Kenneth R. Hoffmann

Generalizing the MTF and DQE to include x-ray scatter and focal spot unsharpness: application to a new microangiographic system.

Detector characterization with modulation transfer function (MTF) and detective quantum efficiency (DQE) inadequately predicts image quality when the imaging system includes focal spot unsharpness and patient scatter. The concepts of MTF, noise power spectrum, noise equivalent quanta and DQE were referenced to the object plane and generalized to include the effect of geometric unsharpness due to the finite size of the focal spot and the effect of the spatial distribution and magnitude of x-ray scatter due to the patient. The generalized quantities provide performance characteristics that consider the complete imaging system, but reduce to a description of the detector properties without magnification or scatter. We have evaluated a new neurovascular angiography imaging system based on a region of interest (ROI) microangiographic detector using these generalized quantities. A uniform head-equivalent phantom was used as a filter and x-ray scatter source. This allowed the study of all properties of the detector under clinically relevant x-ray spectra and x-ray scatter conditions. Realistic focal spots (0.8 mm nominal), beam energies (60-100 kVp), and detector exposures (0.8-2.3 mR) were used, and the effects of different scatter fractions (0-0.62) resulting from changing the beam size (0-100 cm2) were investigated. The generalized MTF and DQE were found to have very little dependence on the tube voltage and the detector entrance exposure. Magnification, with the focal spot used, results in a large decrease of the generalized DQE at higher frequencies (about 100-fold at 10 cycles/mm), but a significantly smaller decrease at lower frequencies. Scatter on the other hand, causes a constant drop in the generalized DQE (factor of 3 for scatter fraction 0.3) for all frequencies. Our results show that there are tradeoffs in the choice of the different system parameters; therefore this methodology of studying the imaging system as a whole could provide guidance in system design.

Determination of 3D imaging geometry and object configurations from two biplane views: An enhancement of the Metz–Fencil technique

We present a new technique based on the method developed by Metz and Fencil for estimation of the 3D imaging geometry and 3D object configurations from biplane angiographic acquisitions. The new method employs the 3D configuration of points calculated by the Metz-Fencil technique as an initial estimate. A 3D Procrustes algorithm is employed to translate, rotate, and scale the configuration until it aligns optimally with the set of lines that connects a focal spot with the corresponding set of image points. This alignment procedure is applied independently for each view. The rotation and translation that relate the two aligned data sets are then determined by an additional 3D Procrustes calculation. These steps are applied iteratively. Evaluations were based on Monte Carlo simulation and phantom studies. With this new technique, the mean absolute errors in magnification, in the relative position of the points, and in the angles defining the rotation and translation matrices were approximately 3.0%, 1.5 mm, and 5 degrees and 3 degrees, respectively, for rms input errors in the image data up to 2.0 pixels (0.7 mm). Errors in the results can be as small as 0.5%, 0.16 mm, 0.6 degrees, and 0.3 degrees, respectively, if input image-data error is 0.035 mm. The improvement of the Metz-Fencil technique described here may provide a basis for precise estimation of the biplane imaging geometry and the 3D positions of vessel bifurcation points.

Noninvasive identification of human central sulcus: a comparison of gyral morphology, functional MRI, dipole localization, and direct cortical mapping.

The locations of the human primary hand cortical somatosensory and motor areas were estimated using structural and functional MRI, scalp-recorded somatosensory-evoked potential dipole localization, expert judgments based on cortical anatomy, and direct cortical stimulation and recording studies. The within-subject reliability of localization (across 3 separate days) was studied for eight normal subjects. Intraoperative validation was obtained from five neurosurgical patients. The mean discrepancy between the different noninvasive functional imaging methods ranged from 6 to 26 mm. Quantitative comparison of the noninvasive methods with direct intraoperative stimulation and recording studies did not reveal a significant mean difference in accuracy. However, the expert judgments of the location of the sensory hand areas were significantly more variable (maximum error, 39 mm) than the dipole or functional MRI techniques. It is concluded that because expert judgments are less reliable for identifying the cortical hand area, consideration of the findings of noninvasive functional MRI and dipole localization studies is desirable for preoperative surgical planning.

GPU-based cone beam computed tomography

The use of cone beam computed tomography (CBCT) is growing in the clinical arena due to its ability to provide 3D information during interventions, its high diagnostic quality (sub-millimeter resolution), and its short scanning times (60 s). In many situations, the short scanning time of CBCT is followed by a time-consuming 3D reconstruction. The standard reconstruction algorithm for CBCT data is the filtered backprojection, which for a volume of size 256(3) takes up to 25 min on a standard system. Recent developments in the area of Graphic Processing Units (GPUs) make it possible to have access to high-performance computing solutions at a low cost, allowing their use in many scientific problems. We have implemented an algorithm for 3D reconstruction of CBCT data using the Compute Unified Device Architecture (CUDA) provided by NVIDIA (NVIDIA Corporation, Santa Clara, California), which was executed on a NVIDIA GeForce GTX 280. Our implementation results in improved reconstruction times from minutes, and perhaps hours, to a matter of seconds, while also giving the clinician the ability to view 3D volumetric data at higher resolutions. We evaluated our implementation on ten clinical data sets and one phantom data set to observe if differences occur between CPU and GPU-based reconstructions. By using our approach, the computation time for 256(3) is reduced from 25 min on the CPU to 3.2 s on the GPU. The GPU reconstruction time for 512(3) volumes is 8.5 s.

Automated calculation of the centerline of the human colon on CT images

RATIONALE AND OBJECTIVES This article presents an evaluation of an automated technique for determining the colon centerline with computed tomographic (CT) data sets. MATERIALS AND METHODS The technique proceeds as follows. After indication of a voxel in the rectum, voxels corresponding to air were segmented. Points along the colon centerline were estimated on the basis of centers of mass of grown voxels. A second segmentation and centerline calculation was initiated at the cecum. These two centerlines were then averaged. The resulting average was refined by using lumen data obtained perpendicular to the average centerline. The accuracy of the technique was investigated with simulation phantoms. The technique was also evaluated for 40 clinical colon cases. Calculated centerline points were compared with those indicated by radiologists for a randomly selected clinical case. RESULTS In the simulation studies, the calculated centerline points were, on average, within 2.5 mm of the true centerlines but differed by up to 4 mm in regions of deep folds or sharp turns. In the clinical colon study, 40% of the centerlines were computed with a single seed point and 25% with two seed points. Average centerlines were computed in 1 minute. The root mean square difference between the computed centerline points and those indicated by the radiologists was 4-5 mm (comparable to interobserver variations). CONCLUSION Accurate centerlines can be determined from colon CT data with this automated technique.

A system for determination of 3D vessel tree centerlines from biplane images

With the increasing number and complexity of therapeutic coronary interventions, there is an increasing need for accurate quantitative measurements. These interventions and measurements may be facilitated by accurate and reproducible magnifications and orientations of the vessel structures, specifically by accurate 3D vascular tree centerlines. A number of methods have been proposed to calculate 3D vascular tree centerlines from biplane images. In general, the calculated magnifications and orientations are accurate to within approximately 1–3% and 2–5°, respectively. Here, we present a complete system for determination of the 3D vessel centerlines from biplane angiograms without the use of a calibration object. Subsequent to indication of the vessel centerlines, the imaging geometry and 3D centerlines are calculated automatically and within approximately 2 min. The system was evaluated in terms of the intra- and inter-user variations of the various calculated quantities. The reproducibilities obtained with this system are comparable to or better than the accuracies and reproducibilities quoted for other proposed methods. Based on these results and those reported in earlier studies, we believe that this system will provide accurate and reproducible vascular tree centerlines from biplane images while the patient is still on the table, and thereby will facilitate interventions and associated quantitative analyses of the vasculature.

Automatic detection of pulmonary nodules in low-dose screening thoracic CT examinations

Computed tomography of the chest can be used as a screening method for lung cancer in a high-risk population. However, the detection of lung nodules is a difficult and time-consuming task for radiologists. The developed technique should improve the sensitivity of the detection of lung nodules without showing too many false positive nodules. In a study, which should evaluate the feasibility of screening lung cancer, about 1400 thoracic studies were acquired. Scanning parameters were 120 kVp, 5 mm collimation pitch of 2, and a reconstruction index of 5 mm. This results in a data set of about 60 to 70 images per exam. In the images the detection technique first eliminates all air outside the patient, then soft tissue and bony structures are removed. In the remaining lung fields a three-dimensional region detection is performed and rule-based analysis is used to detect possible lung nodules. This technique was applied to a small subset (n equals 17) of above studies. Computation time is about 5 min on an O2 workstation. The use of low-dose exams proved not be a hindrance in the detection of lung nodules. All of the nodules (n equals 23), except one with a size of 3 mm, were detected. The false positive rate was less than 0.3 per image. We have developed a technique, which might help the radiologist in the detection of pulmonary nodules in CT exams of the chest.

Micro‐angiography for neuro‐vascular imaging. II. Cascade model analysis

A micro-angiographic detector was designed and its performance was previously tested to evaluate its feasibility as an improvement over current x-ray detectors for neuro-interventional imaging. The detector was shown to have a modulation transfer function value of about 2% at the Nyquist frequency of 10 cycles/mm and a zero frequency detective quantum efficiency [DQE(0)] value of about 55%. An assessment of the system was required to evaluate whether the current system was performing at its full potential and to determine if any of its components could be optimized to further improve the output. For the purpose, in this study, the parallel cascade theory was used to analyze the performance of the detector under neuro-angiographic conditions by studying the output at the various stages in the imaging chain. A simple model for the spread of light in the CsI(Tl) entrance phosphor was developed and the resolution degradation due to K-fluorescence absorption was calculated. The total gain of the system was found to result in 21 e(-) (rms) detected at the charge coupled device per absorbed x-ray photon. The gain and the spread of quanta in the imaging chain were used to calculate theoretically the DQE using the parallel cascade model. The results of the model-based calculations matched fairly well with the experimental data previously obtained. This model was then used to optimize the phosphor thickness for the detector. The results showed that the area under the DQE curve had a maximum value at 150 microm of CsI(Tl), though when weighted by the squared signal in frequency space of a 100-microm-diam iodinated vessel, the integral DQE reached a maximum at 250 microm of CsI(Tl). Further, possible locations for gain increase in the imaging chain were determined, and the output of the improved system was simulated. Thus a theoretical analysis for the micro-angiographic detector was performed to better assess its potential.

Three-dimensional reconstruction of coronary arterial tree based on biplane angiograms

A method has been developed for in-room computer reconstruction of the three-dimensional (3-D) coronary arterial tree from routine biplane angiograms acquired at arbitrary angles and without using calibration objects. The proposed method consists of four major steps: (1) segmentation of vessel centerlines and bifurcation points and measurement of vessel diameters in coronary angiograms, (2) determination of biplane imaging parameters in terms of a rotation matrix R and a translation vector t based on the identified bifurcation points, (3) recovery of 3-D coronary arterial tree based on the calculated biplane imaging parameters, correspondences of vessel centerlines, and diameters, and (4) rendering of reconstructed 3-D coronary tree and estimation of optimal view of selected arterial segments. Angiograms from fifteen patients were utilized for 3-D reconstruction for each patients coronary arterial tree. The biplane imaging geometry was first determined without a calibration object, and the 3-D coronary arterial trees were reconstructed including both left and right coronary artery systems. Various 2-D projection images of the reconstructed 3-D coronary arterial tree were generated and compared to other viewing angles obtained in the actual patient study. Similarity between the real and reconstructed arterial structures was excellent. The accuracy of this method was evaluated by using a computer-simulated coronary arterial tree. Root-mean-square (RMS) errors in the 3-D position and 3-D configuration of vessel centerlines and in the angles defining the R matrix and

Automated Tracking Of The Vascular Tree In Dsa Images Using A Double-Square-Box Region-Of-Search Algorithm

t vector were 1.2 - 5.5 mm, 0.3 - 2.0 mm, and less than 1.5 and 2.0 degrees, respectively, when using 2-D vessel centerlines with RMS normally distributed errors varying from 0.7 - 4.2 pixels (0.25 - 1.26 mm).