Search NASA⌕ Search

SEARCH · Search NASA

Results for “LEAD 179”

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 19 records

Nuclear Data Sheets for A = 179

Nuclear structure data pertaining to all nuclei with mass number A = 179 (Yb, Lu, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl) have been compiled and evaluated, and incorporated into the ENSDF database. This evaluation for A = 179, which considers data available by 15 November 2008, supersedes the previous complete publication (E. Browne, Nuclear Data Sheets55, 483 (1988), cutoff September 1987) and the update-mode publication (C. Baglin, Nuclear Data Sheets72, 617 (1994), cutoff date September 1994), as well as the subsequent evaluations of Ta, Ir and Tl by C. Baglin (literature cutoff dates January 2000, October 1998 and February 2001, respectively). The newly evaluated literature provides new Coulomb excitation data for Hf, transfer reaction data for Hf and Ta, new information on Au and Hg from {alpha} decay and significant (HI,xn{gamma}) reaction data for Hf, Ta, Re, Au and Hg. At the time of the 1994 evaluation, no {gamma}-ray information was available for Au or Hg; that situation has now changed dramatically, although additional work on details of the low-lying structure of Au could be instructive.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Pitfalls of tungsten multileaf collimator in proton beam therapy

Purpose: Particle beam therapy is associated with significant startup and operational cost. Multileaf collimator (MLC) provides an attractive option to improve the efficiency and reduce the treatment cost. A direct transfer of the MLC technology from external beam radiation therapy is intuitively straightforward to proton therapy. However, activation, neutron production, and the associated secondary cancer risk in proton beam should be an important consideration which is evaluated. Methods: Monte Carlo simulation with FLUKA particle transport code was applied in this study for a number of treatment models. The authors have performed a detailed study of the neutron generation, ambient dose equivalent [H*(10)], and activation of a typical tungsten MLC and compared with those obtained from a brass aperture used in a typical proton therapy system. Brass aperture and tungsten MLC were modeled by absorber blocks in this study, representing worst-case scenario of a fully closed collimator. Results: With a tungsten MLC, the secondary neutron dose to the patient is at least 1.5 times higher than that from a brass aperture. The H*(10) from a tungsten MLC at 10 cm downstream is about 22.3 mSv/Gy delivered to water phantom by noncollimated 200 MeV beam of 20 cm diameter compared to 14 mSv/Gy for the brass aperture. For a 30-fraction treatment course, the activity per unit volume in brass aperture reaches 5.3 x 10{sup 4} Bq cm{sup -3} at the end of the last treatment. The activity in brass decreases by a factor of 380 after 24 h, additional 6.2 times after 40 days of cooling, and is reduced to background level after 1 yr. Initial activity in tungsten after 30 days of treating 30 patients per day is about 3.4 times higher than in brass that decreases only by a factor of 2 after 40 days and accumulates to 1.2 x 10{sup 6} Bq cm{sup -3} after a full year of operation. The daily utilization of the MLC leads to buildup of activity with time. The overall activity continues to increase due to {sup 179}Ta with a half-life of 1.82 yr and thus require prolonged storage for activity cooling. The H*(10) near the patient side of the tungsten block is about 100 {mu}Sv/h and is 27 times higher at the upstream side of the block. This would lead to an accumulated dose for therapists in a year that may exceed occupational maximum permissible dose (50 mSv/yr). The value of H*(10) at the upstream surface of the tungsten block is about 220 times higher than that of the brass. Conclusions: MLC is an efficient way for beam shaping and overall cost reduction device in proton therapy. However, based on this study, tungsten seems to be not an optimal material for MLC in proton beam therapy. Usage of tungsten MLC in clinic may create unnecessary risks associated with the secondary neutrons and induced radioactivity for patients and staff depending on the patient load. A careful selection of material for manufacturing of an optimal MLC for proton therapy is thus desired.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Extraction of effective solid-liquid interfacial free energies for full 3D solid crystallites from equilibrium MD simulations

Molecular dynamics simulations of an embedded atom copper system in the NPH ensemble are used to study the e ective solid-liquid interfacial free energy of quasispherical solid crystals within a liquid. This is within the larger context of MD simulations of this system undergoing solidi cation, where single individually-prepared crystallites of di erent sizes grow until they reach a thermodynamically stable nal state. The resulting equilibrium shapes possess the full structural details expected for solids with weakly anisotropic surface free energies (in these cases, ~5 % radial attening and rounded [111] octahedral faces). The simplifying assumption of sphericity and perfect isotropy leads to an e ective interfacial free energy as appearing in the Gibbs-Thomson equation, which we determine to be ~179 erg/cm 2 , roughly independent of crystal size for radii in the 50 - 250 A range. This quantity may be used in atomistically-informed models of solidi cation kinetics for this system.

36 MATERIALS SCIENCE↗

CHANDRA ACIS SURVEY OF X-RAY POINT SOURCES IN 383 NEARBY GALAXIES. I. THE SOURCE CATALOG

The Chandra data archive is a treasure for various studies, and in this paper we exploit this valuable resource to study the X-ray point source populations in nearby galaxies. By 2007 December 14, 383 galaxies within 40 Mpc with isophotal major axis above 1 arcmin had been observed by 626 public ACIS observations, most of which were for the first time analyzed by this survey to study the X-ray point sources. Uniform data analysis procedures are applied to the 626 ACIS observations and lead to the detection of 28,099 point sources, which belong to 17,599 independent sources. These include 8700 sources observed twice or more and 1000 sources observed 10 times or more, providing us a wealth of data to study the long-term variability of these X-ray sources. Cross-correlation of these sources with galaxy isophotes led to 8519 sources within the D{sub 25} isophotes of 351 galaxies, 3305 sources between the D{sub 25} and 2D{sub 25} isophotes of 309 galaxies, and additionally 5735 sources outside 2D{sub 25} isophotes of galaxies. This survey has produced a uniform catalog, by far the largest, of 11,824 X-ray point sources within 2D{sub 25} isophotes of 380 galaxies. Contamination analysis using the log N-log S relation shows that 74% of sources within 2D{sub 25} isophotes above 10{sup 39} erg s{sup -1}, 71% of sources above 10{sup 38} erg s{sup -1}, 63% of sources above 10{sup 37} erg s{sup -1}, and 56% of all sources are truly associated with galaxies. Meticulous efforts have identified 234 X-ray sources with galactic nuclei of nearby galaxies. This archival survey leads to 300 ultraluminous X-ray sources (ULXs) with L{sub X} (0.3-8 keV) {>=} 2 x 10{sup 39} erg s{sup -1}within D{sub 25} isophotes, 179 ULXs between D{sub 25} and 2D{sub 25} isophotes, and a total of 479 ULXs within 188 host galaxies, with about 324 ULXs truly associated with host galaxies based on the contamination analysis. About 4% of the sources exhibited at least one supersoft phase, and 70 sources are classified as ultraluminous supersoft sources with L{sub X} (0.3-8 keV) {>=} 2 x 10{sup 38} erg s{sup -1}. With a uniform data set and good statistics, this survey enables future works on various topics, such as X-ray luminosity functions for the ordinary X-ray binary populations in different types of galaxies, and X-ray properties of galactic nuclei.

79 ASTRONOMY AND ASTROPHYSICS↗

Pressure-Induced Superconductivity in Elemental Ytterbium Metal

Ytterbium (Yb) metal is divalent and nonmagnetic (4f 14 configuration). Under pressure its valence increases significantly leading to the expectation that magnetic instabilities and other highly correlated electron effects may appear before a stable trivalent state is reached (4f 13 configuration). We carried out electrical resistivity and ac magnetic susceptibility measurements to 179 GPa over the temperature range 1.4 - 295 K. No evidence for magnetic order is observed. However, Yb becomes a superconductor at 86 GPa with T c ≃1.4 K, increasing to 4.6 K at 179 GPa. X-ray absorption spectroscopy shows that Yb remains mixed-valent to at least 125 GPa, pointing to an active role of $f$ electrons in the emergence of superconductivity in this simple, elemental solid.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Avalanche boron fusion by laser picosecond block ignition with magnetic trapping for clean and economic reactor

Measured highly elevated gains of proton–boron (HB11) fusion (Picciotto et al. , Phys. Rev. X 4 , 031030 (2014)) confirmed the exceptional avalanche reaction process (Lalousis et al. , Laser Part. Beams 32 , 409 (2014); Hora et al. , Laser Part. Beams 33 , 607 (2015)) for the combination of the non-thermal block ignition using ultrahigh intensity laser pulses of picoseconds duration. The ultrahigh acceleration above $10^{20}~\text{cm}~\text{s}^{-2}$ for plasma blocks was theoretically and numerically predicted since 1978 (Hora, Physics of Laser Driven Plasmas (Wiley, 1981), pp. 178 and 179) and measured (Sauerbrey, Phys. Plasmas 3 , 4712 (1996)) in exact agreement (Hora et al. , Phys. Plasmas 14 , 072701 (2007)) when the dominating force was overcoming thermal processes. This is based on Maxwell’s stress tensor by the dielectric properties of plasma leading to the nonlinear (ponderomotive) force $f_{\text{NL}}$ resulting in ultra-fast expanding plasma blocks by a dielectric explosion. Combining this with measured ultrahigh magnetic fields and the avalanche process opens an option for an environmentally absolute clean and economic boron fusion power reactor. Finally, this is supported also by other experiments with very high HB11 reactions under different conditions (Labaune et al. , Nature Commun. 4 , 2506 (2013)).

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

A Catalog of GALEX Ultraviolet Emission from Asymptotic Giant Branch Stars

We have performed a comprehensive study of the UV emission detected from asymptotic giant branch (AGB) stars by the Galaxy Evolution Explorer (GALEX). Of the 468 AGB stars in our sample, 316 were observed by GALEX. In the near-UV (NUV) bandpass (λ{sub eff}∼2310 A-ring ), 179 AGB stars were detected and 137 were not detected. Only 38 AGB stars were detected in the far-UV (FUV) bandpass (λ{sub eff}∼1528 A-ring ). We find that NUV emission is correlated with optical to near-infrared emission, leading to higher detection fractions among the brightest, and hence closest, AGB stars. Comparing the AGB time-variable visible phased light curves to corresponding GALEX NUV phased light curves, we find evidence that for some AGB stars the NUV emission varies in phase with the visible light curves. We also find evidence that the NUV emission and possibly the FUV emission are anticorrelated with the circumstellar envelope density. These results suggest that the origin of the GALEX-detected UV emission is an inherent characteristic of the AGB stars that can most likely be traced to a combination of photospheric and chromospheric emission. In most cases, UV detections of AGB stars are not likely to be indicative of the presence of binary companions.

79 ASTRONOMY AND ASTROPHYSICS↗

Nanoporous separators for supercapacitor using activated carbon monolith electrode from oil palm empty fruit bunches

Activated porous carbon electrode prepared from fibres of oil palm empty fruit bunches was used for preparing the carbon based supercapacitor cells. The symmetrical supercapacitor cells were fabricated using carbon electrodes, stainless steel current collector, H{sub 2}SO{sub 4} electrolyte, and three types of nanoporous separators. Cells A, B and C were fabricated using polypropylene, eggshell membrane, and filter paper, respectively. Electrochemical characterizations data from Electrochemical Impedance Spectroscopy, Cyclic Voltammetry, and Galvanic Charge Discharge techniques showed that specific capacitance, specific power and specific energy for cell A were 122 F g{sup −1}, 177 W kg{sup −1}, 3.42 Wh kg{sup −1}, cell B; 125 F g{sup −1}, 179 W kg{sup −1}, and 3.64 Wh kg{sup −1}, and cell C; 180 F g{sup −1}, 178 W kg{sup −1}, 4.27 Wh kg{sup −1}. All the micrographs from Field Emission Scanning Electron Microscope showed that the different in nanoporous structure of the separators lead to a significant different in influencing the values of specific capacitance, power and energy of supercapacitors, which is associated with the mobility of ion into the pore network. These results indicated that the filter paper was superior than the eggshell membrane and polypropylene nanoporous separators. However, we found that in terms of acidic resistance, polypropylene was the best nanoporous separator for acidic medium.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Impact Behavior of Polyetheretherketone/Nickel Foam Co-continuous Composites

In this study, we prepared eight types of polyetheretherketone/nickel (PEEK/Ni) foam co-continuous phase composites (named “PEEK-Nis”) in which virgin PEEK was infiltrated into homologous hollow Ni structures by using hot pressing. Their Charpy impact behavior was investigated in accordance with ISO: 179-1, and the effect of the structure of the foam Ni on the impact properties of the PEEK-Nis was also studied. Compared with unfilled PEEK, the Ni prisms can effectively disperse, transfer, and absorb the energy of the pendulum, further improve the impact toughness of the PEEK-Nis and eventually lead to different fracture modes. The impact toughness of PEEK-Nis was mainly affected by Ni volume fraction, coating thickness, and pores per inch (PPI) value, which are mutually restricted. When the Ni content was less than 8.5%, the impact toughness of PEEK-Nis increased with an increase in PPI value. When the Ni content was between 8.5 and 18%, the impact toughness of PEEK-Nis increased with the coating thickness. When the Ni content exceeded 18%, the impact toughness decreased dramatically possibly because of the mechanical mismatch of the two-phase materials in PEEK-Nis. For further optimization, PEEK-NiB-2 was fabricated and had the most reasonable structure of the reinforcing phase. Among all PEEK-Nis, PEEK-NiB-2 exhibited the best impact toughness, with a value over two times that of unfilled PEEK.

36 MATERIALS SCIENCE↗

Dirac-Fock-Breit-Gaunt calculations for tungsten hexacarbonyl W(CO){sub 6}

The first all-electron fully relativistic Dirac-Fock-Breit-Gaunt (DFBG), Dirac–Fock (DF), and nonrelativistic (NR) Hartree-Fock (HF) calculations are reported for octahedral (O{sub h}) tungsten hexacarbonyl W(CO){sub 6}. Our DF and NR HF calculations predict atomization energy of 73.76 and 70.33 eV, respectively. The relativistic contribution of ∼3.4 eV to the atomization energy of W(CO){sub 6} is fairly significant. The DF and NR energy for the reaction W + 6CO → W(CO){sub 6} is calculated as −7.90 and −8.86 eV, respectively. The mean bond energy predicted by our NR and DF calculations is 142.5 kJ/mol and 177.5 kJ/mol, respectively, and our predicted DF mean bond energy is in excellent agreement with the experimental value of 179 kJ/mol quoted in the literature. The relativistic effects contribute ∼35 kJ/mol to the mean bond energy and the calculated BSSE is 1.6 kcal/mol, which indicates that the triple zeta basis set used here is fairly good. The mean bond energy and the atomization energy calculated in our DFBG SCF calculations, which include variationally both the relativistic and magnetic Breit effects, is 157.4 kJ/mol and 68.84 eV, respectively. The magnetic Breit effects lead to a decrease of ∼20 kJ/mol and ∼4.9 eV for the mean bond energy and atomization energy, respectively, for W(CO){sub 6}. Our calculated magnetic Breit interaction energy of −9.79 eV for the energy of reaction (ΔE) for W + 6CO → W(CO){sub 6} is lower by ∼1.90 eV as compared to the corresponding DF value (ΔE) and contributes significantly to the ΔE. A detailed discussion is presented of electronic structure, bonding, and molecular energy levels at various levels of theory for W(CO){sub 6}.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Life cycle greenhouse gas emissions and carbon intensity of U.S. fuel use and projection for the next 10 years-based on built capacity and expansion plans

The U.S. Inflation Reduction Act of 2022 supports biofuel production expansion through the 45Z clean fuel production tax credit, replacing previous 40A and 40B credits. This follows on the Renewable Fuel Standard from the Energy Policy Act of 2005 and its expansion in 2007. States like California, Oregon, and Washington also offer clean fuel credits. Meanwhile, federal agencies, including the U.S. Department of Energy, have advanced alternative fuel technologies through research and development funding. The surging interest in the biofuel industry has spurred the demand for biofuel supplies in the markets, although achieving profitability for advanced biofuels and low-carbon e-fuels remains challenging. This study aims to track U.S. alternative fuel production capacity expansion plans over the next 10 years and estimate impacts on greenhouse gas (GHG) emissions. By tracking built capacity and industry announcements of planned expansion, this study complements other studies which use models to predict changes in energy technologies and the associated GHG implications. Modeled projections of future technologies are often criticized for over or underestimating the cost and potential role of new technologies. The study focuses on sustainable aviation fuel, renewable diesel, ethanol, biodiesel, and renewable natural gas. Using facility-level data, we conducted a bottom-up analysis linking biofuel production pathways with corresponding pathways and parameterizations in the Argonne R&D GREET model. Results indicate that biofuel capacity could reach 3.8 exajoules in 2035, potentially reducing U.S. GHG emissions by 179 million tonnes, including the full life cycle. This corresponds to a 20% reduction in transportation and 5% in industry sector emissions by 2035, or a 3.6% reduction in economy-wide emissions. Overall, this study shows that while biofuel production capacity in the U.S. is expanding, the capacities remain limited compared to fuel demand. Uncertainty regarding the durability and extension of incentives may be dampening the pace of growth. Meanwhile, demonstrating the commercial potential for alternative fuels and climbing the learning curve for new technologies could lead to an increased pace of expansion in later years. This study offers insights for bioenergy stakeholders, highlighting biofuel technologies' contribution to U.S. energy system and emissions reduction over time based on producers' plans.

Biofuel Producers↗

Reinforcement Learning as a Parsimonious Alternative to Prediction Cascades: A Case Study on Image Segmentation

Deep learning architectures have achieved state-of-the-art (SOTA) performance on computer vision tasks such as object detection and image segmentation. This may be attributed to the use of over-parameterized, monolithic deep learning architectures executed on large datasets. Although such large architectures lead to increased accuracy, this is usually accompanied by a larger increase in computation and memory requirements during inference. While this is a non-issue in traditional machine learning (ML) pipelines, the recent confluence of machine learning and fields like the Internet of Things (IoT) has rendered such large architectures infeasible for execution in low-resource settings. For some datasets, large monolithic pipelines may be overkill for simpler inputs. To address this problem, previous efforts have proposed decision cascades where inputs are passed through models of increasing complexity until the desired performance is achieved. However, we argue that cascaded prediction leads to sub-optimal throughput and increased computational cost due to wasteful intermediate computations. To address this, we propose PaSeR (Parsimonious Segmentation with Reinforcement Learning) a non-cascading, cost-aware learning pipeline as an efficient alternative to cascaded decision architectures. Through experimental evaluation on both real-world and standard datasets, we demonstrate that PaSeR achieves better accuracy while minimizing computational cost relative to cascaded models. Further, we introduce a new metric IoU/GigaFlop to evaluate the balance between cost and performance. On the real-world task of battery material phase segmentation, PaSeR yields 179% improvement over SOTA MatPhase model and a 196% improvement over IDK Cascades under the IoU/GigaFlop metric. We also demonstrate PaSeR’s adaptability to complementary models trained on a noisy MNIST dataset, where it outperforms all baselines on IoU/GigaFlop by an average of 44%.

Srikshan, Bharat↗

Effect of residual stress on the ferroelectric property of (Pb{sub 0.90}La{sub 0.10})Ti{sub 0.975}O{sub 3} thin films

Highly (001)-, (100)-, (101)-, and (110)-oriented (Pb{sub 0.90}La{sub 0.10})Ti{sub 0.975}O{sub 3} (PLT) thin films were deposited on the LaNiO{sub 3}(001)/Pt(111)/Ti/SiO{sub 2}/Si(100), LaNiO{sub 3}/SiO{sub 2}/Si(100), PbTiO{sub 3}/Pt(111)/Ti/SiO{sub 2}/Si(100), and LaNiO{sub 3}(110)/Pt(111)/Ti/SiO{sub 2}/Si(100) substrates by rf magnetron sputtering. The orientation dependence of ferroelectric properties of the PLT films was investigated. The result shows that the (001)-oriented PLT thin films exhibit enhanced ferroelectric properties (2P{sub r}=61.1 {mu}C/cm{sup 2} and 2E{sub c}=179 kV/cm). The residual stress of the PLT thin films with different orientations was measured by x-ray diffraction (sin{sup 2} {psi} method) for illuminating the related physical mechanisms. The result shows that the enhanced ferroelectric properties of (001)-oriented PLT thin films should be attributed to low residual stress.

36 MATERIALS SCIENCE↗

Phosphors for near UV-Emitting LED's for Efficacious Generation of White Light

1) We studied phosphors for near-UV (nUV) LED application as an alternative to blue LEDs currently being used in SSL systems. We have shown that nUV light sources could be very efficient at high current and will have significantly less binning at both the chip and phosphor levels. We identified phosphor blends that could yield 4100K lamps with a CRI of approximately 80 and LPWnUV,opt equal to 179 for the best performing phosphor blend. Considering the fact that the lamps were not optimized for light coupling, the results are quite impressive. The main bottleneck is an optimum blue phosphor with a peak near 440 nm with a full width half maximum of about 25 nm and a quantum efficiency of >95%. Unfortunately, that may be a very difficult task when we want to excite a phosphor at ~400 nm with a very small margin for Stokes shift. Another way is to have all the phosphors in the blend having the excitation peak at 400 nm or slightly shorter wavelength. This could lead to a white light source with no body color and optimum efficacy due to no self-absorption effects by phosphors in the blend. This is even harder than finding an ideal blue phosphor, but not necessarily impossible. 2) With the phosphor blends identified, light sources using nUV LEDs at high current could be designed with comparable efficacy to those using blue LEDs. It will allow us to design light sources with multiple wattages using the same chips and phosphor blends simply by varying the input current. In the case of blue LEDs, this is not currently possible because varying the current will lower the efficacy at high current and alter the color point. With improvement of phosphor blends, control over CRI could improve. Less binning at the chip level and also at the phosphor blend level could reduce the cost of SSL light sources. 3) This study provided a deeper understanding of phosphor characteristics needed for LEDs in general and nUV LEDs in particular. Two students received Ph.D. degrees and three undergraduates participated in this work. Two of the undergraduate students are now in graduate school. The results were widely disseminated – 20 archival journal publications (published, accepted or in preparation) and three conference proceedings resulted. The students presented their work at 11 different national/international conferences (32 oral or poster presentations) and the PI’s delivered 12 invited, keynote or plenary lectures.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Glauber model for {alpha}-nucleus total reaction cross section

The Coulomb-modified Glauber model is employed to calculate the total reaction cross section ({sigma}{sub R}) for {alpha} particles from {sup 9}Be, {sup 12}C, {sup 16}O, {sup 28}Si, {sup 40}Ca, {sup 58,60}Ni, {sup 112,116,120,124}Sn, and {sup 208}Pb at 117.2, 163.9, and 192.4 MeV and from the lighter nuclei also at 69.6 MeV. Our main focus in this work is to assess the suitability of semiphenomenological parametrization of the NN amplitude (SPNN), used recently [Deeksha Chauhan and Z. A. Khan, Eur. Phys. J. A 41, 179 (2009)], in the analysis of {sigma}{sub R} at the energies under consideration. Using the realistic form factors for the colliding nuclei, it is found that the SPNN works reasonably well and we have quite a satisfactory account of the {sigma}{sub R} data in all the cases. Moreover, our analysis suggests that the SPNN could be taken as fairly stable to describe simultaneously the elastic angular distribution and the {sigma}{sub R} for a wide range of target nuclei in the relatively low-energy region.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

PLD2 has both enzymatic and cell proliferation-inducing capabilities, that are differentially regulated by phosphorylation and dephosphorylation

Phospholipase D2 (PLD2) overexpression in mammalian cells results in cell transformation. We have hypothesized that this is due to an increase of de novo DNA synthesis. We show here that overexpression of PLD2-WT leads to an increased DNA synthesis, as measured by the expression levels of the proliferation markers PCNA, p27{sup KIP1} and phospho-histone-3. The enhancing effect was even higher with phosphorylation-deficient PLD2-Y179F and PLD2-Y511F mutants. The mechanism for this did not involve the enzymatic activity of the lipase, but, rather, the presence of the protein tyrosine phosphatase CD45, as silencing with siRNA for CD45 abrogated the effect. The two Y{yields}F mutants had in common a YxN consensus site that, in the phosphorylated counterparts, could be recognized by SH2-bearing proteins, such as Grb2. Even though Y179F and Y511F cannot bind Grb2, they could still find other protein partners, one of which, we have reasoned, could be CD45 itself. Affinity purified PLD2 is indeed activated by Grb2 and deactivated by CD45 in vitro. We concluded that phosphorylated PLD2, aided by Grb2, mediates lipase activity, whereas dephosphorylated PLD2 mediates an induction of cell proliferation, and the specific residues involved in this newly discovered regulation of PLD2 are Y{sup 179} and Y{sup 511}.

60 APPLIED LIFE SCIENCES↗

Deep learning and radiomics: the utility of Google TensorFlow™ Inception in classifying clear cell renal cell carcinoma and oncocytoma on multiphasic CT

Purpose: Currently, all solid enhancing renal masses without microscopic fat are considered malignant until proven otherwise and there is substantial overlap in the imaging findings of benign and malignant renal masses, particularly between clear cell RCC (ccRCC) and benign oncocytoma (ONC). Radiomics has attracted increased attention for its utility in pre-operative work-up on routine clinical images. Radiomics based approaches have converted medical images into mineable data and identified prognostic imaging signatures that machine learning algorithms can use to construct predictive models by learning the decision boundaries of the underlying data distribution. The TensorFlow™ framework from Google is a state-of-the-art open-source software library that can be used for training deep learning neural networks for performing machine learning tasks. The purpose of this study was to investigate the diagnostic value and feasibility of a deep learning-based renal lesion classifier using open-source Google TensorFlow™ Inception in differentiating ccRCC from ONC on routine four-phase MDCT in patients with pathologically confirmed renal masses. Methods: With institutional review board approval for this 1996 Health Insurance Portability and Accountability Act compliant retrospective study and a waiver of informed consent, we queried our institution’s pathology, clinical, and radiology databases for histologically proven cases of ccRCC and ONC obtained between January 2000 and January 2016 scanned with a an intravenous contrast-enhanced four-phase renal mass protocol (unenhanced (UN), corticomedullary (CM), nephrographic (NP), and excretory (EX) phases). To extract features to be used for the machine learning model, the entire renal mass was contoured in the axial plane in each of the four phases, resulting in a 3D volume of interest (VOI) representative of the entire renal mass. We investigated thirteen different approaches to convert the acquired VOI data into a set of images that adequately represented each tumor which was used to train the final layer of the neural network model. Training was performed over 4000 iterations. In each iteration, 90% of the data were designated as training data and the remaining 10% served as validation data and a leave-one-out cross-validation scheme was implemented. Accuracy, sensitivity, specificity, positive (PPV) and negative predictive (NPV) values, and CIs were calculated for the classification of the thirteen processing modes. Results: We analyzed 179 consecutive patients with 179 lesions (128 ccRCC and 51 ONC). The ccRCC cohort had a mean size of 3.8 cm (range 0.8–14.6 cm) and the ONC cohort had a mean lesion size of 3.9 cm (range 1.0–13.1 cm). The highest specificity and PPV (52.9% and 80.3%, respectively) were achieved in the EX phase when we analyzed the single mid-slice of the tumor in the axial, coronal and sagittal plane, and when we increased the number of mid-slices of the tumor to three, with an accuracy of 75.4%, which also increased the sensitivity to 88.3% and the PPV to 79.6%. Using the entire tumor volume also showed that classification performance was best in the EX phase with an accuracy of 74.4%, a sensitivity of 85.8% and a PPV of 80.1%. When the entire tumor volume, plus mid-slices from all phases and all planes presented as tiled images, were submitted to the final layer of the neural network we achieved a PPV of 82.5%. Conclusions: The best classification result was obtained in the EX phase among the thirteen classification methods tested. Our proof of concept study is the first step towards understanding the utility of machine learning in the differentiation of ccRCC from ONC on routine CT images. We hope this could lead to future investigation into the development of a multivariate machine learning model which may augment our ability to accurately predict renal lesion histology on imaging.

62 RADIOLOGY AND NUCLEAR MEDICINE↗