Biomedical subjects

C E Floyd

Publications and source records attributed to C E Floyd.

80 records · Page 5Linked to original sources

PubMedDate not supplied

Tc-99m attenuation coefficients in water-filled phantoms determined with gamma cameras.

Quantitative imaging with gamma cameras requires compensation for attenuation of source photons. Some methods of compensation make use of a constant or average estimated attenuation coefficient mu. A value for mu of 0.15 cm-1 for 140.5-keV photons in water or tissue is commonly used. This value, however, neglects scattered photons which are detected within the energy window in gamma camera imaging. Values for mu of 0.12 cm-1 used in attenuation compensation of Tc-99m single-photon emission computed tomography scans of uniform cylindrical sources have been shown to give improved results compared with use of mu = 0.15 cm-1. In this study, gamma cameras and a multichannel pulse-height analyzer were used to determine effective values of mu for photons in water as a function of energy window. Two cylindrical water-filled phantoms, circular and elliptical, were used with a point source of Tc-99m at depths up to 18 cm. Energy data were integrated over the top half of the photopeak, and over 10%, 20%, and 30% windows centered on the photopeak. Attenuation curves were exponential for all photopeak windows with values of mu of 0.12 +/- 0.014 cm-1 for all windows up to 20% and 0.1 cm-1 for a 30% window. This study suggests that a value of mu of 0.11-0.12 cm-1 is, in fact, appropriate for use in attenuation compensations where an average is required.

Models, Structural↗

PubMedDate not supplied

Monte Carlo evaluation of Compton scatter subtraction in single photon emission computed tomography.

In single photon emission computed tomography (SPECT), Compton scattering produces a background that degrades the image quality and contributes erroneously to quantitative measurements. A clinically implementable compensation algorithm has previously been reported that subtracts a Compton scatter image, acquired in an energy window set below the energy of the photopeak, from the primary image acquired in the photopeak window. We present an evaluation and justification of the assumptions made in the previous empirical development of the subtraction algorithm. A Monte Carlo model of the SPECT system in which the Compton scattered vents may be followed independently of the nonscattered events was used to evaluate this subtraction technique. Simulation shows that the assumptions made in the experimental application of this algorithm were valid. Specifically (1) the "scatter" energy window used in the experiment (91-125 keV for imaging Tc-99m) contains only scattered events, (2) the shape of the line spread function (LSF) for the events in the scatter window is a reasonable approximation to the shape of the scatter in the photopeak window, and (3) the ratio of the number of scattered events in the photopeak window to the number of events in the scatter window is 0.57, close to the value of 0.5 derived heuristically. Thus, Monte Carlo simulation validates the basic assumptions underlying the empirical implementation of the scatter subtraction algorithm.

Humans↗

PubMedDate not supplied

Cone beam collimation for single photon emission computed tomography: analysis, simulation, and image reconstruction using filtered backprojection.

This paper presents an analysis of two cone beam configurations (having focal lengths of 40 and 60 cm) for the acquisition of single photon emission computed tomography (SPECT) projection data. A three-dimensional filtered backprojection algorithm is used to reconstruct SPECT images of cone beam projection data obtained using Monte Carlo simulations. The mathematical analysis resulted in on-axis point source sensitivities (calculated for a distance of 15 cm from the collimator surface) for cone beam configurations that were 1.4-3 times the sensitivities of parallel-hole and fan beam geometries having similar geometric resolutions. Cone beam collimation offers the potential for improved sensitivity for SPECT devices using large-field-of-view scintillation cameras.

Humans↗

PubMedDate not supplied

Quantitative radiographic imaging using a photostimulable phosphor system.

We have evaluated a photostimulable phosphor x-ray imaging system [Philips Computed Radiography (PCR) system] for use in quantification of x-ray exposure in diagnostic radiography. An exponential function was fitted to data yielding quantitative x-ray exposure values as a function of digital pixel values. We investigated several factors that affect the accuracy of exposure measurement using the PCR including repeatability, background noise as a function of time delay between plate erasure and use, sensitivity variation between different plates, nonuniformity of sensitivity within a plate, decay of the latent image between time of exposure and readout (observed as a change in sensitivity), and the accuracy with which the (exponential) calibration function yields exposure values as a function of digital pixel values. The calibration was performed over the exposure range from 5.1 X 10(-9) to 2.0 X 10(-5) C/kg (0.02-75 mR). The accuracy of exposure measurements made with a single imaging plate is between 1.6% and 4.2%. If measurements from several plates are involved, the uncertainty in the final measurement will be between 5% and 5.9%.

Evaluation Studies as Topic↗

PubMedDate not supplied

Quantitative scatter measurement in digital radiography using a photostimulable phosphor imaging system.

X-ray scatter fractions measured with two detectors are compared: a photostimulable phosphor system (PSP) and a conventional film-screen technique. For both detection methods, a beam-stop technique was used to estimate the scatter fraction in polystyrene phantoms. These scatter fraction measurements are compared to previously reported film-based measurements. Scatter fractions obtained with the PSP were in good agreement both with measurements using film as well as with most previously reported measurements. For the PSP measurements, repeatability was better than 1%. It was found that the PSP provides a precise x-ray detector for quantitative scatter measurement in digital radiography.

Luminescent Measurements↗

PubMedDate not supplied

An artificial neural network for estimating scatter exposures in portable chest radiography.

An adaptive linear element (Adaline) was developed to estimate the two-dimensional scatter exposure distribution in digital portable chest radiographs (DPCXR). DPCXRs and quantitative scatter exposure measurements at 64 locations throughout the chest were acquired for ten radiographically normal patients. The Adaline is an artificial neural network which has only a single node and linear thresholding. The Adaline was trained using DPCXR-scatter measurement pairs from five patients. The spatially invariant network would take a portion of the image as its input and estimate the scatter content as output. The trained network was applied to the other five images, and errors were evaluated between estimated and measured scatter values. Performance was compared against a convolution scatter estimation algorithm. The network was evaluated as a function of network size, initial values, and duration of training. Network performance was evaluated qualitatively by the correlation of network weights to physical models, and quantitatively by training and evaluation errors. Using DPCXRs as input, the network learned to describe known scatter exposures accurately (7% error) and estimate scatter in new images (< 8% error) slightly better than convolution methods. Regardless of size and initial shape, all networks adapted into radial exponentials with magnitude of 0.75, perhaps implying an ideal point spread function and average scatter fraction, respectively. To implement scatter compensation, the two-dimensional scatter distribution estimated by the neural network is subtracted from the original DPCXR.

Algorithms↗

PubMedDate not supplied

A vectorized Monte Carlo code for modeling photon transport in SPECT.

A vectorized Monte Carlo computer code has been developed for modeling photon transport in single photon emission computed tomography (SPECT). The code models photon transport in a uniform attenuating region and photon detection by a gamma camera. It is adapted from a history-based Monte Carlo code in which photon history data are stored in scalar variables and photon histories are computed sequentially. The vectorized code is written in FORTRAN77 and uses an event-based algorithm in which photon history data are stored in arrays and photon history computations are performed within DO loops. The indices of the DO loops range over the number of photon histories, and these loops may take advantage of the vector processing unit of our Stellar GS1000 computer for pipelined computations. Without the use of the vector processor the event-based code is faster than the history-based code because of numerical optimization performed during conversion to the event-based algorithm. When only the detection of unscattered photons is modeled, the event-based code executes 5.1 times faster with the use of the vector processor than without; when the detection of scattered and unscattered photons is modeled the speed increase is a factor of 2.9. Vectorization is a valuable way to increase the performance of Monte Carlo code for modeling photon transport in SPECT.

Algorithms↗

PubMedDate not supplied

The effect of data sampling on the performance evaluation of artificial neural networks in medical diagnosis.

PURPOSE: To study the effect of data sampling on the predictive assessment of artificial neural networks (ANNs) for medical diagnostic tasks. METHODS: Three statistical techniques were used to evaluate the diagnostic performances of ANNs: 1) cross validation, 2) round robin, and 3) bootstrap. These techniques are different sampling plans designed to reduce the small-sample estimation bias and variance contributions. The study was based on two networks, one developed for the diagnosis of pulmonary embolism (1,064 cases) and the other developed for the diagnosis of breast cancer (206 cases). RESULTS: The three sampling techniques produced different performance estimates for both networks. The estimates varied substantially depending on the training sample size and the training-stopping criterion. CONCLUSION: The predictive assessment of ANNs in medical diagnosis can vary substantially based on the complexity of the problem, the data sampling technique, and the number of cases available.

Breast Neoplasms↗