[Show abstract][Hide abstract] ABSTRACT: To present a method for the dosimetric analysis of permanent prostate brachytherapy implants using a combination of stereoscopic X-ray radiography and magnetic resonance (MR) imaging (XMR) in an XMR facility, and to compare the clinical results between XMR- and computed tomography (CT)-based dosimetry.
Patients who had received nonstranded iodine-125 permanent prostate brachytherapy implants underwent XMR and CT imaging 4 weeks later. Four observers outlined the prostate gland on both sets of images. Dose-volume histograms (DVHs) were derived, and agreement was compared among the observers and between the modalities.
A total of 30 patients were evaluated. Inherent XMR registration based on prior calibration and optical tracking required a further automatic seed registration step that revealed a median root mean square registration error of 4.2 mm (range, 1.6-11.4). The observers agreed significantly more closely on prostate base and apex positions as well as outlining contours on the MR images than on those from CT. Coefficients of variation were significantly higher for observed prostate volumes, D90, and V100 parameters on CT-based dosimetry as opposed to XMR. The XMR-based dosimetry showed little agreement with that from CT for all observers, with D90 95% limits of agreement ranges of 65, 118, 79, and 73 Gy for Observers 1, 2, 3, and 4, respectively.
The study results showed that XMR-based dosimetry offers an alternative to other imaging modalities and registration methods with the advantages of MR-based prostate delineation and confident three-dimensional reconstruction of the implant. The XMR-derived dose-volume histograms differ from the CT-derived values and demonstrate less interobserver variability.
International Journal of Radiation OncologyBiologyPhysics 09/2008; 71(5):1518-25. DOI:10.1016/j.ijrobp.2008.03.065 · 4.26 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: The three-year UK NeuroGrid project aims to develop a Grid-based collaborative research environment to support the data and compute needs for a UK Neurosciences community. This paper describes the challenges in developing this architecture and details initial results from the development of its first prototype to support psychosis, dementia and stroke research and the social challenges of such a collaborative research project. The paper discusses approaches being taken to explore the collaborative science process to inform the requirements for follow on prototypes and methods utilized to develop an effective project team.
Proceedings of the IEEE Symposium on Computer-Based Medical Systems 06/2006; DOI:10.1109/CBMS.2006.156
[Show abstract][Hide abstract] ABSTRACT: The paper draws on a number of Grid projects, particularly on the experience of NeuroGrid, a UK project in the Neurosciences tasked with developing a Grid-based collaborative research environment to support the sharing of digital images and patient data across multiple distributed sites. It outlines recurrent socio-technical issues, highlighting the challenges of scaling up technological networks in advance of the regulatory networks which normally regulate their use in practice.
Studies in health technology and informatics 02/2006; 120:336-47.
[Show abstract][Hide abstract] ABSTRACT: The European Union (EU) has recently introduced a directive that aims to protect workers from adverse effects of exposure to electromagnetic fields. All countries within the EU are required to incorporate this directive into their national law by 2008. This legislation applies to all types of occupational exposure to electromagnetic fields with frequencies from 0 Hz to 300 GHz. It has dramatic implications for interventional magnetic resonance (MR) imaging, because workers who are close to the MR scanner while scanning is taking place are exposed at levels substantially above the exposure limits. This is especially the case for time-varying magnetic fields in the 110 Hz-5 kHz range, which includes the field from the imaging gradients. In this article, the scientific evidence on which the limits are based is brought into question. It is concluded that an urgent review of the directive is required and that more research, especially at MR gradient frequencies, is needed.
[Show abstract][Hide abstract] ABSTRACT: In magnetic resonance imaging (MRI), there is always a drive toward reducing the acquisition time. In volume imaging, time is often spent in acquiring data where there exists no signal because the imaging volume is larger than the object. In this paper, a method is presented for scan time reduction using an adaptive field of view (FOV). Multislice images are acquired with the FOV in the phase encoding direction of each slice determined by measurements made on the initial localization survey scan. Depending on the region of interest, an optimized FOV is also determined so that scan time is reduced in comparison to a normal scan while improving image resolution. The method is simple to implement and requires no additional hardware. Typical reductions in scan time are on the order 9-14%.
Magnetic Resonance Imaging 02/2005; 23(1):47-52. DOI:10.1016/j.mri.2004.09.007 · 2.09 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: There is considerable interest in performing free-breathing acquisitions of the heart in order to obtain high-quality images without the need for multiple, long breathholds. In this article a 3D motion-correction method is described that is based on image registration of in-plane data and through-plane slice tracking. A number of fast radial undersampled images are acquired, each of which is free of motion artifacts. Initially, in-plane translational and rotational motion between each image was corrected before combining the data to give a fully sampled image. At the next stage, correction of in-plane deformation, in addition to translations and rotations, was performed in the image domain. Through-plane translational motion was compensated using a navigator echo to move the acquisition plane. Using this method, information on the motion of the heart was captured at the same time as acquiring the image data. No motion model, assumptions about the motion, or training data are required. The method is demonstrated on phantom data and cardiac images acquired on free-breathing volunteers.
Magnetic Resonance in Medicine 11/2004; 52(5):1127-35. DOI:10.1002/mrm.20252 · 3.57 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: This paper describes a quantitative assessment of respiratory motion of the heart and the construction of a model of respiratory motion. Three-dimensional magnetic resonance scans were acquired on eight normal volunteers and ten patients. The volunteers were imaged at multiple positions in the breathing cycle between full exhalation and full inhalation while holding their breath. The exhalation volume was segmented and used as a template to which the other volumes were registered using an intensity-based rigid registration algorithm followed by nonrigid registration. The patients were imaged at inhale and exhale only. The registration results were validated by visual assessment and consistency measurements indicating subvoxel registration accuracy. For all subjects, we assessed the nonrigid motion of the heart at the right coronary artery, right atrium, and left ventricle. We show that the rigid-body motion of the heart is primarily in the craniocaudal direction with smaller displacements in the right-left and anterior-posterior directions; this is in agreement with previous studies. Deformation was greatest for the free wall of the right atrium and the left ventricle; typical deformations were 3-4 mm with deformations of up to 7 mm observed in some subjects. Using the registration results, landmarks on the template surface were mapped to their correct positions through the breathing cycle. Principal component analysis produced a statistical model of the motion and deformation of the heart. We discuss how this model could be used to assist motion correction.
IEEE Transactions on Medical Imaging 10/2002; 21(9):1142-50. DOI:10.1109/TMI.2002.804427 · 3.39 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We describe a dynamic atlas that can be customized to an individual study subject in near-real-time. The atlas comprises 180
brain volumes each of which has been automatically segmented into grey matter, white matter and CSF, and also non-rigidly
registered to the Montreal BrainWeb reference dataset providing automatic delineation of brain structures of interest. To
create a dynamic atlas, the user loads a study dataset (eg: a patient) and queries the atlas database to identify similar
subjects. All selected database subjects are then aligned with the study subject using affine registration, and average tissue
probability maps and structure delineations produced. The system can run on distributed data and distributed CPUs illustrating
the potential of computational grids in medical image analysis.
[Show abstract][Hide abstract] ABSTRACT: A specially designed phantom consisting of a 3D array of 427 accurately manufactured spheres together with a point-based registration algorithm was used to detect distortion described by polynomial orders 1-4. More than thirty 3D gradient echo GRE and multi-slice spin echo SE phantom scans were acquired with a Philips 1.5T Gyroscan ACS2. Distortion was measured as a function of: readout gradient strength 0:72 = G r = 1 :7mTTm, TRRTEEEip angle, shim settings, and temporal distortion change for 11 weekly scans for the FFE sequence and TRRTEEslice gap for SE. Precision measurements for linear distortion were: scale = 0 :03, shear = 0 :04 degrees. Linear distortion in the readout dependent directions increased with decreased readout strength r r 0:93. There was a signiicantly higher p p 0:01 sagittal shear for 5 SE scans compared with 5 FFE ones with the same G r -possibly because of slice selection. Diierent shim settings produced only linear distortion change: up to 2 scale and 1 degree shear. There was negligible distortion change over time: scale 0:1, shear = 0 :05 degrees. There was a decrease in distortion as a function of polynomial order r r 0:9; n = 33, 75 of the distortion was either rst or second order.
Proceedings of SPIE - The International Society for Optical Engineering 07/2001; 4322(1). DOI:10.1117/12.431151 · 0.20 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Introduction In traditional MR acquisitions, k-space is sampled in a pre-determined fashion eg: raster, radial or spiral, with each measurement placed in a pre-allocated part of k-space. By analogy with the radar technique inverse synthetic aperture radar (ISAR), it is possible to reconstruct an image of an object moving in an unknown way, by estimating the position of the object for each measurement. Rotation of the object rotates the k-space position of the associated measurement, so by determining the motion of the object, each measurement can be correctly placed in k-space, and the resulting non-uniformly sampled k-space can be reconstructed. The novelty of this approach is that unknown object motion is seen as an aid rather than a hindrance to image acquisition. A potential advantage is that a small, cheap MR scanner could be built without phase-encoding gradients. By analogy with inverse SAR, we call this technique inverse MRI. In this proof-of-concept paper, we demonstrate that non-uniform rotational motion of an object can be used to reconstruct images, using conventional slice select and readout gradients, but no phase-encoding. Methods Images were acquired using a conventional 1.5T MR scanner (Philips Gyroscan ACS2), using a 256 line dual echo spin-echo acquisition TR=800ms,TE 1 =11ms, TE 2 =200ms) with the phase encode gradient turned off. The object was attached to a specially designed rig that fixed the center of rotation, and was rotated by a minimum of 180° free-hand (ie: with variable angular velocity). A 5mm diameter marker containing 0.5mM GdDTPA was attached to the rig. Due to choice of echo times, only the marker had high intensity in the second echo. Using software written in Matlab, each echo underwent a 1D Fourier transform to generate a space-time hybrid space. The marker was automatically tracked in the second echo hybrid space also known as a sinogram (Figure 1), as the path of the marker is an approximation to a sinusoid. To calculate the center of rotation of the object the half way point between the maximum and minimum displacement of the marker path is found. The angle of acquistion at each point is found from the inverse cosine of the ratio of the displacement at that point to the maximum displacement from the centre.