SEARCH · Engineering Papers

Results for “NORMAL”

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.

SourceResearch connectionFrom yearThrough year

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

At least 91 records · Page 5

All sources NASA NTRS DOE OSTI

NASA NTRS1998

Decorin and biglycan of normal and pathologic human corneas

PURPOSE: Corneas with scars and certain chronic pathologic conditions contain highly sulfated dermatan sulfate, but little is known of the core proteins that carry these atypical glycosaminoglycans. In this study the proteoglycan proteins attached to dermatan sulfate in normal and pathologic human corneas were examined to identify primary genes involved in the pathobiology of corneal scarring. METHODS: Proteoglycans from human corneas with chronic edema, bullous keratopathy, and keratoconus and from normal corneas were analyzed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), quantitative immunoblotting, and immunohistology with peptide antibodies to decorin and biglycan. RESULTS: Proteoglycans from pathologic corneas exhibit increased size heterogeneity and binding of the cationic dye alcian blue compared with those in normal corneas. Decorin and biglycan extracted from normal and diseased corneas exhibited similar molecular size distribution patterns. In approximately half of the pathologic corneas, the level of biglycan was elevated an average of seven times above normal, and decorin was elevated approximately three times above normal. The increases were associated with highly charged molecular forms of decorin and biglycan, indicating modification of the proteins with dermatan sulfate chains of increased sulfation. Immunostaining of corneal sections showed an abnormal stromal localization of biglycan in pathologic corneas. CONCLUSIONS: The increased dermatan sulfate associated with chronic corneal pathologic conditions results from stromal accumulation of decorin and particularly of biglycan in the affected corneas. These proteins bear dermatan sulfate chains with increased sulfation compared with normal stromal proteoglycans.

Non-NASA Center↗

NASA NTRS2016

Normalized Temperature Contrast Processing in Flash Infrared Thermography

The paper presents further development in normalized contrast processing of flash infrared thermography method by the author given in US 8,577,120 B1. The method of computing normalized image or pixel intensity contrast, and normalized temperature contrast are provided, including converting one from the other. Methods of assessing emissivity of the object, afterglow heat flux, reflection temperature change and temperature video imaging during flash thermography are provided. Temperature imaging and normalized temperature contrast imaging provide certain advantages over pixel intensity normalized contrast processing by reducing effect of reflected energy in images and measurements, providing better quantitative data. The subject matter for this paper mostly comes from US 9,066,028 B1 by the author. Examples of normalized image processing video images and normalized temperature processing video images are provided. Examples of surface temperature video images, surface temperature rise video images and simple contrast video images area also provided. Temperature video imaging in flash infrared thermography allows better comparison with flash thermography simulation using commercial software which provides temperature video as the output. Temperature imaging also allows easy comparison of surface temperature change to camera temperature sensitivity or noise equivalent temperature difference (NETD) to assess probability of detecting (POD) anomalies.

Koshti, Ajay M.↗

NASA NTRS2016

Normalized Temperature Contrast Processing in Infrared Flash Thermography

The paper presents further development in normalized contrast processing used in flash infrared thermography method. Method of computing normalized image or pixel intensity contrast, and normalized temperature contrast are provided. Methods of converting image contrast to temperature contrast and vice versa are provided. Normalized contrast processing in flash thermography is useful in quantitative analysis of flash thermography data including flaw characterization and comparison of experimental results with simulation. Computation of normalized temperature contrast involves use of flash thermography data acquisition set-up with high reflectivity foil and high emissivity tape such that the foil, tape and test object are imaged simultaneously. Methods of assessing other quantitative parameters such as emissivity of object, afterglow heat flux, reflection temperature change and surface temperature during flash thermography are also provided. Temperature imaging and normalized temperature contrast processing provide certain advantages over normalized image contrast processing by reducing effect of reflected energy in images and measurements, therefore providing better quantitative data. Examples of incorporating afterglow heat-flux and reflection temperature evolution in flash thermography simulation are also discussed.

Koshti, Ajay M.↗

NASA NTRS2016

A Method to Measure and Estimate Normalized Contrast in Infrared Flash Thermography

Koshti, Ajay M.↗

DOE OSTI2019

In Vitro Bovine Liver Experiment of Cisplatin-Infused and Normal Saline-Infused Radiofrequency Ablation with an Internally Cooled Perfusion Electrode

Purpose: To evaluate the efficacy of cisplatin-infused and normal saline-infused radiofrequency ablation (RFA) with internally cooled perfusion (ICP) electrode. Materials and Methods: Using a 200 W generator, thirty ablation zones were created and divided into three groups of 10 each as follows: group A, RFA alone with 16 gauge monopolar internally cooled (IC) electrode; group B, cisplatin-infused RFA with 16 gauge ICP electrode; and group C, normal saline-infused RFA with 16 gauge ICP electrode. Radiofrequency was applied to the explanted bovine liver for 12 min. During RFA, cisplatin and normal saline were injected into tissue at a rate of 0.5 mL/min through the ICP electrode by injection pump. Dimensions of the ablation zone and technical parameters were compared between the three groups. Result: In the cisplatin-infused RFA group, the ablation zone size was significantly larger than that of the RFA-alone group but significantly smaller than normal saline-infused RFA group. The width of longitudinal section and volume were 3.39 ± 0.22 cm{sup 2} and 26.55 ± 4.62 cm{sup 3} in RFA-alone group, 3.88 ± 0.32 cm{sup 2} and 36.45 ± 5.46 cm{sup 3} in cisplatin-infused RFA group, and 4.52 ± 0.50 cm{sup 2} and 49.44 ± 7.55 cm{sup 3} in normal saline-infused RFA group, respectively (p < 0.05 between any two groups). The mean impedance in group A, B, and C were 60.0 ± 7.2, 50.3 ± 2.5, and 40.3 ± 4.0 Ω, respectively (p < 0.05 between any two groups). Conclusion: Cisplatin-infused RFA with ICP electrode created the larger size of ablation zone than that of monopolar RFA with an IC electrode, but created the smaller size of ablation zone than that of normal saline-infused RFA.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2017

Effect of Normalizing Temperature on Fracture Characteristic of Tensile and Impact Tested Creep Strength-Enhanced Ferritic P92 Steel

The high-temperature Cr-Mo creep strength-enhanced ferritic (CSEF) steels are mainly used in nuclear and thermal power plants. In the present investigation, a systematic study on fracture surface morphologies of tensile and impact tested specimens and mechanical properties of cast and forged (C&F) P92 steel was performed for various heat treatment conditions. The heat treatment was carried out in normalizing temperature range of 950-1150 °C and then tempered to a fixed tempering temperature of 760 °C. The effect of varying normalizing temperatures before and after tempering on microstructure evolution, tensile properties, Vicker’s hardness and Charpy toughness was studied. The normalizing temperature before and after tempering was having a noticeable effect on mechanical properties of as-received P92 steel. The fracture surface of impact and tensile tested samples was also studied for various normalizing temperatures with or without tempering. Fracture surface morphology was affected by the presence of secondary phase carbide particles. The fraction area of cleavage facets on the tensile fracture surface was found to be increased with an increase in the normalizing temperature. The fractured tensile specimens were characterized by transgranular ductile dimples, tear ridges and transgranular cleavage facets for various heat treatments. The fracture mode of impact tested samples was more complex. It showed both quasi-cleavage facets and ductile dimple tearing for various normalizing temperatures.

36 MATERIALS SCIENCE↗

DOE OSTI2015

An analytical elastic plastic contact model with strain hardening and frictional effects for normal and oblique impacts

Impact between metallic surfaces is a phenomenon that is ubiquitous in the design and analysis of mechanical systems. We found that to model this phenomenon, a new formulation for frictional elastic–plastic contact between two surfaces is developed. The formulation is developed to consider both frictional, oblique contact (of which normal, frictionless contact is a limiting case) and strain hardening effects. The constitutive model for normal contact is developed as two contiguous loading domains: the elastic regime and a transitionary region in which the plastic response of the materials develops and the elastic response abates. For unloading, the constitutive model is based on an elastic process. Moreover, the normal contact model is assumed to only couple one-way with the frictional/tangential contact model, which results in the normal contact model being independent of the frictional effects. Frictional, tangential contact is modeled using a microslip model that is developed to consider the pressure distribution that develops from the elastic–plastic normal contact. This model is validated through comparisons with experimental results reported in the literature, and is demonstrated to be significantly more accurate than 10 other normal contact models and three other tangential contact models found in the literature.

36 MATERIALS SCIENCE↗

DOE OSTI2013

On the continuity of mean total normal stress in geometrical multiscale cardiovascular problems

In this work an iterative strategy to implicitly couple dimensionally-heterogeneous blood flow models accounting for the continuity of mean total normal stress at interface boundaries is developed. Conservation of mean total normal stress in the coupling of heterogeneous models is mandatory to satisfy energetic consistency between them. Nevertheless, existing methodologies are based on modifications of the Navier–Stokes variational formulation, which are undesired when dealing with fluid–structure interaction or black box codes. The proposed methodology makes possible to couple one-dimensional and three-dimensional fluid–structure interaction models, enforcing the continuity of mean total normal stress while just imposing flow rate data or even the classical Neumann boundary data to the models. This is accomplished by modifying an existing iterative algorithm, which is also able to account for the continuity of the vessel area, when required. Comparisons are performed to assess differences in the convergence properties of the algorithms when considering the continuity of mean normal stress and the continuity of mean total normal stress for a wide range of flow regimes. Finally, examples in the physiological regime are shown to evaluate the importance, or not, of considering the continuity of mean total normal stress in hemodynamics simulations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2024

A reduced-order model for nonlinear radiative transfer problems based on moment equations and POD-Petrov-Galerkin projection of the normalized Boltzmann transport equation

A data-driven projection-based reduced-order model (ROM) for nonlinear thermal radiative transfer (TRT) problems is presented. The TRT ROM is formulated by (i) a hierarchy of low-order quasidiffusion (aka variable Eddington factor) equations for moments of the radiation intensity and (ii) the normalized Boltzmann transport equation (BTE). The multilevel system of moment equations is derived by projection of the BTE onto a sequence of subspaces which represent elements of the phase space of the problem. Exact closure for the moment equations is provided by the Eddington tensor. A Petrov-Galerkin (PG) projection of the normalized BTE is formulated using a proper orthogonal decomposition (POD) basis representing the normalized radiation intensity over the whole phase space and time. The Eddington tensor linearly depends on the solution of the normalized BTE. By linear superposition of the POD basis functions, a low-rank expansion of the Eddington tensor is constructed with coefficients defined by the PG projected normalized BTE. The material energy balance (MEB) equation is coupled with the effective gray low-order equations which exist on the same dimensional scale as the MEB equation. The resulting TRT ROM is structure and asymptotic preserving. A detailed analysis of the ROM is performed on the classical Fleck-Cummings (F-C) TRT multigroup test problem in 2D geometry. Numerical results are presented to demonstrate the ROM's effectiveness in the simulation of radiation wave phenomena. Importantly, the ROM is shown to produce solutions with sufficiently high accuracy while using low-rank approximation of the normalized BTE solution. Essential physical characteristics of supersonic radiation wave are preserved in the ROM solutions.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2017

Nickel-induced HIF-1α promotes growth arrest and senescence in normal human cells but lacks toxic effects in transformed cells

Nickel is a human carcinogen that acts as a hypoxia mimic by activating the transcription factor HIF-1α and hypoxia-like transcriptomic responses. Hypoxia and elevated HIF-1α are typically associated with drug resistance in cancer cells, which is caused by increased drug efflux and other mechanisms. Here we examined the role of HIF-1α in uptake of soluble Ni(II) and Ni(II)-induced cell fate outcomes using si/shRNA knockdowns and gene deletion models. We found that HIF-1α had no effect on accumulation of Ni(II) in two transformed (H460, A549) and two normal human cell lines (IMR90, WI38). The loss of HIF-1α also produced no significant impact on p53-dependent and p53-independent apoptotic responses or clonogenic survival of Ni(II)-treated transformed cells. In normal human cells, HIF-1α enhanced the ability of Ni(II) to inhibit cell proliferation and cause a permanent growth arrest (senescence). Consistent with its growth-suppressive effects, HIF-1α was important for upregulation of the cell cycle inhibitors p21 (CDKN1A) and p27 (CDKN1B). Irrespective of HIF-1α status, Ni(II) strongly increased levels of MYC protein but did not change protein expression of the cell cycle-promoting phosphatase CDC25A or the CDK inhibitor p16. Our findings indicate that HIF-1α limits propagation of Ni(II)-damaged normal cells, suggesting that it may act in a tumor suppressor-like manner during early stages of Ni(II) carcinogenesis. - Highlights: • HIF-1α enhances growth arrest and senescence by Ni(II) in normal human cells. • HIF-1α is important for upregulation of the cell cycle inhibitors p21 and p27 in Ni(II)-treated normal cells. • Apoptotic responses and clonogenic toxicity by Ni(II) in transformed human cells are HIF-1α independent. • HIF-1α does not affect Ni(II) accumulation in normal or transformed human cells.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2019

Deducing Phonon Scattering from Normal Mode Excitations

While the quantum scattering theory has provided the theoretical underpinning for phonon interactions, the correspondence between the phonon modes and normal modes of vibrations has never been fully established; for example, the nature of energy exchange during elementary normal mode interactions remains largely unknown. In this work, by adopting a set of real asymmetric normal mode amplitudes, we first discriminate the normal and Umklapp processes directly from atomistic dynamics. We then demonstrate that the undulating harmonic and anharmonic potentials, which allow a number of interaction pathways, generate several total-energy-conserving forward and backward scattering events including those which are traditionally considered as quantum-forbidden. Although the normal mode energy is proportional to the square of the eigen-frequency, we deduce that the energy exchanged from one mode to another in each elementary interaction is proportional to the frequency – a quantum-like restriction. We anticipate that the current approach can be utilized profitably to discover unbiased scattering channels, many traditionally quantum forbidden, with complex anharmonicities. Our discovery will aid in the development of next-generation Peierls-Boltzmann transport simulations that access normal mode scattering pathways from finite temperature ab initio simulations.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2014

Controlling the number of graphene sheets exfoliated from graphite by designed normal loading and frictional motion

We use molecular dynamics to study the exfoliation of patterned nanometer-sized graphite under various normal loading conditions for friction-induced exfoliation. Using highly ordered pyrolytic graphite (HOPG) as well as both amorphous and crystalline SiO{sub 2} substrate as example systems, we show that the exfoliation process is attributed to the corrugation of the HOPG surface and the atomistic roughness of the substrate when they contact under normal loading. The critical normal strain, at which the exfoliation occurs, is higher on a crystalline substrate than on an amorphous substrate. This effect is related to the atomistic flatness and stiffness of the crystalline surface. We observe that an increase of the van der Waals interaction between the graphite and the substrate results in a decrease of the critical normal strain for exfoliation. We find that the magnitude of the normal strain can effectively control the number of exfoliated graphene layers. This mechanism suggests a promising approach of applying designed normal loading while sliding to pattern controlled number of graphene layers or other two-dimensional materials on a substrate surface.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2022

Frictional properties of Opalinus Clay: influence of humidity, normal stress and grain size on frictional stability

SUMMARY The Opalinus Clay (OPA) is a clay-rich formation considered as a potential host rock for radioactive waste repositories and as a caprock for carbon storage in Switzerland. Its very low permeability (10−19 to 10−21 m2) makes it a potential sealing horizon, however the presence of faults that may be activated during the lifetime of a repository project can compromise the long-term hydrological confinement, and lead to mechanical instability. Here, we have performed laboratory experiments to test the effect of relative humidity (RH), grain size (g.s.) and normal stress on rate-and-state frictional properties and stability of fault laboratory analogues corresponding to powders of OPA shaly facies. The sifted host rock powders at different grain size fractions (<63 μm and 63 < g.s. < 125 μm), at room (∼25 per cent) and 100 per cent humidity, were slid in double-direct shear configuration, under different normal stresses (5–70 MPa). We observe that peak friction, μpeak and steady-state friction, μss, depend on water vapour content and applied normal stress. Increasing relative humidity from ∼25 per cent RH (room humidity) to 100 per cent RH causes a decrease of frictional coefficient from 0.41 to 0.35. The analysis of velocity-steps in the light of rate-and-state friction framework shows that the stability parameter (a–b) is always positive (velocity-strengthening), and it increases with increasing sliding velocity and humidity. The dependence of (a–b) on slip rate is lost as normal stress increases, for each humidity condition. By monitoring the variations of the layer thickness during the velocity steps, we observe that dilation (Δh) is directly proportional to the sliding velocity, decreases with normal stress and is unaffected by humidity. Microstructural analysis shows that most of the deformation is accommodated within B-shear zones, and the increase of normal stress (σn) promotes the transition from strain localization and grain size reduction to distributed deformation on a well-developed phyllosilicate network. These results suggest that: (1) the progressive loss of velocity dependence of frictional stability parameter (a–b) at σn > 35 MPa is dictated by a transition from localized to distributed deformation and (2) water vapour content does not affect the deformation mechanisms and dilation, whereas it decreases steady-state friction (μss), and enhances fault stability.

58 GEOSCIENCES↗

DOE OSTI2014

SU-E-J-178: A Normalization Method Can Remove Discrepancy in Ventilation Function Due to Different Breathing Patterns

Purpose: To develop a normalization method to remove discrepancy in ventilation function due to different breathing patterns. Methods: Twenty five early stage non-small cell lung cancer patients were included in this study. For each patient, a ten phase 4D-CT and the voluntarily maximum inhale and exhale CTs were acquired clinically and retrospectively used for this study. For each patient, two ventilation maps were calculated from voxel-to-voxel CT density variations from two phases of the quiet breathing and two phases of the extreme breathing. For the quiet breathing, 0% (inhale) and 50% (exhale) phases from 4D-CT were used. An in-house tool was developed to calculate and display the ventilation maps. To enable normalization, the whole lung of each patient was evenly divided into three parts in the longitude direction at a coronal image with a maximum lung cross section. The ratio of cumulated ventilation from the top one-third region to the middle one-third region of the lung was calculated for each breathing pattern. Pearson's correlation coefficient was calculated on the ratios of the two breathing patterns for the group. Results: For each patient, the ventilation map from the quiet breathing was different from that of the extreme breathing. When the cumulative ventilation was normalized to the middle one-third of the lung region for each patient, the normalized ventilation functions from the two breathing patterns were consistent. For this group of patients, the correlation coefficient of the normalized ventilations for the two breathing patterns was 0.76 (p < 0.01), indicating a strong correlation in the ventilation function measured from the two breathing patterns. Conclusion: For each patient, the ventilation map is dependent of the breathing pattern. Using a regional normalization method, the discrepancy in ventilation function induced by the different breathing patterns thus different tidal volumes can be removed.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2014

SU-E-T-568: Improving Normal Brain Sparing with Increasing Number of Arc Beams for Volume Modulated Arc Beam Radiosurgery of Multiple Brain Metastases

Purpose: Intensity modulated arc beams have been newly reported for treating multiple brain metastases. The purpose of this study was to determine the variations in the normal brain doses with increasing number of arc beams for multiple brain metastases treatments via the TrueBeam Rapidarc system (Varian Oncology, Palo Alto, CA). Methods: A patient case with 12 metastatic brain lesions previously treated on the Leksell Gamma Knife Perfexion (GK) was used for the study. All lesions and organs at risk were contoured by a senior radiation oncologist and treatment plans for a subset of 3, 6, 9 and all 12 targets were developed for the TrueBeam Rapidarc system via 3 to 7 intensity modulated arc-beams with each target covered by at least 99% of the prescribed dose of 20 Gy. The peripheral normal brain isodose volumes as well as the total beam-on time were analyzed with increasing number of arc beams for these targets. Results: All intensisty modulated arc-beam plans produced efficient treatment delivery with the beam-on time averaging 0.6–1.5 min per lesion at an output of 1200 MU/min. With increasing number of arc beams, the peripheral normal brain isodose volumes such as the 12-Gy isodose line enclosed normal brain tissue volumes were on average decreased by 6%, 11%, 18%, and 28% for the 3-, 6-, 9-, 12-target treatment plans respectively. The lowest normal brain isodose volumes were consistently found for the 7-arc treatment plans for all the cases. Conclusion: With nearly identical beam-on times, the peripheral normal brain dose was notably decreased when the total number of intensity modulated arc beams was increased when treating multiple brain metastases. Dr Sahgal and Dr Ma are currently serving on the board of international society of stereotactic radiosurgery.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2015

SU-E-J-212: MR Diffusion Tensor Imaging for Assessment of Tumor and Normal Brain Tissue Responses of Juvenile Pilocytic Astrocytoma Treated by Proton Therapy

Purpose: Diffusion tensor imaging (DTI) can measure molecular mobility at the cellular level, quantified by the apparent diffusion coefficient (ADC). DTI may also reveal axonal fiber directional information in the white matter, quantified by the fractional anisotropy (FA). Juvenile pilocytic astrocytoma (JPA) is a rare brain tumor that occurs in children and young adults. Proton therapy (PT) is increasingly used in the treatment of pediatric brain tumors including JPA. However, the response of both tumors and normal tissues to PT is currently under investigation. We report tumor and normal brain tissue responses for a pediatric case of JPA treated with PT assessed using DTI. Methods: A ten year old male with JPA of the left thalamus received passive scattered PT to a dose of 50.4 Gy (RBE) in 28 fractions. Post PT, the patient has been followed up in seven years. At each follow up, MRI imaging including DTI was performed to assess response. MR images were registered to the treatment planning CT and the GTV mapped onto each MRI. The GTV contour was then mirrored to the right side of brain through the patient’s middle line to represent normal brain tissue. ADC and FA were measured within the ROIs. Results: Proton therapy can completely spare contra lateral brain while the target volume received full prescribed dose. From a series of MRI ADC images before and after PT at different follow ups, the enhancement corresponding to GTV had nearly disappeared more than 2 years after PT. Both ADC and FA demonstrate that contralateral normal brain tissue were not affect by PT and the tumor volume reverted to normal ADC and FA values. Conclusion: DTI allowed quantitative evaluation of tumor and normal brain tissue responses to PT. Further study in a larger cohort is warranted.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2015

SU-E-T-573: Normal Tissue Dose Effect of Prescription Isodose Level Selection in Lung Stereotactic Body Radiation Therapy

Purpose: To evaluate dose fall-off in normal tissue for lung stereotactic body radiation therapy (SBRT) cases planned with different prescription isodose levels (IDLs), by calculating the dose dropping speed (DDS) in normal tissue on plans computed with both Pencil Beam (PB) and Monte-Carlo (MC) algorithms. Methods: The DDS was calculated on 32 plans for 8 lung SBRT patients. For each patient, 4 dynamic conformal arc plans were individually optimized for prescription isodose levels (IDL) ranging from 60% to 90% of the maximum dose with 10% increments to conformally cover the PTV. Eighty non-overlapping rind structures each of 1mm thickness were created layer by layer from each PTV surface. The average dose in each rind was calculated and fitted with a double exponential function (DEF) of the distance from the PTV surface, which models the steep- and moderate-slope portions of the average dose curve in normal tissue. The parameter characterizing the steep portion of the average dose curve in the DEF quantifies the DDS in the immediate normal tissue receiving high dose. Provided that the prescription dose covers the whole PTV, a greater DDS indicates better normal tissue sparing. The DDS were compared among plans with different prescription IDLs, for plans computed with both PB and MC algorithms. Results: For all patients, the DDS was found to be the lowest for 90% prescription IDL and reached a highest plateau region for 60% or 70% prescription. The trend was the same for both PB and MC plans. Conclusion: Among the range of prescription IDLs accepted by lung SBRT RTOG protocols, prescriptions to 60% and 70% IDLs were found to provide best normal tissue sparing.

61 RADIATION PROTECTION AND DOSIMETRY↗

DOE OSTI2016

SU-F-R-08: Can Normalization of Brain MRI Texture Features Reduce Scanner-Dependent Effects in Unsupervised Machine Learning?

Purpose: To reduce differences in features calculated from MRI brain scans acquired at different field strengths with or without Gadolinium contrast. Methods: Brain scans were processed for 111 epilepsy patients to extract hippocampus and thalamus features. Scans were acquired on 1.5 T scanners with Gadolinium contrast (group A), 1.5T scanners without Gd (group B), and 3.0 T scanners without Gd (group C). A total of 72 features were extracted. Features were extracted from original scans and from scans where the image pixel values were rescaled to the mean of the hippocampi and thalami values. For each data set, cluster analysis was performed on the raw feature set and for feature sets with normalization (conversion to Z scores). Two methods of normalization were used: The first was over all values of a given feature, and the second by normalizing within the patient group membership. The clustering software was configured to produce 3 clusters. Group fractions in each cluster were calculated. Results: For features calculated from both the non-rescaled and rescaled data, cluster membership was identical for both the non-normalized and normalized data sets. Cluster 1 was comprised entirely of Group A data, Cluster 2 contained data from all three groups, and Cluster 3 contained data from only groups 1 and 2. For the categorically normalized data sets there was a more uniform distribution of group data in the three Clusters. A less pronounced effect was seen in the rescaled image data features. Conclusion: Image Rescaling and feature renormalization can have a significant effect on the results of clustering analysis. These effects are also likely to influence the results of supervised machine learning algorithms. It may be possible to partly remove the influence of scanner field strength and the presence of Gadolinium based contrast in feature extraction for radiomics applications.

60 APPLIED LIFE SCIENCES↗