355–425 µm, were used to simulate MCs. MCs of the same size were arranged to form a 5×5 MC cluster and embedded in the breast phantoms. These MC clusters were positioned at 2.8 cm away from the center of the breast phantoms. The phantoms were imaged at 60, 80, and 100 kVp. With a single scan (360 degrees), 300 projection images were acquired with 0.5×, 1×, and 2× mean glandular dose limit for 10.5-cm phantom and with 1×, 2×, and 4× for 14.5-cm phantom. Feldkamp algorithm with a pure ramp filter was used for image reconstruction. The normalized noise level was calculated for each x-ray tube voltage and dose level. The image quality of CBCT images was evaluated by counting the number of visible MCs for each MC cluster for various conditions. The average percentage of the visible MCs were computed and plotted as a function of the MGD, the kVp, and the average MC size. The results show that the MC visibility increased with the MGD significantly but decreased with the breast size. The results also show the x-ray tube voltage affects the detection of MCs under different circumstances. With a 50% threshold, the minimum detectable MC sizes for the 10.5-cm phantom were 348 (±2), 288 (±7), 257 (±2) µm at 3, 6, and 12 mGy respectively. Those for the 14.5-cm phantom were 355 (±1), 307 (±7), 275 (±5) µm at 6, 12, and 24 mGy, respectively. With a 75% threshold, the minimum detectable MC sizes for the 10.5-cm phantom were 367 (±1), 316 (±7), 265 (±3) µm at 3, 6, and 12 mGy, respectively. Those for the 14.5-cm phantom were 377 (±3), 334 (±5), 300 (±2) µm at 6, 12, and 24 mGy, respectively.
cone-beam computed tomography; breast imaging; flat-panel detector; microcalcifications; mean glandular dose
The estimation of organ residence time is essential for high-dose myeloablative regimens in radioimmunotherapy (RIT). Frequently, this estimation is based on a series of simple planar scans and planar processing. The authors previously performed a simulation study which demonstrated that the accuracy of this methodology is limited compared to a hybrid planar/SPECT residence time estimation method. In this work the authors applied this hybrid method to data from a clinical trial of high-dose myeloablative yttrium-90 ibritumomab tiuxetan therapy. Image data acquired from 18 patients were comprised of planar scans at five time points ranging from 1 to 144 h postinjection and abdominal and thoracic SPECT/CT scans obtained at 24 h postinjection. The simple planar processing method used in this work was based on the geometric mean method with energy window based scatter compensation. No explicit background subtraction nor object or source thickness corrections were performed. The SPECT projections were reconstructed using iterative reconstruction with compensations for attenuation, scatter, and full collimator-detector response. Large differences were observed when residence times were estimated using the simple planar method compared to the hybrid method. The differences were not constant but varied in magnitude and sign. For the dose-limiting organ (liver), the average difference was −18% and variation in the difference was 19%, similar to the differences observed in a previously reported simulation study. The authors also looked at the relationship between the weight of the patient and the liver residence time and found that there was no meaningful correlation for either method. This indicates that weight would not be an adequate proxy for an experimental estimate of residence time when choosing the activity to administer for therapy. The authors conclude that methods such as the simple planar method used here are inadequate for RIT treatment planning. More sophisticated methods, such as the hybrid SPECT/planar method investigated here, are likely to be better predictors of organ dose and, as a result, organ toxicities.
absolute quantitation; quantitative SPECT; radioimmunotherapy
This work is intended to investigate the spatial resolution properties in cone beam CT by estimating the point spread functions (PSFs) in the reconstructed 3D images through simulation. The point objects were modeled as 3D delta functions. Their projections onto the detector plane were analytically derived and blurred with 2D PSFs estimated and used to represent the detector and focal spot blurring effects. The 2D PSF for detector blurring was computed from the line spread function measured for a typical a-Si/CsI flat panel detector used for general radiography. The focal spot blurring effect was simulated for an x-ray source with a nominal focal spot size of 0.6 mm and 1.33× magnification at the rotating center. Projection images were computed and sampled with an interval significantly smaller than the detector pixel size to avoid aliasing. Images were reconstructed using the Feldkamp algorithm with the five different filter functions. Reconstructed PSFs were plotted and analyzed to investigate the effects of detector blurring alone, focal spot blurring alone, or a combination of the two on the PSFs and their variations with the radial distance and z-level. Effects of binning and reconstruction filters were also studied. Our results show that the PSFs due to detector blurring are largely symmetric and vary little with the locations of the point objects. With focal spot blurring only or added to detector blurring, the PSFs along the rotation axis were largely symmetric but became increasingly asymmetric as the point objects were moved away from the rotation axis. The PSFs were found to become wider in the axial (anode to cathode) direction as the objects were moved toward the cathode side. The 3D PSFs may be approximated by an ellipsoid with three different axial lengths. They were found to point upright along the rotating axis but tilt toward the rotating axis as the point object was moved away from the axis.
cone beam CT; spatial resolution; point spread function; detector blurring; focal spot blurring; Feldkamp reconstruction algorithm
In this article, the authors evaluate a merit function for 2D/3D registration called stochastic rank correlation (SRC). SRC is characterized by the fact that differences in image intensity do not influence the registration result; it therefore combines the numerical advantages of cross correlation (CC)-type merit functions with the flexibility of mutual-information-type merit functions. The basic idea is that registration is achieved on a random subset of the image, which allows for an efficient computation of Spearman’s rank correlation coefficient. This measure is, by nature, invariant to monotonic intensity transforms in the images under comparison, which renders it an ideal solution for intramodal images acquired at different energy levels as encountered in intrafractional kV imaging in image-guided radiotherapy. Initial evaluation was undertaken using a 2D/3D registration reference image dataset of a cadaver spine. Even with no radiometric calibration, SRC shows a significant improvement in robustness and stability compared to CC. Pattern intensity, another merit function that was evaluated for comparison, gave rather poor results due to its limited convergence range. The time required for SRC with 5% image content compares well to the other merit functions; increasing the image content does not significantly influence the algorithm accuracy. The authors conclude that SRC is a promising measure for 2D/3D registration in IGRT and image-guided therapy in general.