Search NASA⌕ Search

SEARCH · Search NASA

Results for “gradient methods”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 703 records · Page 39

An extension of the localized artificial diffusivity method for immiscible and high density ratio flows

The localized artificial diffusivity (LAD) method is widely regarded as the preferred multi-material regularization scheme for the compact finite difference method, because it is conservative, easy to implement, and generally robust for a wide range of multi-material problems. However, traditional LAD methods face significant challenges when applied to flows with large density ratios and when maintaining thermodynamic equilibrium across material interfaces. These limitations arise from the formulation of the artificial diffusivity flux and the reliance on enthalpy diffusion for interface regularization. Additionally, traditional LAD methods struggle to ensure stability under large density ratio conditions, fail to maintain a finite interface thickness, and are therefore unsuitable for modeling immiscible interfaces. Here, in this work, we discuss the origins of these issues in traditional LAD methods and propose modifications which enable the simulation of large density ratio and immiscible flows. The proposed method targets the artificial diffusion fluxes at gradients and ringing in the volume fraction, rather than the mass fraction in traditional methods, to consistently regularize large density ratio interfaces. Furthermore, the proposed method introduces an artificial bulk density diffusion term to enforce equilibrium conditions across interfaces. To address the challenge of modeling immiscible flows, a conservative diffuse interface term is incorporated into the formulation to ensure a finite interface thickness. Specific consideration is taken in the design of the method to ensure that these crucial properties are maintained for N -material flows. The effectiveness of the proposed method is demonstrated through a series of canonical test cases, and its accuracy is validated by comparison with experimental data on micro-bubble collapse in water. These results highlight the method’s robustness and its ability to overcome the limitations of traditional LAD approaches.

Artificial diffusivity↗

General-Purpose Bayesian Tensor Learning With Automatic Rank Determination and Uncertainty Quantification

A major challenge in many machine learning tasks is that the model expressive power depends on model size. Low-rank tensor methods are an efficient tool for handling the curse of dimensionality in many large-scale machine learning models. The major challenges in training a tensor learning model include how to process the high-volume data, how to determine the tensor rank automatically, and how to estimate the uncertainty of the results. While existing tensor learning focuses on a specific task, this paper proposes a generic Bayesian framework that can be employed to solve a broad class of tensor learning problems such as tensor completion, tensor regression, and tensorized neural networks. We develop a low-rank tensor prior for automatic rank determination in nonlinear problems. Our method is implemented with both stochastic gradient Hamiltonian Monte Carlo (SGHMC) and Stein Variational Gradient Descent (SVGD). We compare the automatic rank determination and uncertainty quantification of these two solvers. We demonstrate that our proposed method can determine the tensor rank automatically and can quantify the uncertainty of the obtained results. We validate our framework on tensor completion tasks and tensorized neural network training tasks.

Bayesian inference↗

Renormalized quark masses using gradient flow

We propose a new and simple method for determining the renormalized quark masses from lattice simulations. Renormalized quark masses are an important input to many phenomenological applications, including searching and modeling physics beyond the Standard Model. The nonperturbative renormalization is performed using gradient flow combined with the short-flow-time expansion that is improved by renormalization group (RG) running to match to the $\overline{MS}$ scheme. Implementing the RG running perturbatively, we demonstrate this method works reliably at least up to the charm-quark mass and exhibits an easily attainable “windowing condition.” Using RBC/UKQCD’s (2+1)-flavor Shamir domain-wall fermion ensembles with Iwasaki gauge action, we find 𝑚 $\overline{MS}$ 𝑠⁡ (𝜇 = 2 GeV) = 89⁢(3) MeV and 𝑚 $\overline{MS}$ 𝑐 ⁡(𝜇 = 3 GeV) = 972⁢(16) MeV. These results predict the scale-independent ratio 𝑚 𝑐 /𝑚 𝑠 = 12.1⁢(4). Generalization to other observables is possible, providing an efficient approach to determine nonperturbatively renormalized fermionic observables like form factors or bag parameters from lattice simulations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Stability of compressible reduced magnetohydrodynamic equilibria-Analogy with magnetorotational instability

Stability analyses for equilibria of the compressible reduced magnetohydrodynamics (CRMHD) model are carried out by means of the Energy-Casimir (EC) method. Stability results are compared with those obtained for ideal magnetohydrodynamics (MHD) from the classical {delta}W criterion. An identification of the terms in the second variation of the free energy functional for CRMHD with those of {delta}W is made: two destabilizing effects present for CRMHD turn out to correspond to the kink and interchange instabilities in usual MHD, while the stabilizing roles of field line bending and compressibility are also identified in the reduced model. Also, using the EC method, stability conditions in the presence of toroidal flow are obtained. A formal analogy between CRMHD and a reduced incompressible model for magnetized rotating disks, due to Julien and Knobloch [EAS Pub. Series, 21, 81 (2006)], is discovered. In light of this analogy, energy stability analysis shows that the condition for magnetorotational instability (MRI) for the latter model corresponds to the condition for interchange instability in CRMHD, with the Coriolis term and shear velocity playing the roles of the curvature term and pressure gradient, respectively. Using the EC method, stability conditions for the rotating disk model, for a large class of equilibria with possible non-uniform magnetic fields, are obtained. In particular, this shows it is possible for the MRI system to undergo, in addition to the MRI, another instability that is analogous to the kink instability. For vanishing magnetic field, the Rayleigh hydrodynamical stability condition is recovered.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Efficient Generalizable Deep Learning

Stochastic optimization is a fundamental field of research for machine learning. Stochastic gradient descent (SGD) and related methods provide a feasible means to train complicated prediction models over large datasets. SGD, however, does not explicitly address the problem of overfitting, which can lead to predictions that perform poorly on new data. This difference between loss performance on unseen testing data verses that of training data defines the generalization gap of a model. We introduce a new computational kernel called Stochastic Hessian Projection (SHP) that uses a maximum likelihood framework to simultaneously estimate gradient noise covariance and local curvature of the loss function. Our analysis illustrates that these quantities affect the evolution of parameter uncertainty and therefore generalizability. We show how these computations allow us to predict the generalization gap without requiring holdout data. Explicitly assessing this metric for generalizability during training may improve machine learning predictions when data is scarce and understanding prediction variability is critical.

42 ENGINEERING↗

DAKOTA Design Analysis Kit for Optimization and Terascale

The DAKOTA (Design Analysis Kit for Optimization and Terascale Applications) toolkit provides a flexible and extensible interface between simulation codes (computational models) and iterative analysis methods. By employing object-oriented design to implement abstractions of the key components required for iterative systems analyses, the DAKOTA toolkit provides a flexible and extensible problem-solving environment for design and analysis of computational models on high performance computers.A user provides a set of DAKOTA commands in an input file and launches DAKOTA. DAKOTA invokes instances of the computational models, collects their results, and performs systems analyses. DAKOTA contains algorithms for optimization with gradient and nongradient-based methods; uncertainty quantification with sampling, reliability, polynomial chaos, stochastic collocation, and epistemic methods; parameter estimation with nonlinear least squares methods; and sensitivity/variance analysis with design of experiments and parameter study methods. These capabilities may be used on their own or as components within advanced strategies such as hybrid optimization, surrogate-based optimization, mixed integer nonlinear programming, or optimization under uncertainty. Services for parallel computing, simulation interfacing, approximation modeling, fault tolerance, restart, and graphics are also included.

SciDAC↗

Method for computationally efficient design of dielectric laser accelerator structures

Here, dielectric microstructures have generated much interest in recent years as a means of accelerating charged particles when powered by solid state lasers. The acceleration gradient (or particle energy gain per unit length) is an important figure of merit. To design structures with high acceleration gradients, we explore the adjoint variable method, a highly efficient technique used to compute the sensitivity of an objective with respect to a large number of parameters. With this formalism, the sensitivity of the acceleration gradient of a dielectric structure with respect to its entire spatial permittivity distribution is calculated by the use of only two full-field electromagnetic simulations, the original and ‘adjoint’. The adjoint simulation corresponds physically to the reciprocal situation of a point charge moving through the accelerator gap and radiating. Using this formalism, we perform numerical optimizations aimed at maximizing acceleration gradients, which generate fabricable structures of greatly improved performance in comparison to previously examined geometries.

43 PARTICLE ACCELERATORS↗

Volumetric Additive Manufacturing of Gradient Composition Glass

Advanced optics, such as lasers or cameras, are currently limited by geometric or compositional restrictions on the silica glass components that act as focusing or refracting lenses. Currently, these glass components must be made of one homogenous material in a range of shapes and sizes limited by a few specialized vendors. However, additional flexibility and customization of these optical components could significantly improve the performance of these devices and expand the design space available to optics engineers, as would the capability to locally change the optical properties of these glass components by changing the composition of the glass as a function of location. To this end, we have developed a method to print and process gradient composition glass, using volumetric additive manufacturing (VAM) and direct ink writing (DIW), on-site at Lawrence Livermore National Lab.

36 MATERIALS SCIENCE↗

Harnessing Machine Learning to Predict MoS 2 Solid Lubricant Performance

Physical vapor deposited (PVD) molybdenum disulfide (MoS 2 ) solid lubricant coatings are an exemplar material system for machine learning methods due to small changes in process variables often causing large variations in microstructure and mechanical/tribological properties. Here, in this work, a gradient boosted regression tree machine learning method is applied to an existing experimental data set containing process, microstructure, and property information to create deeper insights into the process-structure–property relationships for molybdenum disulfide (MoS 2 ) solid lubricant coatings. The optimized and cross-validated models show good predictive capabilities for density, reduced modulus, hardness, wear rate, and initial coefficients of friction. The contribution of individual deposition variables (i.e., argon pressure, deposition power, target conditioning) on coating properties is highlighted through feature importance. The process-property relationships established herein show linear and non-linear relationships and highlight the influence of uncontrolled deposition variables (i.e., target conditioning) on the tribological performance.

MoS2↗

Composition gradient effects on strain relaxation in Sr-doped LaMnO{sub 3} epitaxial thin films

The authors report on a novel method to fabricate Sr-doped composition gradient epitaxial La{sub 1−x}Sr{sub x}MnO{sub 3} thin films by radio frequency magnetron sputtering. Biaxially strained epitaxial La{sub 1−x}Sr{sub x}MnO{sub 3} thin films were grown on (001) LaAlO{sub 3} substrates by following a cosputtering procedure from LaMnO{sub 3} and La{sub 0.67}Sr{sub 0.33}MnO{sub 3} targets. Three depositions were conducted by varying the substrate temperature (750 and 850 °C) and controlling the relative deposition rate from the two targets by varying their power rate during sputtering. The thickness of the thin films was about 20 and 30 nm for the short and long duration deposition, respectively. The films were studied by symmetric θ–2θ x-ray diffraction, pole figure analysis, atomic force microscopy, and x-ray photoelectron spectroscopy. Fabrication of smooth, composition gradient films of high epitaxial quality was achieved at a substrate temperature of 850 °C and low sputtering rate. A novel strain relaxation mechanism is also found that decreases significantly the mismatch between the film and substrate as the Sr doping level increases. The reported deposition procedure can produce new possibilities of designing nanoscale structures with cross coupled properties that may result in new materials.

42 ENGINEERING↗

Out-of-time-order correlators bridge classical transport and quantum dynamics

The out-of-time-order correlator (OTOC) has emerged as a central tool for quantifying decoherence across wide-ranging physical platforms. Here, we demonstrate its direct measurement in a classical ensemble using nuclear magnetic resonance with a modulated gradient spin echo sequence and extend the method into a multidimensional correlation to track exchange phenomena. Position is encoded through magnetic field gradients and momentum through the velocity autocorrelation function, enabling experimental access to OTOCs for proton motion confined within the self-similar lattice of the metal–organic framework MOF-808. Here, water confined to specified geometries within the MOF pores gives rise to spatially distinct diffusive eigenmodes with characteristic relative entropies. We demonstrate that periodic radio frequency driving combined with gradient modulation yields entropy evolution through the selection of distinct diffusion modes. Frequency-resolved diffusion spectra connect these entropy dynamics to classical heat exchange laws, revealing how operational features of quantum systems are mirrored in confined, macroscopic spin ensembles.

Fricke, Sophia N. [University of California, Berke↗

The use of seismic spatial gradients in a single layer neural network for seismic source discrimination: proof of concept

This report describes a proof-of-concept method of seismic source discrimination using seismic gradiometry and a common machine learning technique. The tests described here are purely numerical, using synthetic seismic data and well understood mathematical techniques. The primary innovation described here is the application of a richer seismic data set derived from seismic gradiometry. Seismic gradiometry is a method to estimate the time variable spatial gradient of the wavefield to compute various wavefield attributes such as slowness, dynamic strain, and rotational motions. With the addition of these wavefield attributes, we are afforded up to twenty "compo- nents" of time series data measured at a single point on, or in, the Earth. This is in direct contrast to conventional three-component seismic data collected at several locations using a seismic network. Using the gradiometrically-derived wavefield components directly in a single-layer neural network, I show that it is possible to discriminate between three common seismic source types (earthquakes, explosions, and opening fractures) for various noise conditions and gradiometry configurations.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Generalized Procedure for Improved Accuracy of Thermal Contact Resistance Measurements for Materials With Arbitrary Temperature-Dependent Thermal Conductivity

Thermal contact resistance (TCR) is most commonly measured using one-dimensional steady-state calorimetric techniques. In the experimental methods we utilized, a temperature gradient is applied across two contacting beams and the temperature drop at the interface is inferred from the temperature profiles of the rods that are measured at discrete points. During data analysis, thermal conductivity of the beams is typically taken to be an average value over the temperature range imposed during the experiment. Our generalized theory is presented and accounts for temperature-dependent changes in thermal conductivity. The procedure presented enables accurate measurement of TCR for contacting materials whose thermal conductivity is any arbitrary function of temperature. For example, it is shown that the standard technique yields TCR values that are about 15% below the actual value for two specific examples of copper and silicon contacts. Conversely, the generalized technique predicts TCR values that are within 1% of the actual value. The method is exact when thermal conductivity is known exactly and no other errors are introduced to the system.

42 ENGINEERING↗

Towards optimized viral metagenomes for double-stranded and single-stranded DNA viruses from challenging soils

Soils impact global carbon cycling and their resident microbes are critical to their biogeochemical processing and ecosystem outputs. Based on studies in marine systems, viruses infecting soil microbes likely modulate host activities via mortality, horizontal gene transfer, and metabolic control. However, their roles remain largely unexplored due to technical challenges with separating, isolating, and extracting DNA from viruses in soils. Some of these challenges have been overcome by using whole genome amplification methods and while these have allowed insights into the identities of soil viruses and their genomes, their inherit biases have prevented meaningful ecological interpretations. Here we experimentally optimized steps for generating quantitatively-amplified viral metagenomes to better capture both ssDNA and dsDNA viruses across three distinct soil habitats along a permafrost thaw gradient. First, we assessed differing DNA extraction methods (PowerSoil, Wizard mini columns, and cetyl trimethylammonium bromide) for quantity and quality of viral DNA. This established PowerSoil as best for yield and quality of DNA from our samples, though ~1/3 of the viral populations captured by each extraction kit were unique, suggesting appreciable differential biases among DNA extraction kits. Second, we evaluated the impact of purifying viral particles after resuspension (by cesium chloride gradients; CsCl) and of viral lysis method (heat vs bead-beating) on the resultant viromes. DNA yields after CsCl particle-purification were largely non-detectable, while unpurified samples yielded 1–2-fold more DNA after lysis by heat than by bead-beating. Virome quality was assessed by the number and size of metagenome-assembled viral contigs, which showed no increase after CsCl-purification, but did from heat lysis relative to bead-beating. We also evaluated sample preparation protocols for ssDNA virus recovery. In both CsCl-purified and non-purified samples, ssDNA viruses were successfully recovered by using the Accel-NGS 1S Plus Library Kit. While ssDNA viruses were identified in all three soil types, none were identified in the samples that used bead-beating, suggesting this lysis method may impact recovery. Further, 13 ssDNA vOTUs were identified compared to 582 dsDNA vOTUs, and the ssDNA vOTUs only accounted for ~4% of the assembled reads, implying dsDNA viruses were dominant in these samples. This optimized approach was combined with the previously published viral resuspension protocol into a sample-to-virome protocol for soils now available at protocols.io, where community feedback creates ‘living’ protocols. This collective approach will be particularly valuable given the high physicochemical variability of soils, which will may require considerable soil type-specific optimization. This optimized protocol provides a starting place for developing quantitatively-amplified viromic datasets and will help enable viral ecogenomic studies on organic-rich soils.

Trubl, Gareth↗

Towards optimized viral metagenomes for double-stranded and single-stranded DNA viruses from challenging soils

Soils impact global carbon cycling and their resident microbes are critical to their biogeochemical processing and ecosystem outputs. Based on studies in marine systems, viruses infecting soil microbes likely modulate host activities via mortality, horizontal gene transfer, and metabolic control. However, their roles remain largely unexplored due to technical challenges with separating, isolating, and extracting DNA from viruses in soils. Some of these challenges have been overcome by using whole genome amplification methods and while these have allowed insights into the identities of soil viruses and their genomes, their inherit biases have prevented meaningful ecological interpretations. Here we experimentally optimized steps for generating quantitatively-amplified viral metagenomes to better capture both ssDNA and dsDNA viruses across three distinct soil habitats along a permafrost thaw gradient. First, we assessed differing DNA extraction methods (PowerSoil, Wizard mini columns, and cetyl trimethylammonium bromide) for quantity and quality of viral DNA. This established PowerSoil as best for yield and quality of DNA from our samples, though ~1/3 of the viral populations captured by each extraction kit were unique, suggesting appreciable differential biases among DNA extraction kits. Second, we evaluated the impact of purifying viral particles after resuspension (by cesium chloride gradients; CsCl) and of viral lysis method (heat vs bead-beating) on the resultant viromes. DNA yields after CsCl particle-purification were largely non-detectable, while unpurified samples yielded 1–2-fold more DNA after lysis by heat than by bead-beating. Virome quality was assessed by the number and size of metagenome-assembled viral contigs, which showed no increase after CsCl-purification, but did from heat lysis relative to bead-beating. We also evaluated sample preparation protocols for ssDNA virus recovery. In both CsCl-purified and non-purified samples, ssDNA viruses were successfully recovered by using the Accel-NGS 1S Plus Library Kit. While ssDNA viruses were identified in all three soil types, none were identified in the samples that used bead-beating, suggesting this lysis method may impact recovery. Further, 13 ssDNA vOTUs were identified compared to 582 dsDNA vOTUs, and the ssDNA vOTUs only accounted for ~4% of the assembled reads, implying dsDNA viruses were dominant in these samples. This optimized approach was combined with the previously published viral resuspension protocol into a sample-to-virome protocol for soils now available at protocols.io, where community feedback creates ‘living’ protocols. This collective approach will be particularly valuable given the high physicochemical variability of soils, which will may require considerable soil type-specific optimization. This optimized protocol provides a starting place for developing quantitatively-amplified viromic datasets and will help enable viral ecogenomic studies on organic-rich soils.

Trubl, Gareth↗

Towards optimized viral metagenomes for double-stranded and single-stranded DNA viruses from challenging soils

Soils impact global carbon cycling and their resident microbes are critical to their biogeochemical processing and ecosystem outputs. Based on studies in marine systems, viruses infecting soil microbes likely modulate host activities via mortality, horizontal gene transfer, and metabolic control. However, their roles remain largely unexplored due to technical challenges with separating, isolating, and extracting DNA from viruses in soils. Some of these challenges have been overcome by using whole genome amplification methods and while these have allowed insights into the identities of soil viruses and their genomes, their inherit biases have prevented meaningful ecological interpretations. Here we experimentally optimized steps for generating quantitatively-amplified viral metagenomes to better capture both ssDNA and dsDNA viruses across three distinct soil habitats along a permafrost thaw gradient. First, we assessed differing DNA extraction methods (PowerSoil, Wizard mini columns, and cetyl trimethylammonium bromide) for quantity and quality of viral DNA. This established PowerSoil as best for yield and quality of DNA from our samples, though ∼1/3 of the viral populations captured by each extraction kit were unique, suggesting appreciable differential biases among DNA extraction kits. Second, we evaluated the impact of purifying viral particles after resuspension (by cesium chloride gradients; CsCl) and of viral lysis method (heat vs bead-beating) on the resultant viromes. DNA yields after CsCl particle-purification were largely non-detectable, while unpurified samples yielded 1–2-fold more DNA after lysis by heat than by bead-beating. Virome quality was assessed by the number and size of metagenome-assembled viral contigs, which showed no increase after CsCl-purification, but did from heat lysis relative to bead-beating. We also evaluated sample preparation protocols for ssDNA virus recovery. In both CsCl-purified and non-purified samples, ssDNA viruses were successfully recovered by using the Accel-NGS 1S Plus Library Kit. While ssDNA viruses were identified in all three soil types, none were identified in the samples that used bead-beating, suggesting this lysis method may impact recovery. Further, 13 ssDNA vOTUs were identified compared to 582 dsDNA vOTUs, and the ssDNA vOTUs only accounted for ∼4% of the assembled reads, implying dsDNA viruses were dominant in these samples. This optimized approach was combined with the previously published viral resuspension protocol into a sample-to-virome protocol for soils now available at protocols.io , where community feedback creates ‘living’ protocols. This collective approach will be particularly valuable given the high physicochemical variability of soils, which will may require considerable soil type-specific optimization. This optimized protocol provides a starting place for developing quantitatively-amplified viromic datasets and will help enable viral ecogenomic studies on organic-rich soils.

dsDNA viruses↗

Depth‐dependent concentrations of hematopoietic stem cells in the adult skeleton: Implications for active marrow dosimetry

Purpose The hematopoietically active (or red) bone marrow is the target tissue assigned in skeletal dosimetry models for assessment of stochastic effects (leukemia induction) as well as tissue reactions (marrow toxicity). Active marrow, however, is in reality a surrogate tissue region for specific cell populations, namely the hematopoietic stem and progenitor cells. Present models of active marrow dosimetry implicitly assume that these cells are uniformly localized throughout the marrow spaces of trabecular spongiosa. Data from Watchman et al. and Bourke et al., however, clearly indicate that there is a substantial spatial concentration gradient of these cells with the highest concentrations localized near the bone trabeculae surfaces. The purpose of the present study was thus to explore the dosimetric implications of these spatial gradients on active marrow dosimetry. Methods Images of several bone sites from a 45‐yr female were retagged to group active marrow voxels into 50 μm increments of marrow depth, after which electron and alpha‐particle depth‐dependent specific absorbed fractions were computed for four source tissues — active marrow, inactive marrow, bone trabeculae volumes, and bone trabeculae surfaces. Corresponding depth‐dependent S values (dose to a target tissue per decay in a source tissue) were computed and further weighted by the relative target cell concentration. These depth‐weighted radionuclide S values were systematically compared to the more traditional volume‐averaged radionuclide S values of the MIRD schema for both individual bones of the skeleton and their skeletal‐averaged quantities. Results For both beta‐emitters and alpha‐emitters localized in the active and inactive marrow, depth‐weighted S values were shown to differ from volume‐averaged S values by only a few percent, as dose gradients across the marrow tissues are nonexistent. For bone volume and bone surface sources of alpha‐emitters and lower energy beta‐emitters, when marrow dose gradients are expected, explicit consideration of target cell spatial concentration gradients are shown to significantly impact marrow dosimetry. Conclusions For medical isotopes currently utilized for treatment of skeletal metastases, namely 153 Sm and 223 Ra, accounting for hematopoietic stem and progenitor cell concentration gradients resulted in maximum percent differences to reference skeletal‐averaged S values of ~21% and 55%, respectively.

Geyer, Amy M.↗

SU-E-J-275: Review - Computerized PET/CT Image Analysis in the Evaluation of Tumor Response to Therapy

Purpose: To review the literature in using computerized PET/CT image analysis for the evaluation of tumor response to therapy. Methods: We reviewed and summarized more than 100 papers that used computerized image analysis techniques for the evaluation of tumor response with PET/CT. This review mainly covered four aspects: image registration, tumor segmentation, image feature extraction, and response evaluation. Results: Although rigid image registration is straightforward, it has been shown to achieve good alignment between baseline and evaluation scans. Deformable image registration has been shown to improve the alignment when complex deformable distortions occur due to tumor shrinkage, weight loss or gain, and motion. Many semi-automatic tumor segmentation methods have been developed on PET. A comparative study revealed benefits of high levels of user interaction with simultaneous visualization of CT images and PET gradients. On CT, semi-automatic methods have been developed for only tumors that show marked difference in CT attenuation between the tumor and the surrounding normal tissues. Quite a few multi-modality segmentation methods have been shown to improve accuracy compared to single-modality algorithms. Advanced PET image features considering spatial information, such as tumor volume, tumor shape, total glycolytic volume, histogram distance, and texture features have been found more informative than the traditional SUVmax for the prediction of tumor response. Advanced CT features, including volumetric, attenuation, morphologic, structure, and texture descriptors, have also been found advantage over the traditional RECIST and WHO criteria in certain tumor types. Predictive models based on machine learning technique have been constructed for correlating selected image features to response. These models showed improved performance compared to current methods using cutoff value of a single measurement for tumor response. Conclusion: This review showed that computerized PET/CT image analysis holds great potential to improve the accuracy in evaluation of tumor response. This work was supported in part by the National Cancer Institute Grant R01CA172638.

60 APPLIED LIFE SCIENCES↗