Application of a long short-term memory neural network: a burgeoning method of deep learning in forecasting HIV incidence in Guangxi, China.

Guangxi, a province in southwestern China, has the second highest reported number of HIV/AIDS cases in China. This study aimed to develop an accurate and effective model to describe the tendency of HIV and to predict its incidence in Guangxi. HIV incidence data of Guangxi from 2005 to 2016 were obtained from the database of the Chinese Center for Disease Control and Prevention. Long short-term memory (LSTM) neural network models, autoregressive integrated moving average (ARIMA) models, generalised regression neural network (GRNN) models and exponential smoothing (ES) were used to fit the incidence data. Data from 2015 and 2016 were used to validate the most suitable models. The model performances were evaluated by evaluating metrics, including mean square error (MSE), root mean square error, mean absolute error and mean absolute percentage error. The LSTM model had the lowest MSE when the N value (time step) was 12. The most appropriate ARIMA models for incidence in 2015 and 2016 were ARIMA (1, 1, 2) (0, 1, 2)12 and ARIMA (2, 1, 0) (1, 1, 2)12, respectively. The accuracy of GRNN and ES models in forecasting HIV incidence in Guangxi was relatively poor. Four performance metrics of the LSTM model were all lower than the ARIMA, GRNN and ES models. The LSTM model was more effective than other time-series models and is important for the monitoring and control of local HIV epidemics.

Stationary intraoral digital tomosynthesis using a carbon nanotube X-ray source array.

OBJECTIVES: Intraoral dental tomosynthesis and closely related tuned-aperture CT (TACT) are low-dose three-dimensional (3D) imaging modalities that have shown improved detection of multiple dental diseases. Clinical interest in implementing these technologies waned owing to their time-consuming nature. Recently developed carbon nanotube (CNT) X-ray sources allow rapid multi-image acquisition without mechanical motion, making tomosynthesis a clinically viable technique. The objective of this investigation was to evaluate the feasibility of and produce high-quality images from a digital tomosynthesis system employing CNT X-ray technology. METHODS: A test-bed stationary intraoral tomosynthesis unit was constructed using a CNT X-ray source array and a digital intraoral sensor. The source-to-image distance was modified to make the system comparable in image resolution to current two-dimensional intraoral radiography imaging systems. Anthropomorphic phantoms containing teeth with simulated and real caries lesions were imaged using a dose comparable to D-speed film dose with a rectangular collimation. Images were reconstructed and analysed. RESULTS: Tomosynthesis images of the phantom and teeth specimen demonstrated perceived image quality equivalent or superior to standard digital images with the added benefit of 3D information. The ability to "scroll" through slices in a buccal-lingual direction significantly improved visualization of anatomical details. In addition, the subjective visibility of dental caries was increased. CONCLUSIONS: Feasibility of the stationary intraoral tomosynthesis is demonstrated. The results show clinical promise and suitability for more robust observer and clinical studies.

A first generation compact microbeam radiation therapy system based on carbon nanotube X-ray technology.

We have developed a compact microbeam radiation therapy device using carbon nanotube cathodes to create a linear array of narrow focal line segments on a tungsten anode and a custom collimator assembly to select a slice of the resulting wedge-shaped radiation pattern. Effective focal line width was measured to be 131 µm, resulting in a microbeam width of ∼300 µm. The instantaneous dose rate was projected to be 2 Gy/s at full-power. Peak to valley dose ratio was measured to be >17 when a 1.4 mm microbeam separation was employed. Finally, multiple microbeams were delivered to a mouse with beam paths verified through histology.

TU-E-217BCD-11: Evaluating the Performance of a Stationary Digital Breast Tomosynthesis System.

PURPOSE: In conventional Digital Breast Tomosynthesis (DBT) systems a single x-ray source moves over a limited angle arc. This leads to motion blurring in the projection images associated with x-ray source motion and total scan times. We have developed a stationary DBT (s-DBT) system which forgoes a rotating source for an array of carbon nanotube (CNT) based x-ray sources. Here we report the results of evaluating the performance and the optimization of image acquisition parameters of the s-DBT system. METHODS: The s-DBT system consists of a linear source array with 31 x-ray generating focal spots distributed over a 30 degree angular span. The source array has been retrofitted onto a Hologic Selenia Dimensions DBT system. An American College of Radiology accreditation phantom was imaged to assess the quality of the reconstruction images in different configurations. A line wire phantom is used to measure the modulation transfer function (MTF). RESULTS: For the standard imaging protocol, the system resolution along the scanning direction is increased from 3.0 cycles/mm in DBT to 4.2 cycles/mm in s-DBT at a magnification factor of 1.08. The MTF did not have a noticeable change between different configurations, whereas in DBT the MTF can be degraded for larger angular spans due to faster x-ray source motion. The overall image quality factor is found to be best for the configuration with a large angular span and intermittent number of projection views. CONCLUSIONS: We demonstrated successful construction and operation of the s-DBT system integrating a CNT x-ray source array with a Hologic DBT system. The spatial resolution of the s-DBT system is demonstrated to be substantially increased over the corresponding DBT system. It was found that a configuration with a large angular span, an intermittent number of projection views, and an even dose distribution resulted in the best overall image quality. Hologic INC has provided the Hologic Selenia Dimensions used in the research. The project is supported by the National Cancer Institute under grant number U54CA119343 and R01CA134598 and the UNC University Cancer Research Fund. Dr. Xin Qian is supported by a fellowship from the Department of Defense under grant number BC087505.

WE-C-217BCD-03: Restricted Data Set Reconstruction Based on Respiration Quality to Improve Prospectively Gated in Vivo Micro-CT of Mice.

PURPOSE: Micro-CT is commonly employed for lung imaging of mice; prospective gating allows for in-vivo imaging of free-breathing subjects. While this technique is successfully executed for healthy animals, results are less consistent for some disease models whose symptoms include irregular or unstable respiration. The purpose of this work is to repair the quality of high-blur images that arise from respiration instability using a retrospective method of motion reduction which identifies the individual x-ray projection images contributing most to the motion blur. Reconstructions were performed after the exclusion of these projections (the so-called restricted set). METHODS: Sixteen mice were imaged using field emission cone beam micro-CT and prospective gating with a bellows-type respiration sensor. The scanner was operated in step-and-shoot mode; 400 projection images were acquired per scan. An algorithm was developed to analyze the respiration trace file and segment the individual breath corresponding to each projection image. We tested three different criteria to define a bad breath shape (correlation, mean breath height, or mode breath height), and restricted data set reconstructions were performed using each of these criteria to exclude projections corresponding to bad breaths. Each restricted set was compared against the full unrestricted data set image; the slope perpendicular to the diaphragm was used as a quantitative assessment of motion blur. RESULTS: All image sets saw a reduction in motion blur with at least one restriction technique. In 22 of 27 images, improvement was measured regardless of the removal criterion. Five percent total projection removal is optimal; a more aggressive correction increases the likelihood of under-sampling artifacts. CONCLUSIONS: Removing a subset of bad projections from otherwise complete image sets measurably decreases motion blur in respiratory-gated imaging. An approach based on breath height generally provides the best results. The technique is applicable to a variety of imaging modalities.

SU-D-217A-01: A High-Resolution in Vivo Molecular Imaging Technique Based on X- Ray Fluorescence.

PURPOSE: Traditional molecular imaging techniques such as PET/SPECT have many limitations, including relatively low spatial resolution, short lifetime of radioisotope probes, limited availability due to reliance on cyclotron, relatively high dose, and lack of effective molecular probes for certain tumor cells. In this work we demonstrate the feasibility of a novel x- ray fluorescence molecular imaging (XFMI) technique using high-power carbon nanotube (CNT) x-ray array technology. The XFMI overcomes some limitations and will be a significant advance in molecular imaging technology for cancer drug development and cancer biology research. METHODS: A testing chamber was constructed containing two Amptek energy-resolving detectors analyzing a copper collimated fluorescence beam placed opposite each other symmetric about the sample and perpendicular to the primary incoming x-ray beam, also using the copper collimator. Different concentrations of indium trichloride and iodine were tested to determine the minimum detectable concentration (MDC) for each. RESULTS: The MDC for indium was found to be 80 ug/mL and 100 ug/mL for iodine at 50 kVp, 30mAs (5 minutes imaging time). This is on the order of magnitude of the MDCs determined at large synchrotron facilities using XFI. Higher concentrations above 1mg/mL of both elements were detectable at 1.5 mAs (15 seconds). CONCLUSIONS: These results show that it is possible to not only measure low concentrations of the above elements, but also distinguish between similar L-alpha peaks. This allows for future work of obtaining 2D and 3D imaging to determine element types and concentrations diffused in different parts of the body. Carolina Center of Cancer Nanotechnology, University Cancer Research Fund.

Design and feasibility studies of a stationary digital breast tomosynthesis system.

Studies have shown that digital breast tomosynthesis (DBT) can improve breast cancer diagnosis by reconstructing 3D images. However, DBT scanners based on rotation gantry prolong the imaging time and reduce spatial resolution due to motion comparing with the regular two-view mammography. To obtain three dimension reconstruction images and maintain the high image quality of conventional mammography, we proposed a prototype stationary digital breast tomosynthesis system (s-DBT). The proposed s-DBT system acquires projection images without mechanical movement. The core component of the s-DBT system is a specially designed spatially distributed multi-beam x-ray tube based on the carbon nanotube field emission x-ray technology. The multi-beam x-ray source array enables collection of all projection images from different viewing angles without mechanical motion. Preliminary results show the s-DBT system can achieve a scan time comparable to the regular two-view mammography, and improve the spatial resolution comparing with rotating gantry DBT.

Stationary digital breast tomosynthesis with distributed field emission X-ray tube.

Tomosynthesis requires projection images from different viewing angles. Using a distributed x-ray source this can be achieved without mechanical motion of the source with the potential for faster image acquisition speed. A distributed x-ray tube has been designed and manufactured specifically for breast tomosynthesis. The x-ray tube consists of 31 field emission x-ray sources with an angular range of 30°. The total dose is up to 100mAs with an energy range between 27 and 45 kVp. We discuss the source geometry and results from the characterization of the first prototype. The x-ray tube uses field emission cathodes based on carbon nanotubes (CNT) as electron source. Prior to the manufacturing of the sealed x-ray tube extensive testing on the field emission cathodes has been performed to verify the requirements for commercial tomosynthesis systems in terms of emission current, focal spot size and tube lifetime.

Distributed source x-ray tube technology for tomosynthesis imaging.

Tomosynthesis imaging requires projection images from different viewing angles. Conventional systems use a moving xray source to acquire the individual projections. Using a stationary distributed x-ray source with a number of sources that equals the number of required projections, this can be achieved without any mechanical motion. Advantages are a potentially faster image acquisition speed, higher spatial and temporal resolution and simple system design. We present distributed x-ray sources based on carbon nanotube (CNT) field emission cathodes. The field emission cathodes deliver the electrons required for x-ray production. CNT emitters feature a stable emission at high current density, a cold emission, excellent temporal control of the emitted electrons and good configurability. We discuss the use of stationary sources for two applications: (i) a linear tube for stationary digital breast tomosynthesis (sDBT), and (ii) a square tube for on-board tomosynthesis image-guided radiation therapy (IGRT). Results from high energy distributed sources up to 160kVp are also presented.

A dynamic micro-CT scanner based on a carbon nanotube field emission x-ray source.

Current commercial micro-CT scanners have the capability of imaging objects ex vivo with high spatial resolution, but performing in vivo micro-CT on free-breathing small animals is still challenging because their physiological motions are non-periodic and much faster than those of humans. In this paper, we present a prototype physiologically gated micro-computed tomography (micro-CT) scanner based on a carbon nanotube field emission micro-focus x-ray source. The novel x-ray source allows x-ray pulses and imaging sequences to be readily synchronized and gated to non-periodic physiological signals from small animals. The system performance is evaluated using phantoms and sacrificed and anesthetized mice. Prospective respiratory-gated micro-CT images of anesthetized free-breathing mice were collected using this scanner at 50 ms temporal resolution and 6.2 lp mm(-1) at 10% system MTF. The high spatial and temporal resolutions of the micro-CT scanner make it well suited for high-resolution imaging of free-breathing small animals.

Development of a nanotechnology based low-LET multi-microbeam array single cell irradiation system.

A novel single cell irradiation system using carbon nanotube (CNT) based field emission technology is proposed. The system can produce electron microbeam at a large range of pulsation frequencies and dose rates with energy between 20 and 60 keV. Different from any existing single beam microbeam device, the CNT-based system can have 10,000 microbeam pixels, each is approximately 10 microm in size and individually controlled. Microscope imaging will be used for targeting cell(s) and the coordinate(s) identification. A single cell or large number of individually selected cells can be simultaneously irradiated under real time microscope observation. This poster reports our preliminary results in the initial stage of the CNT multipixel microbeam array development-prototype single pixel CNT microbeam device development.

Lithium intercalation into opened single-wall carbon nanotubes: storage capacity and electronic properties.

The effects of structure and morphology on lithium storage in single-wall carbon nanotube (SWNT) bundles were studied by electrochemistry and nuclear magnetic resonance techniques. SWNTs were chemically etched to variable lengths and were intercalated with Li. The reversible Li storage capacity increased from LiC(6) in close-end SWNTs to LiC(3) after etching, which is twice the value observed in intercalated graphite. All the nanotubes became metallic upon intercalation of Li, with the density of states at the Fermi level increasing with increasing Li concentration. The enhanced capacity is attributed to Li diffusion into the interior of the SWNTs through the opened ends and sidewall defects.

Electronic structures of single-walled carbon nanotubes determined by NMR

Single-walled carbon nanotubes were studied by (13)C nuclear magnetic resonance (NMR). Two types of (13)C nuclear spins were identified with different spin-lattice relaxation rates. The fast-relaxing component, assigned to metallic tubes, followed the relaxation behavior expected in metals, and the density-of-states at the Fermi level increased with decreasing tube diameter. The slow-relaxing component has a significantly lower density-of-states at the Fermi level. Exposure to oxygen has a substantial effect on relaxation rates of both components.

Magnetic susceptibility of molecular carbon: nanotubes and fullerite.

Elemental carbon can be synthesized in a variety of geometrical forms, from three-dimensional extended structures (diamond) to finite molecules (C(60) fullerite). Results are presented here on the magnetic susceptibility of the least well-understood members of this family, nanotubes and C(60) fullerite. (i) Nanotubes represent the cylindrical form of carbon, intermediate between graphite and fullerite. They are found to have significantly larger orientation-averaged susceptibility, on a per carbon basis, than any other form of elemental carbon. This susceptibility implies an average band structure among nanotubes similar to that of graphite. (ii) High-resolution magnetic susceptibility data on C(60) fullerite near the molecular orientational-ordering transition at 259 K show a sharp jump corresponding to 2.5 centimeter-gram-second parts per million per mole of C(60). This jump directly demonstrates the effect of an intermolecular cooperative transition on an intramolecular electronic property, where the susceptibility jump may be ascribed to a change in the shape of the molecule due to lattice forces.

New phases of c60 synthesized at high pressure.

The fullerene C(60) can be converted into two different structures by high pressure and temperature. They are metastable and revert to pristine C(60) on reheating to 300 degrees C at ambient pressure. For synthesis temperatures between 300 degrees and 400 degrees C and pressures of 5 gigapascals, a nominal face-centered-cubic structure is produced with a lattice parameter a(o) = 13.6 angstroms. When treated at 500 degrees to 800 degrees C at the same pressure, C(60) transforms into a rhombohedral structure with hexagonal lattice parameters of a(o) = 9.22 angstroms and c(o) = 24.6 angstroms. The intermolecular distance is small enough that a chemical bond can form, in accord with the reduced solubility of the pressure-induced phases. Infrared, Raman, and nuclear magnetic resonance studies show a drastic reduction of icosahedral symmetry, as might occur if the C(60) molecules are linked.

Partial characterization of the proteolytic secretions of Acanthamoeba polyphaga.

The ocular pathology associated with acanthamoebiasis may result, at least in part, from the excretory and secretory (E&S) products of the active Acanthamoeba trophozoites. To test this hypothesis, the ability of A. polyphaga (ATCC Strain 30461) trophozoite E&S products to digest collagen, the major constituent of the corneal stroma, was evaluated. The secreted proteinases of A. polyphaga were identified using in vitro azocoll degradation, activity-PAGE, radiolabeled extracellular matrix (ECM) degradation, and collagen degradation assays. Inhibitors of serine (phenylmethylsulfonyl fluoride, diisopropyl fluorophosphate), cysteine [benzyloxyphenylalanyl-analyl fluoromethyl ketone, N-ethylmaleamide, ethylenediamine tetraacetic acid (EDTA), L-trans-3-carboxyiran-2-carbonyl-L-leucylagmatine], metallo- (1,10-phenanthroline, EDTA, phosphoramidon), and aspartyl (pepstatin A) proteinases were incorporated into the assays. Proteinase activity was detected in trophozoites and the E&S products of trophozoites but not in cysts. The azocoll and activity-PAGE assays indicated the presence of serine and cysteine proteinases, while degradation of the radiolabeled ECM by live trophozoites confirmed not only the presence of serine and cysteine proteinases but also metalloproteinase(s). Most proteinase activity occurred at neutral pH. Incubation of E&S with type I collagen did not yield the typical 3/4:1/4 products characteristic of vertebrate collagenases. These data suggest that E&S products of A. polyphaga contain multiple serine and cysteine proteinases with nonspecific collagenolytic activity and that metallproteinases form an additional minor constituent.

Defects in carbon nanostructures.

Previous high-resolution electron microscopy (HREM) observations of the carbon nanotubes have led to a "Russian doll" structural model that is based on hollow concentric cylinders capped at both ends. The structures of the carbon nanotubes and particles were characterized here by bulk physical and chemical property measurements. The individual nanostructure is as compressible as graphite in the c axis, and such nanostructures can be intercalated with potassium and rubidium, leading to a saturation composition of "MC(8)." These results are counter to expectations that are based on a Russian doll structure. HREM after intercalation with potassium and deintercalation indicates that individual nanoparticles are a "paper-mache" of smaller graphite layers. Direct current magnetization and electron spin resonance measurements indicate that the electronic properties of the nanostructures are distinctly different from those of graphite. Although the nanostructures have distinct morphologies and electronic properties, they are highly defective and have a local structure similar to turbostratic graphite.