Search NASA⌕ Search

SEARCH · Search NASA

Results for “false data”

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 1,315 records · Page 73

Robo-AO Kepler Survey. IV. The Effect of Nearby Stars on 3857 Planetary Candidate Systems

We present the overall statistical results from the Robo-AO Kepler planetary candidate survey, comprising of 3857 high-angular resolution observations of planetary candidate systems with Robo-AO, an automated laser adaptive optics system. These observations reveal previously unknown nearby stars blended with the planetary candidate host stars that alter the derived planetary radii or may be the source of an astrophysical false positive transit signal. In the first three papers in the survey, we detected 440 nearby stars around 3313 planetary candidate host stars. In this paper, we present observations of 532 planetary candidate host stars, detecting 94 companions around 88 stars; 84 of these companions have not previously been observed in high resolution. We also report 50 more-widely separated companions near 715 targets previously observed by Robo-AO. We derive corrected planetary radius estimates for the 814 planetary candidates in systems with a detected nearby star. If planetary candidates are equally likely to orbit the primary or secondary star, the radius estimates for planetary candidates in systems with likely bound nearby stars increase by a factor of 1.54, on average. We find that 35 previously believed rocky planet candidates are likely not rocky due to the presence of nearby stars. From the combined data sets from the complete Robo-AO KOI survey, we find that 14.5 ± 0.5% of planetary candidate hosts have a nearby star with 4″, while 1.2% have two nearby stars, and 0.08% have three. We find that 16% of Earth-sized, 13% of Neptune-sized, 14% of Saturn-sized, and 19% of Jupiter-sized planet candidates have detected nearby stars.

79 ASTRONOMY AND ASTROPHYSICS↗

AmeriFlux US-Ton Tonzi Ranch

This is the AmeriFlux version of the carbon flux data for the site US-Ton Tonzi Ranch. Site Description - The Tonzi Ranch is located on the lower foothills of the Sierra Nevada Mountains. The site is classified as an oak savanna woodland (Davis et al, 2016) and is situated on privately owned land. This site is representative of the oak savanna ecosystem that rings the Great Central Valley of California and inhabits the lower reaches of the Sierra Nevada foothills. This ecosystem is a natural laboratory for studying how ecosystem scale carbon and water fluxes respond to environmental change. The temperatures range between 0 and 40 C, the autumn, winter and spring are wet and cool, while the summer growing season is hot, dry and rainless. The overstory is dominated by blue oak trees (40% of total vegetation) with intermittent grey pine trees (3 trees/ha). Understory species include a variety of grasses and herbs, including purple false brome, smooth cat's ear, and rose clover. These two distinctive layers operate in and out of phase from one another. Growing season of the understory is confined to the wet season only, typically from October to early May. In contrast, the deciduous blue oak trees are dormant and leafless during the rainy winter months. The trees tend to leaf out in March and reach maximum LAI in April. During the rainless summers, the grasses in the understory die and the leaves experience soil moisture deficit stress References Davis, F., et al. (2016). "Oak woodlands." Ecosystems of California: 509–529.

Ma, Siyan↗

Sensitive and Specific Peak Detection for SELDI-TOF Mass Spectrometry Using a Wavelet/Neural-Network Based Approach

SELDI-TOF mass spectrometer’s compact size and automated, high throughput design have been attractive to clinical researchers, and the platform has seen steady-use in biomarker studies. Despite new algorithms and preprocessing pipelines that have been developed to address reproducibility issues, visual inspection of the results of SELDI spectra preprocessing by the best algorithms still shows miscalled peaks and systematic sources of error. This suggests that there continues to be problems with SELDI preprocessing. In this work, we study the preprocessing of SELDI in detail and introduce improvements. While many algorithms, including the vendor supplied software, can identify peak clusters of specific mass (or m/z) in groups of spectra with high specificity and low false discover rate (FDR), the algorithms tend to underperform estimating the exact prevalence and intensity of peaks in those clusters. Thus group differences that at first appear very strong are shown, after careful and laborious hand inspection of the spectra, to be less than significant. Here we introduce a wavelet/neural network based algorithm which mimics what a team of expert, human users would call for peaks in each of several hundred spectra in a typical SELDI clinical study. The wavelet denoising part of the algorithm optimally smoothes the signal in each spectrum according to an improved suite of signal processing algorithms previously reported (the LibSELDI toolbox under development). The neural network part of the algorithm combines those results with the raw signal and a training dataset of expertly called peaks, to call peaks in a test set of spectra with approximately 95% accuracy. The new method was applied to data collected from a study of cervical mucus for the early detection of cervical cancer in HPV infected women. The method shows promise in addressing the ongoing SELDI reproducibility issues.

59 BASIC BIOLOGICAL SCIENCES↗

False vacuum energy dominated inflation with large r and the importance of κ{sub s}

We investigate to which extent and under which circumstances false vacuum energy (V{sub 0}) dominated slow-roll inflation is compatible with a large tensor-to-scalar ratio r = O(0.1), as indicated by the recent BICEP2 measurement. With V{sub 0} we refer to a constant contribution to the inflaton potential, present before a phase transition takes place and absent in the true vacuum of the theory, like e.g. in hybrid inflation. Based on model-independent considerations, we derive an upper bound on the possible amount of V{sub 0} domination and highlight the importance of higher-order runnings of the scalar spectral index (beyond α{sub s}) in order to realise scenarios of V{sub 0} dominated inflation. We study the conditions for V{sub 0} domination explicitly with an inflaton potential reconstruction around the inflaton field value 50 e-folds before the end of inflation, taking into account the present observational data. To this end, we provide the up-to-date parameter constraints within ΛCDM + r + α{sub s} + κ{sub s} using the cosmological parameter estimation code Monte Python together with the Boltzmann code CLASS.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

ASSESSING THE MILKY WAY SATELLITES ASSOCIATED WITH THE SAGITTARIUS DWARF SPHEROIDAL GALAXY

Numerical models of the tidal disruption of the Sagittarius (Sgr) dwarf galaxy have recently been developed that for the first time simultaneously satisfy most observational constraints on the angular position, distance, and radial velocity trends of both leading and trailing tidal streams emanating from the dwarf. We use these dynamical models in combination with extant three-dimensional position and velocity data for Galactic globular clusters and dSph galaxies to identify those Milky Way satellites that are likely to have originally formed in the gravitational potential well of the Sgr dwarf, and have been stripped from Sgr during its extended interaction with the Milky Way. We conclude that the globular clusters Arp 2, M 54, NGC 5634, Terzan 8, and Whiting 1 are almost certainly associated with the Sgr dwarf, and that Berkeley 29, NGC 5053, Pal 12, and Terzan 7 are likely to be as well (albeit at lower confidence). The initial Sgr system therefore may have contained five to nine globular clusters, corresponding to a specific frequency S{sub N} = 5-9 for an initial Sgr luminosity M{sub V} = -15.0. Our result is consistent with the 8 {+-} 2 genuine Sgr globular clusters expected on the basis of statistical modeling of the Galactic globular cluster distribution and the corresponding false-association rate due to chance alignments with the Sgr streams. The globular clusters identified as most likely to be associated with Sgr are consistent with previous reconstructions of the Sgr age-metallicity relation, and show no evidence for a second-parameter effect shaping their horizontal branch morphologies. We find no statistically significant evidence to suggest that any of the recently discovered population of ultrafaint dwarf galaxies are associated with the Sgr tidal streams, but are unable to rule out this possibility conclusively for all systems.

79 ASTRONOMY AND ASTROPHYSICS↗

Multichannel response analysis on 2D projection views for detection of clustered microcalcifications in digital breast tomosynthesis

Purpose: To investigate the feasibility of a new two-dimensional (2D) multichannel response (MCR) analysis approach for the detection of clustered microcalcifications (MCs) in digital breast tomosynthesis (DBT). Methods: With IRB approval and informed consent, a data set of two-view DBTs from 42 breasts containing biopsy-proven MC clusters was collected in this study. The authors developed a 2D approach for MC detection using projection view (PV) images rather than the reconstructed three-dimensional (3D) DBT volume. Signal-to-noise ratio (SNR) enhancement processing was first applied to each PV to enhance the potential MCs. The locations of MC candidates were then identified with iterative thresholding. The individual MCs were decomposed with Hermite–Gaussian (HG) and Laguerre–Gaussian (LG) basis functions and the channelized Hotelling model was trained to produce the MCRs for each MC on the 2D images. The MCRs from the PVs were fused in 3D by a coincidence counting method that backprojects the MC candidates on the PVs and traces the coincidence of their ray paths in 3D. The 3D MCR was used to differentiate the true MCs from false positives (FPs). Finally a dynamic clustering method was used to identify the potential MC clusters in the DBT volume based on the fact that true MCs of clinical significance appear in clusters. Using two-fold cross validation, the performance of the 3D MCR for classification of true and false MCs was estimated by the area under the receiver operating characteristic (ROC) curve and the overall performance of the MCR approach for detection of clustered MCs was assessed by free response receiver operating characteristic (FROC) analysis. Results: When the HG basis function was used for MCR analysis, the detection of MC cluster achieved case-based test sensitivities of 80% and 90% at the average FP rates of 0.65 and 1.55 FPs per DBT volume, respectively. With LG basis function, the average FP rates were 0.62 and 1.57 per DBT volume at the same sensitivity levels. The difference in the two sets of basis functions for detection of MCs did not show statistical significance. Conclusions: The authors' experimental results indicate that the MCR approach is promising for the detection of MCs on PV images. The HG or LG basis functions are both effective in characterizing the signal response of MCs using the channelized Hotelling model. The coincidence counting method for fusion of the 2D MCR in 3D is an important step for FP reduction. Further study is underway to improve the MCR approach for microcalcification detection in DBT.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Defining the Clinical Target Volume for Bladder Cancer Radiotherapy Treatment Planning

Purpose: There are currently no data for the expansion margin required to define the clinical target volume (CTV) around bladder tumors. This information is particularly relevant when perivesical soft tissue changes are seen on the planning scan. While this appearance may reflect extravesical extension (EVE), it may also be an artifact of previous transurethral resection (TUR). Methods and Materials: Eighty patients with muscle-invasive bladder cancer who had undergone radical cystectomy were studied. All patients underwent preoperative TUR and staging computed tomography (CT) scans. The presence and extent of tumor growth beyond the outer bladder wall was measured radiologically and histopathologically. Results: Forty one (51%) patients had histologically confirmed tumor extension into perivesical fat. The median and mean extensions beyond the outer bladder wall were 1.7 and 3.1 mm, respectively. Thirty five (44%) patients had EVE, as seen on CT scans. The sensitivity and specificity of CT scans for EVE were 56% and 79%, respectively. False-positive results were infrequent and not affected by either the timing or the amount of tissue resected at TUR. CT scans consistently tended to overestimate the extent of EVE. Tumor size and the presence of either lymphovascular invasion or squamoid differentiation predict a greater extent of EVE. Conclusions: In patients with radiological evidence of extravesical disease, the CTV should comprise the outer bladder wall plus a 10-mm margin. In patients with no evidence of extravesical disease on CT scans, the CTV should be restricted to the outer bladder wall plus a 6-mm margin. These recommendations would encompass microscopic disease extension in 90% of cases.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

A Machine Learning Approach to Objective Identification of Dust in Satellite Imagery

Airborne dust has broad adverse effects on human activity, including aviation, human health, and agriculture. Remote sensing observations are used to detect dust and aerosols in the atmosphere using long established techniques. False color Red-Green-Blue (RGB) imagery using band differences sensitive to dust absorption (Dust RGB) is currently used operationally to assist forecasters and decision-makers in identifying dust at night, but there are still limitations, subjectivity, and nuances to image interpretation making night-time dust identification difficult even for experts. This study applies machine learning to the problem of night-time dust detection with a simple random forest (RF) model using Geostationary Operational Environmental Satellite-16 (GOES-16) Advanced Baseline Imager (ABI) infrared imagery, band differences sensitive to dust absorption, and Dust RGB color components as inputs to the model. The RF model achieves an Area-Under-Curve (AUC) of 0.97 with a standard deviation of 0.04 for dust cases. For images with dust present, the model correctly labels 85% of dust pixels and 99.96% of no-dust pixels for all dust images in the validation data set. The addition of a single null case to the training data set drastically reduces error in labeling no-dust pixels as dust from 45% to 14.5%. Application of the machine learning model to the April 13–14, 2019 dust event demonstrates the ability of the model to identify dust during night-time hours when visual dust detection is limited by the cooling ground surface characteristics.

dust↗

Exposing and Reducing Biases of Simulating Mixed-Phase Clouds in the Convection-Permitting E3SM Atmosphere Model: Lessons From an Arctic Cold-Air Outbreak

Mixed-phase clouds modulate the water and energy cycles of high-latitude regions, yet their liquid-ice phase partitioning has long been poorly simulated in climate models. Here, simulations of Arctic mixed-phase clouds by the Simple Cloud-Resolving E3SM Atmosphere Model (SCREAM) are assessed against large-eddy simulations, satellite data, and ground-based observations during the Cold-Air Outbreaks in the Marine Boundary Layer Experiment field campaign. SCREAM simulates nearly completely frozen clouds, which is attributed largely to the unreasonably strong Wegener–Bergeron–Findeisen (WBF) process that converts liquid to ice excessively and partly to the early over-abundant ice production at cold temperatures from a temperature-deterministic deposition ice nucleation scheme. Assuming no subgrid variation for the WBF process in the original formulation particularly conflicts with the instantaneous saturation adjustment assumption in the condensation scheme that assumes subgrid variability, leading to exaggerated WBF process rates. A proposed simple physically-based improvement on the treatment of subgrid cloud overlap substantially increases supercooled liquid water content and notably improves cloud-top phase partitioning, aligning better with observations. Improvement of supercooled liquid water content also converges with increasing horizontal resolution. The deposition ice nucleation scheme is found responsible for a falsely-produced ice cloud aloft that is not observed, biasing the simulated cloud radiative effects and top-of-atmosphere radiative fluxes. This study identifies key deficiencies in cloud parameterizations that continue to challenge convection-permitting models.

Geosciences↗

AmeriFlux FLUXNET-1F US-Ton Tonzi Ranch

This is the AmeriFlux Management Project (AMP) created FLUXNET-1F version of the carbon flux data for the site US-Ton Tonzi Ranch. This is the FLUXNET version of the carbon flux data for the site US-Ton Tonzi Ranch produced by applying the standard ONEFlux (1F) software. Site Description - The Tonzi Ranch is located on the lower foothills of the Sierra Nevada Mountains. The site is classified as an oak savanna woodland (Davis et al, 2016) and is situated on privately owned land. This site is representative of the oak savanna ecosystem that rings the Great Central Valley of California and inhabits the lower reaches of the Sierra Nevada foothills. This ecosystem is a natural laboratory for studying how ecosystem scale carbon and water fluxes respond to environmental change. The temperatures range between 0 and 40 C, the autumn, winter and spring are wet and cool, while the summer growing season is hot, dry and rainless. The overstory is dominated by blue oak trees (40% of total vegetation) with intermittent grey pine trees (3 trees/ha). Understory species include a variety of grasses and herbs, including purple false brome, smooth cat's ear, and rose clover. These two distinctive layers operate in and out of phase from one another. Growing season of the understory is confined to the wet season only, typically from October to early May. In contrast, the deciduous blue oak trees are dormant and leafless during the rainy winter months. The trees tend to leaf out in March and reach maximum LAI in April. During the rainless summers, the grasses in the understory die and the leaves experience soil moisture deficit stress References Davis, F., et al. (2016). "Oak woodlands." Ecosystems of California: 509–529.

Ma, Siyan↗

Analysis study of multispectral data, ERTS-A, from an area in West Pakistan

The author has identified the following significant results. Geomorphic features related to a known porphyry copper deposit at Saindak, western Chagai District, Pakistan, are easily distinguished on ERTS-1 images. No new prospecting sites were recognized on the basis of the images alone, but new information from the images was used in conjunction with known geology to evaluate two previously known speculative areas and to suggest one additional one. The study showed that Saindak-type deposits are less likely in some extensive areas of the Chagai District than formerly believed. The Saindak deposit is in an area of relatively soft folded sedimentary and volcanic rocks. The deposit is characterized by a linear zone of easily eroded sulfide-rich rock surrounded by a resistant rim of hornfels and propylitically altered rock. Both this rim and the central sulfide-rich valley are conspicuous features on the images. Swarms of dikes are probably useful for distinguishing real rims from other resistant rock types, but there is no expression of them on the image, although they are easily seen on aerial photographs of the Saindak rim. During field mapping, patches of strong red and yellow hue related to the mineral natrojarosite were noted in the central valley. Attempts to detect a color anomaly using simple false color composites were not successful.

Schmidt, R. G.↗

Validation of a treatment plan-based calibration method for 2D detectors used for treatment delivery quality assurance

Purpose: Dosimetry using film, CR, electronic portal imaging, or other 2D detectors requires calibration of the raw image data to obtain dose. Typically, a series of known doses are given to the detector, the raw signal for each dose is obtained, and a calibration curve is created. This calibration curve is then applied to the measured raw signals to convert them to dose. With the advent of IMRT, film dosimetry for quality assurance has become a routine and labor intensive part of the physicist's day. The process of calibrating the film or other 2D detector takes time and additional film or images for performing the calibration, and comes with its own source of errors. This article studies a new methodology for the relative dose calibration of 2D imaging detectors especially useful for IMRT QA, which relies on the treatment plan dose image to provide the dose information which is paired with the raw QA image data after registration of the two images (plan-based calibration). Methods: Validation of the accuracy and robustness of the method is performed on ten IMRT cases performed using EDR2 film with conventional and plan-based calibration. Also, for each of the ten cases, a 5 mm registration error was introduced and the Gamma analysis was reevaluated. In addition, synthetic image tests were performed to test the limits of the method. The Gamma analysis is used as a measure of dosimetric agreement between plan and film for the clinical cases and a dose difference metric for the synthetic cases. Results: The QA image calibrated by the plan-based method was found to more accurately match the treatment plan doses than the conventionally calibrated films and also to reveal dose errors more effectively when a registration error was introduced. When synthetic acquired images were systematically studied, localized and randomly placed dose errors were correctly identified without excessive falsely passing or falsely failing pixels, unless the errors were concentrated in a majority of pixels in a contiguous narrow dose band. Irregularities seen in the calibration curve expose these errors. Conclusions: The plan-based calibration method was found to be an accurate, efficient procedure, capable of detecting IMRT QA relative dosimetry errors as well as, or better than conventional calibration methods.

07 ISOTOPE AND RADIATION SOURCES↗

A New View of the Lunar South Pole from the Lunar Orbiter Laser Altimeter (LOLA)

Upcoming missions to the lunar south pole require detailed maps over large areas to fully characterize landing sites locally and regionally and to place their data into proper geologic context. To that end, we enhanced the Lunar Orbiter Laser Altimeter (LOLA) altimetry dataset for the south polar region, from which we produced new maps of topography, topographic roughness, and permanently shadowed regions (PSRs). The roughness maps reveal a diversity of terrains characterized by hectometer-scale roughness which is controlled in this region primarily by cratering and downslope mass transport. The south polar region is littered with linear roughness features of order ~1 – 10 km wide and ~10s – 100s of km long hypothesized to be secondary impacts within extended ejecta rays. Non uniformities in these features could reflect variations in secondary impactor properties and/or target terrain properties. Poleward of 80° S, the PSR cumulative size-frequency distribution (CSFD) shows an approximately power-law behavior whose exponent has a spatial variation of ~10%. PSRs with areas < 1 km2 contain 15 ± 5% of the total PSR area. Finally, we studied the effect of false positives and false negatives on the accuracy of the measured PSR cumulative size-frequency distribution (CSFD) and on the area for any individual PSR. The new maps presented here have many applications in the science and exploration of the lunar south polar region, such as geologic mapping and traverse planning.

Michael K. Barker↗

Human virus and microbial indicator occurrence in public-supply groundwater systems: meta-analysis of 12 international studies

Groundwater quality is often evaluated using microbial indicators. This study examines data from 12 international groundwater studies (conducted 1992–2013) of 718 public drinking-water systems located in a range of hydrogeological settings. Focus was on testing the value of indicator organisms for identifying virus-contaminated wells. One or more indicators and viruses were present in 37 and 15% of 2,273 samples and 44 and 27% of 746 wells, respectively. Escherichia coli ( E. coli ) and somatic coliphage are 7– 9 times more likely to be associated with culturable viruspositive samples when the indicator is present versus when it is absent, while F-specific and somatic coliphages are 8–9 times more likely to be associated with culturable viruspositive wells. However, single indicators are only marginally associated with viruses detected by molecular methods, and all microbial indicators have low sensitivity and positive predictive values for virus occurrence, whether by culturable or molecular assays, i.e., indicators are often absent when viruses are present and the indicators have a high false-positive rate. Wells were divided into three susceptibility subsets based on presence of (1) total coliform bacteria or (2) multiple indicators, or (3) location of wells in karst, fractured bedrock, or gravel/cobble settings. Better associations of some indicators with viruses were observed for (1) and (3). Findings indicate the best indicators are E. coli or somatic coliphage, although both indicators may underestimate virus occurrence. Repeat sampling for indicators improves evaluation of the potential for viral contamination in a well.

58 GEOSCIENCES↗

Statistical Requirements For Pass-Fail Testing Of Contraband Detection Systems

Contraband detection systems for homeland security applications are typically tested for probability of detection (PD) and probability of false alarm (PFA) using pass-fail testing protocols. Test protocols usually require specified values for PD and PFA to be demonstrated at a specified level of statistical confidence CL. Based on a recent more theoretical treatment of this subject [1], this summary reviews the definition of CL and provides formulas and spreadsheet functions for constructing tables of general test requirements and for determining the minimum number of tests required. The formulas and tables in this article may be generally applied to many other applications of pass-fail testing, in addition to testing of contraband detection systems.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

Digital Camera with Apparatus for Authentication of Images Produced from an Image File

A digital camera equipped with a processor for authentication of images produced from an image file taken by the digital camera is provided. The digital camera processor has embedded therein a private key unique to it, and the camera housing has a public key that is so uniquely related to the private key that digital data encrypted with the private key may be decrypted using the public key. The digital camera processor comprises means for calculating a hash of the image file using a predetermined algorithm, and second means for encrypting the image hash with the private key, thereby producing a digital signature. The image file and the digital signature are stored in suitable recording means so they will be available together. Apparatus for authenticating the image file as being free of any alteration uses the public key for decrypting the digital signature, thereby deriving a secure image hash identical to the image hash produced by the digital camera and used to produce the digital signature. The authenticating apparatus calculates from the image file an image hash using the same algorithm as before. By comparing this last image hash with the secure image hash, authenticity of the image file is determined if they match. Other techniques to address time-honored methods of deception, such as attaching false captions or inducing forced perspectives, are included.

Friedman, Gary L.↗

Elastic Changepoint Detection for Globally-indexed Functional Time Series Data with Climate Applications

Changepoint detection is a vital tool in the application of climate data analysis. Numerous types of climate observation data are most properly represented by functional time series, implying a need for accurate changepoint detection methods applicable to functional time series data. Such data taken at a global scale often contain both spatial heterogeneity and dependence as well as phase (time) misalignment. In this report, we present methods which can detect spatially-dependent changepoints while allowing different estimates of change time and change strength depending on location. Additionally, we provide extensions to this spatially-predicted model which controls for phase variability among observations. Our methods provide the ability to detect a single change, or control for epidemic changes (where a “return-to-normal” change is more likely to be detected than the initial change). We showcase results analyzing the June 1991 eruption of Mt. Pinatubo, where our methods demonstrate the ability to accurately detect both single and epidemic changepoints even in the presence of strong seasonal variability. We find that our spatially-predicted model improves the detection of relevant changepoints versus methods which do not take spatial information into account, and we find that controlling for phase variability helps to control the false discovery rate during the detection process.

54 ENVIRONMENTAL SCIENCES↗

Evaluation of Automatic Atlas-Based Lymph Node Segmentation for Head-and-Neck Cancer

Purpose: To evaluate if automatic atlas-based lymph node segmentation (LNS) improves efficiency and decreases inter-observer variability while maintaining accuracy. Methods and Materials: Five physicians with head-and-neck IMRT experience used computed tomography (CT) data from 5 patients to create bilateral neck clinical target volumes covering specified nodal levels. A second contour set was automatically generated using a commercially available atlas. Physicians modified the automatic contours to make them acceptable for treatment planning. To assess contour variability, the Simultaneous Truth and Performance Level Estimation (STAPLE) algorithm was used to take collections of contours and calculate a probabilistic estimate of the 'true' segmentation. Differences between the manual, automatic, and automatic-modified (AM) contours were analyzed using multiple metrics. Results: Compared with the 'true' segmentation created from manual contours, the automatic contours had a high degree of accuracy, with sensitivity, Dice similarity coefficient, and mean/max surface disagreement values comparable to the average manual contour (86%, 76%, 3.3/17.4 mm automatic vs. 73%, 79%, 2.8/17 mm manual). The AM group was more consistent than the manual group for multiple metrics, most notably reducing the range of contour volume (106-430 mL manual vs. 176-347 mL AM) and percent false positivity (1-37% manual vs. 1-7% AM). Average contouring time savings with the automatic segmentation was 11.5 min per patient, a 35% reduction. Conclusions: Using the STAPLE algorithm to generate 'true' contours from multiple physician contours, we demonstrated that, in comparison with manual segmentation, atlas-based automatic LNS for head-and-neck cancer is accurate, efficient, and reduces interobserver variability.

62 RADIOLOGY AND NUCLEAR MEDICINE↗