Neutron stimulated emission computed tomography: a Monte Carlo simulation approach.

A Monte Carlo simulation has been developed for neutron stimulated emission computed tomography (NSECT) using the GEANT4 toolkit. NSECT is a new approach to biomedical imaging that allows spectral analysis of the elements present within the sample. In NSECT, a beam of high-energy neutrons interrogates a sample and the nuclei in the sample are stimulated to an excited state by inelastic scattering of the neutrons. The characteristic gammas emitted by the excited nuclei are captured in a spectrometer to form multi-energy spectra. Currently, a tomographic image is formed using a collimated neutron beam to define the line integral paths for the tomographic projections. These projection data are reconstructed to form a representation of the distribution of individual elements in the sample. To facilitate the development of this technique, a Monte Carlo simulation model has been constructed from the GEANT4 toolkit. This simulation includes modeling of the neutron beam source and collimation, the samples, the neutron interactions within the samples, the emission of characteristic gammas, and the detection of these gammas in a Germanium crystal. In addition, the model allows the absorbed radiation dose to be calculated for internal components of the sample. NSECT presents challenges not typically addressed in Monte Carlo modeling of high-energy physics applications. In order to address issues critical to the clinical development of NSECT, this paper will describe the GEANT4 simulation environment and three separate simulations performed to accomplish three specific aims. First, comparison of a simulation to a tomographic experiment will verify the accuracy of both the gamma energy spectra produced and the positioning of the beam relative to the sample. Second, parametric analysis of simulations performed with different user-defined variables will determine the best way to effectively model low energy neutrons in tissue, which is a concern with the high hydrogen content in biological tissue. Third, determination of the energy absorbed in tissue during neutron interrogation in order to estimate the dose. Results from these three simulation experiments demonstrate that GEANT4 is an effective simulation platform that can be used to facilitate the future development and optimization of NSECT.

Compression of digital coronary angiograms does not affect visual or quantitative assessment of coronary artery stenosis severity.

Digital coronary angiographic techniques are now widely used in many cardiac catheterization laboratories. However, the full potential of digital imaging technology remains to be achieved because of its enormous storage and exchange requirements. Compression of digital imaging data allows a reduction in the volume of data so that storage and transmission are more efficient and cost-effective. Three angiographers reviewed the original and compressed formats of 96 coronary angiographic sequences in a blinded fashion to assess coronary lesion severity. Compression was achieved using the Joint Photographic Experts Group (JPEG) standard, which resulted in a compression ratio of approximately 15:1. The original format was reviewed in a blinded fashion a second time to assess for intraobserver variability of similar formats. Lesion severity was graded in quartiles. Coronary stenosis >50% was considered "significant." In parallel, the reproducibility of quantitative coronary angiographic (QCA) measurements of coronary artery dimensions was also evaluated. For the visual assessment of lesion severity in the compressed versus original formats, kappa=0.52, suggesting moderate agreement. When lesions were assessed as significant versus "insignificant," however, kappa=0.88, suggesting excellent agreement. In the 2 separate readings of the original data formats, kappa=0.44 for assessment of lesion severity by quartiles and kappa=0.72 for lesions assessed as significant versus insignificant. Analysis of the compressed versus original data sets using QCA resulted in an excellent correlation for the measurement of lesion severity (r=0.99). The correlation was equally strong when the original format was analyzed sequentially (r=0.98). Lossy JPEG (15:1) compression is a valid means for reducing storage and exchange requirements of coronary angiographic data. The variability in assessing lesion severity between the original and compressed formats is comparable to the reported variability in visual assessment of lesion severity in sequential analysis of cine film.