Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “logistic regression”

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 163 records · Page 9

Exploring the environmental drivers of human blastomycosis cases in the Midwestern United States

Blastomycosis is a fungal infection endemic to the eastern United States (US) and Canada caused by the inhalation of the fungi Blastomyces spp. Currently, the environmental drivers of disease dynamics are poorly understood. The goal of our work was to explore what environmental conditions are associated with the annual presence of blastomycosis cases, and therefore are potentially explanatory of the ecological niche of Blastomyces. We examined the relationships between reported cases of blastomycosis in three Midwestern US states (Michigan, Minnesota, and Wisconsin) from 2007–2017 in relation to eleven hypothesized environmental conditions, including climate, stream and soil mineral content, and land cover variables. Then, we fit logistic regression models to explore the relationships between the environmental variables and yearly blastomycosis case occurrence. Mean soil moisture, stream sediment mercury content, percent of water within the county, and woody wetlands land cover were all positively associated with the presence of annual cases, with woody wetlands having the most consistent signal across the three states. We also found significant differences in the likelihood of case presence between US states that were not explained by the variables in our model, suggesting state-level differences in case reporting and disease awareness. Our results provide a perspective on potential biological hypotheses to further test regarding environmental controls on the life cycle and ecological niche of Blastomyces.

54 ENVIRONMENTAL SCIENCES↗

Open Specy 1.0: Automated (Hyper)spectroscopy for Microplastics

Microplastic spectral analysis is one of the most time-consuming processes in studying microplastic pollution, often requiring days per sample. Researchers are transitioning to automated batch and hyperspectral image analysis techniques to enhance efficiency. Open Specy, initially aimed at manual single-spectrum analysis, has now integrated automated methods. This updated version, Open Specy 1.0, introduces several new features, including two algorithms for automated processing (smoothing and particle compression), an extensive library containing over 40,000 open-source Raman and FTIR spectra, and two machine learning classifiers (logistic regression and k medoids) developed from this library. Furthermore, it includes a revamped user interface, an R package, and a benchmark data set for testing future advancements in automated techniques. Researchers evaluated various configurations for hyperspectral smoothing, particle identification, compression, and splitting, to achieve combined recovery rates between 50 and 150% particle counts, identities, and sizes with a coefficient of variation (CV) of less than 40% (the accredited standard). Mean absorbance times the standard deviation provided a consistent particle identification. Hyperspectral smoothing led to a 96% combined recovery rate and reduced variability (CV = 38%) compared to the 86% recovery (CV = 83%) of nonsmoothed controls. Additionally, compressing spectra for particles was significantly faster (>3x) and showed similar accuracy but with reduced variability than processing each pixel individually. Key challenges persist in automating spectral analysis, particularly in refining particle splitting algorithms, and improving identification routines to minimize false positives and negatives. In conclusion, new methods in sample preparation for better stabilization and dispersion of particles could overcome some of these issues.

13 HYDRO ENERGY↗

Bone mineral density and blood metals in premenopausal women

Exposure to metals, specifically cadmium, lead, and mercury, is widespread and is associated with reduced bone mineral density (BMD) in older populations, but the associations among premenopausal women are unclear. Therefore, we evaluated the relationship between these metals in blood and BMD (whole body, total hip, lumbar spine, and non-dominant wrist) quantified by dual energy X-ray absorptiometry in 248 premenopausal women, aged 18-44. Participants were of normal body mass index (mean BMI 24.1), young (mean age 27.4), 60% were white, 20% non-Hispanic black, 15% Asian, and 6% other race group, and were from the Buffalo, New York region. The median (interquartile range) level of cadmium was 0.30 {mu}g/l (0.19-0.43), of lead was 0.86 {mu}g/dl (0.68-1.20), and of mercury was 1.10 {mu}g/l (0.58-2.00). BMD was treated both as a continuous variable in linear regression and dichotomized at the 10th percentile for logistic regression analyses. Mercury was associated with reduced odds of decreased lumbar spine BMD (0.66, 95% confidence interval: 0.44, 0.99), but overall, metals at environmentally relevant levels of exposure were not associated with reduced BMD in this population of healthy, reproductive-aged women. Further research is needed to determine if the blood levels of cadmium, lead, and mercury in this population are sufficiently low that there is no substantive impact on bone, or if effects on bone can be expected only at older ages.

60 APPLIED LIFE SCIENCES↗

Incidence and Treatment Patterns in Hospitalizations for Malignant Spinal Cord Compression in the United States, 1998-2006

Purpose: To characterize patterns in incidence, management, and costs of malignant spinal cord compression (MSCC) hospitalizations in the United States, using population-based data. Methods and Materials: Using the Nationwide Inpatient Sample, an all-payer healthcare database representative of all U.S. hospitalizations, MSCC-related hospitalizations were identified for the period 1998-2006. Cases were combined with age-adjusted Surveillance, Epidemiology and End Results cancer death data to estimate annual incidence. Linear regression characterized trends in patient, treatment, and hospital characteristics, costs, and outcomes. Logistic regression was used to examine inpatient treatment (radiotherapy [RT], surgery, or neither) by hospital characteristics and year, adjusting for confounding. Results: We identified 15,367 MSCC-related cases, representing 75,876 hospitalizations. Lung cancer (24.9%), prostate cancer (16.2%), and multiple myeloma (11.1%) were the most prevalent underlying cancer diagnoses. The annual incidence of MSCC hospitalization among patients dying of cancer was 3.4%; multiple myeloma (15.0%), Hodgkin and non-Hodgkin lymphomas (13.9%), and prostate cancer (5.5%) exhibited the highest cancer-specific incidence. Over the study period, inpatient RT for MSCC decreased (odds ratio [OR] 0.68, 95% confidence interval [CI] 0.61-0.81), whereas surgery increased (OR 1.48, 95% CI 1.17-1.84). Hospitalization costs for MSCC increased (5.3% per year, p < 0.001). Odds of inpatient RT were greater at teaching hospitals (OR 1.41, 95% CI 1.19-1.67), whereas odds of surgery were greater at urban institutions (OR 1.82, 95% CI 1.29-2.58). Conclusions: In the United States, patients dying of cancer have an estimated 3.4% annual incidence of MSCC requiring hospitalization. Inpatient management of MSCC varied over time and by hospital characteristics, with hospitalization costs increasing. Future studies are required to determine the impact of treatment patterns on MSCC outcomes and strategies for reducing MSCC-related costs.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Maternal nutritional status during pregnancy and surma use determine cord lead levels in Karachi, Pakistan

Objectives: To estimate the umbilical cord blood lead levels (BLLs) of Pakistani neonates and to identify determinants for umbilical BLLs. Methods: We conducted a cross-sectional study of mothers and infants at one of the two obstetric units of two tertiary care hospitals in Karachi during January-August 2005. Information from 540 mothers selected randomly from those registered for delivery was obtained about their pregnancy, diet, and current and past lead exposures. We collected umbilical cord blood for lead levels analyzed using graphite furnace atomic absorption spectrophotometry. We computed geometric and arithmetic means. We performed multiple linear regression analysis to identify factors associated with log-transformed umbilical cord BLLs. We also performed logistic regression analysis to identify determinants of high lead cord BLLs ({>=}10 {mu}g/dl). Results: The geometric mean cord BLL of the neonates was 9.6 {mu}g/dl; arithmetic mean (S.D.) was 10.8 {mu}g/dl (5.7) with a median of 9.7 {mu}g/dl and a range of 1.8-48.9 {mu}g/dl. Women who reported intake of less than 58.5 mg of elemental iron supplement per day during pregnancy had cord BLL of 10.0 {mu}g/dl; in comparison those women who had higher iron intake had lower cord BLL (8.4 {mu}g/dl). Those who used surma (an eye cosmetic) daily had higher cord BLL (11.5 {mu}g/dl) as compared to those who used it less frequently (9.4 {mu}g/dl). In multivariable linear regression model, higher iron intake, owning a car, and being in 2nd quartile of mid-arm circumference were associated with low lead levels while father's occupation in lead-based industry was associated with significantly higher umbilical cord BLLs. There was interaction of daily surma use and ethnicity. Geometric mean BLLs were varied among surma users by ethnicity. Conclusions: Umbilical cord BLLs are high in Karachi, Pakistan, in comparison to those in developed countries such as United States. Measures are needed to reduce fetal lead exposure to prevent adverse affect on neurocognitive development. Association of low iron (below RDA of 60 mg per day) with high umbilical cord has implications for strengthening iron supplement intake during pregnancy. Umbilical cord BLLs differed among surma users by ethnicity.

60 APPLIED LIFE SCIENCES↗

MEASUREMENT OF WIND SPEED FROM COOLING LAKE THERMAL IMAGERY

The Savannah River National Laboratory (SRNL) collected thermal imagery and ground truth data at two commercial power plant cooling lakes to investigate the applicability of laboratory empirical correlations between surface heat flux and wind speed, and statistics derived from thermal imagery. SRNL demonstrated in a previous paper [1] that a linear relationship exists between the standard deviation of image temperature and surface heat flux. In this paper, SRNL will show that the skewness of the temperature distribution derived from cooling lake thermal images correlates with instantaneous wind speed measured at the same location. SRNL collected thermal imagery, surface meteorology and water temperatures from helicopters and boats at the Comanche Peak and H. B. Robinson nuclear power plant cooling lakes. SRNL found that decreasing skewness correlated with increasing wind speed, as was the case for the laboratory experiments. Simple linear and orthogonal regression models both explained about 50% of the variance in the skewness - wind speed plots. A nonlinear (logistic) regression model produced a better fit to the data, apparently because the thermal convection and resulting skewness are related to wind speed in a highly nonlinear way in nearly calm and in windy conditions.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Evaluation of diffusion kurtosis and diffusivity from baseline staging MRI as predictive biomarkers for response to neoadjuvant chemoradiation in locally advanced rectal cancer

Purpose: To evaluate the role of diffusion kurtosis and diffusivity as potential imaging biomarkers to predict response to neoadjuvant chemoradiation therapy (CRT) from baseline staging magnetic resonance imaging (MRI) in locally advanced rectal cancer (LARC). Materials and methods: This retrospective study included 45 consecutive patients (31 male/14 female) who underwent baseline MRI with high b-value sequences (up to 1500 mm/s{sup 2}) for LARC followed by neoadjuvant chemoradiation and surgical resection. The mean age was 57.4 years (range 34.2–72.9). An abdominal radiologist using open source software manually segmented T2-weighted images. Segmentations were used to derive diffusion kurtosis and diffusivity from diffusion-weighted images as well as volumetric data. These data were analyzed with regard to tumor regression grade (TRG) using the four-tier American Joint Committee on Cancer (AJCC) classification, TRG 0–3. Proportional odds regression was used to analyze the four-level ordinal outcome. A sensitivity analysis was performed using univariable logistic regression for binary TRG groups, TRG 0/1 (> 90% response), or TRG 2/3 (< 90% response). p < 0.05 was considered significant throughout. Results: In the univariable proportional odds regression analysis, higher diffusivity summary (D{sub sum}) values were observed to be significantly associated with higher odds of being in one or more favorable TRG group (TRG 0 or 1). In other words, on average, patients with higher D{sub sum} values were more likely to be in a more favorable TRG group. These results are mostly consistent with the sensitivity analysis, in which higher values for most D{sub sum} values [all but region of interest (ROI)-max D median (p = 0.08)] were observed to be significantly associated with higher odds of being TRG 0 or 1. Tumor volume of interest (VOI) and ROI volume, ROI kurtosis mean and median, and VOI kurtosis mean and median were not significantly associated with TRG. Conclusion: Diffusivity derived from the baseline staging MRI, but not diffusion kurtosis or volumetric data, is associated with TRG and therefore shows promise as a potential imaging biomarker to predict the response to neoadjuvant chemotherapy in LARC. Clinical relevance statement: Diffusivity shows promise as a potential imaging biomarker to predict AJCC TRG following neoadjuvant CRT, which has implications for risk stratification. Patients with TRG 0/1 have 5-year disease-free survival (DFS) of 90–98%, as opposed to those who are TRG 2/3 with 5-year DFS of 68–73%.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Pretreatment Apparent Diffusion Coefficient of the Primary Lesion Correlates With Local Failure in Head-and-Neck Cancer Treated With Chemoradiotherapy or Radiotherapy

Purpose: This study was performed to evaluate whether the apparent diffusion coefficient (ADC) of a primary lesion correlates with local failure in primary head-and-neck squamous cell carcinoma (HNSCC) treated with chemoradiotherapy or radiotherapy. Methods and Materials: We retrospectively studied 38 patients with primary HNSCC (12 oropharynx, 20 hypopharynx, 4 larynx, 2 oral cavity) treated with chemoradiotherapy or radiotherapy with radiation dose to gross tumor volume equal to or over 60 Gy and who underwent pretreatment magnetic resonance imaging, including diffusion-weighted imaging. Ten patients developed local failure during follow-up periods of 2.0 to 9.3 months, and the remaining 28 showed local control during follow-up periods of 10.5 to 31.7 months. The variables that could affect local failure (age, tumor volume, ADC, T stage, N stage, dose, treatment method, tumor location, and overall treatment time) were analyzed using logistic regression analyses for all 38 patients and for 17 patients with Stage T3 or T4 disease. Results: In univariate logistic analysis for all 38 cases, tumor volume, ADC, T stage, and treatment method showed significant (p < 0.05) associations with local failure. In multivariate analysis, ADC and T stage revealed significance (p < 0.01). In univariate logistic analysis for the 17 patients with Stage T3 or T4 disease, ADC and dose showed significant (p < 0.01) associations with local failure. In multivariate analysis, ADC alone showed significance (p < 0.05). Conclusions: The results suggest that pretreatment ADC, along with T stage, is a potential indicator of local failure in HNSCC treated with chemoradiotherapy or radiotherapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Association Between Prevalence of Peripheral Artery Disease and Radiation Exposure in the Atomic Bomb Survivors

Past reports suggested that total-body irradiation at 0.5 to 1.0 Gy could be responsible for atherosclerosis. Peripheral artery disease (PAD) is a manifestation of systematic atherosclerosis. Whether the consequences of a low-to-moderate dose of radiation include increased risk of PAD remains to be determined. The purpose of this study was to examine the association between radiation exposure and prevalence of PAD among Japanese atomic bomb survivors. Radiation exposure from the atomic bombing was assessed in 3476 participants (41.1% men, mean age 74.8 years with SD 6.4 years) with a cross-sectional survey in 2010 to 2014. Left- and right-side ankle-brachial indexes and upstroke time (UT) were obtained using oscillometric VP-2000. PAD was defined as an ankle-brachial index of 1.0 or less or a prior history related to revascularization. UT was considered a sensitive marker of early-stage PAD. Association between radiation exposure and PAD or UT was assessed using multivariable regression analyses with adjustment for potential confounding factors. Of 3476 participants, 79 (2.3%) were identified as having prevalent PAD. Multivariate logistic regression analysis indicated that radiation dose was unrelated to PAD prevalence (odds ratio, 0.83; 95% confidence interval [0.57-1.21]). UT appeared to increase with radiation dose, but the increase was not statistically significant (1.09 ms/Gy; 95% confidence interval [-0.17 to 2.36]). We found no clear association of radiation dose with PAD, but it remains to be determined whether UT is associated with radiation dose.

60 APPLIED LIFE SCIENCES↗

Functional Data Analysis in NTCP Modeling: A New Method to Explore the Radiation Dose-Volume Effects

Purpose/Objective(s): To describe a novel method to explore radiation dose-volume effects. Functional data analysis is used to investigate the information contained in differential dose-volume histograms. The method is applied to the normal tissue complication probability modeling of rectal bleeding (RB) for patients irradiated in the prostatic bed by 3-dimensional conformal radiation therapy. Methods and Materials: Kernel density estimation was used to estimate the individual probability density functions from each of the 141 rectum differential dose-volume histograms. Functional principal component analysis was performed on the estimated probability density functions to explore the variation modes in the dose distribution. The functional principal components were then tested for association with RB using logistic regression adapted to functional covariates (FLR). For comparison, 3 other normal tissue complication probability models were considered: the Lyman-Kutcher-Burman model, logistic model based on standard dosimetric parameters (LM), and logistic model based on multivariate principal component analysis (PCA). Results: The incidence rate of grade ≥2 RB was 14%. V{sub 65Gy} was the most predictive factor for the LM (P=.058). The best fit for the Lyman-Kutcher-Burman model was obtained with n=0.12, m = 0.17, and TD50 = 72.6 Gy. In PCA and FLR, the components that describe the interdependence between the relative volumes exposed at intermediate and high doses were the most correlated to the complication. The FLR parameter function leads to a better understanding of the volume effect by including the treatment specificity in the delivered mechanistic information. For RB grade ≥2, patients with advanced age are significantly at risk (odds ratio, 1.123; 95% confidence interval, 1.03-1.22), and the fits of the LM, PCA, and functional principal component analysis models are significantly improved by including this clinical factor. Conclusion: Functional data analysis provides an attractive method for flexibly estimating the dose-volume effect for normal tissues in external radiation therapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Associations between cadmium levels in blood and urine, blood pressure and hypertension among Canadian adults

Background: Cadmium has been inconsistently related to blood pressure and hypertension. The present study seeks to clarify the relationship between cadmium levels found in blood and urine, blood pressure and hypertension in a large sample of adults. Methods: The study sample included participants ages 20 through 79 from multiple cycles of the Canadian Health Measures Survey (2007 through 2013) with measured blood cadmium (n=10,099) and urinary cadmium (n=6988). Linear regression models examined the association between natural logarithm transformed cadmium levels and blood pressure (separate models for systolic and diastolic blood pressure) after controlling for known covariates. Logistic regression models were used to examine the association between cadmium and hypertension. Models were run separately by sex, smoking status, and body mass index category. Results: Men had higher mean systolic (114.8 vs. 110.8 mmHg, p<0.01) and diastolic (74.0 vs. 69.6 mmHg, p<0.01) blood pressure compared to women. Although, geometric mean blood (0.46 vs. 0.38 µg/L, p<0.01) and creatinine-adjusted standardized urinary cadmium levels (0.48 vs. 0.38 µg/L, p<0.01) were higher among those with hypertension, these differences were no longer significant after adjustment for age, sex and smoking status. In overall regression models, increases in blood cadmium were associated with increased systolic (0.70 mmHg, 95% confidence interval [CI]=0.25–1.16, p<0.01) and diastolic blood pressure (0.74 mmHg, 95% CI=0.30–1.19, p<0.01). The associations between urinary cadmium, blood pressure and hypertension were not significant in overall models. Model stratification revealed significant and negative associations between urinary cadmium and hypertension among current smokers (OR=0.61, 95% CI=0.44–0.85, p<0.01), particularly female current smokers (OR=0.52, 95% CI=0.32–0.85, p=0.01). Conclusion: This study provides evidence of a significant association between cadmium levels, blood pressure and hypertension. However, the significance and direction of this association differs by sex, smoking status, and body mass index category. - Highlights: • Blood and urinary cadmium levels higher among those with hypertension. • Evidence of association between cadmium levels, blood pressure and hypertension. • Significance and direction of association differs by sex, smoking status, and BMI. • Higher urinary cadmium levels lower hypertension risk for current (female) smokers.

54 ENVIRONMENTAL SCIENCES↗

IoT Intrusion Detection Taxonomy, Reference Architecture, and Analyses

This paper surveys the deep learning (DL) approaches for intrusion-detection systems (IDSs) in Internet of Things (IoT) and the associated datasets toward identifying gaps, weaknesses, and a neutral reference architecture. A comparative study of IDSs is provided, with a review of anomaly-based IDSs on DL approaches, which include supervised, unsupervised, and hybrid methods. All techniques in these three categories have essentially been used in IoT environments. To date, only a few have been used in the anomaly-based IDS for IoT. For each of these anomaly-based IDSs, the implementation of the four categories of feature(s) extraction, classification, prediction, and regression were evaluated. We studied important performance metrics and benchmark detection rates, including the requisite efficiency of the various methods. Four machine learning algorithms were evaluated for classification purposes: Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), and an Artificial Neural Network (ANN). Therefore, we compared each via the Receiver Operating Characteristic (ROC) curve. The study model exhibits promising outcomes for all classes of attacks. The scope of our analysis examines attacks targeting the IoT ecosystem using empirically based, simulation-generated datasets (namely the Bot-IoT and the IoTID20 datasets).

97 MATHEMATICS AND COMPUTING↗

Does the air pollution model influence the evidence of socio-economic disparities in exposure and susceptibility?

Highlights: • Daily exposure to ozone may vary across air pollution models, considerably. • The exposure model influenced disparity of exposure across socio-economic strata. • The air pollution model does not affect disparity of susceptibility. • Multiple air pollution models should be used to assess socio-economic disparities. Studies assessing socio-economic disparities in air pollution exposure and susceptibility are usually based on a single air pollution model. A time stratified case-crossover study was designed to assess the impact of the type of model on differential exposure and on the differential susceptibility in the relationship between ozone exposure and daily mortality by socio-economic strata (SES) in Montreal. Non-accidental deaths along with deaths from cardiovascular and respiratory causes on the island of Montreal for the period 1991–2002 were included as cases. Daily ozone concentration estimates at partictaipants’ residence were obtained from the five following air pollution models: Average value (AV), Nearest station model (NS), Inverse-distance weighting interpolation (IDW), Land-use regression model with back-extrapolation (LUR-BE) and Bayesian maximum entropy model combined with a land-use regression (BME-LUR). The prevalence of a low household income (< 20,000/year) was used as socio-economic variable, divided into two categories as a proxy for deprivation. Multivariable conditional logistic regressions were used considering 3-day average concentrations. Multiplicative and additive interactions (using Relative Excess Risk due to Interaction) as well as Cochran's tests were calculated and results were compared across the different air pollution models. Heterogeneity of susceptibility and exposure according to socio-economic status (SES) were found. Ratio of exposure across SES groups means ranged from 0.75 [0.74–0.76] to 1.01 [1.00–1.02], respectively for the LUR-BE and the BME-LUR models. Ratio of mortality odds ratios ranged from 1.01 [0.96–1.05] to 1.02 [0.97–1.08], respectively for the IDW and LUR-BE models. Cochran's test of heterogeneity between the air pollution models showed important heterogeneity regarding the differential exposure by SES, but the air pollution model was not found to influence heterogeneity regarding the differential susceptibility. The study showed air pollution models can influence the assessment of disparities in exposure according to SES in Montreal but not that of disparities in susceptibility.

54 ENVIRONMENTAL SCIENCES↗

Overhead longwave infrared hyperspectral material identification using radiometric models

Material detection algorithms used in hyperspectral data processing are computationally efficient but can produce relatively high numbers of false positives. Material identification performed as a secondary processing step on detected pixels can help separate true and false positives. This paper presents a material identification processing chain for longwave infrared hyperspectral data of solid materials collected from airborne platforms. The algorithms utilize unwhitened radiance data and an iterative algorithm that determines the temperature, humidity, and ozone of the atmospheric profile. Pixel unmixing is done using constrained linear regression and Bayesian Information Criteria for model selection. The resulting product includes an optimal atmospheric profile and full radiance material model that includes material temperature, abundance values, and several fit statistics. A logistic regression method utilizing all model parameters to improve identification is also presented. This paper details the processing chain and provides justification for the algorithms used. Several examples are provided using modeled data at different noise levels.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Elastic functional principal component regression

We study regression using functional predictors in situations where these functions contains both phase and amplitude variability. In other words, the functions are misaligned due to errors in time measurements, and these errors can significantly degrade both model estimation and prediction performance. The current techniques either ignore the phase variability, or handle it via preprocessing, that is, use an off–the–shelf technique for functional alignment and phase removal. We develop a functional principal component regression model which has a comprehensive approach in handling phase and amplitude variability. The model utilizes a mathematical representation of the data known as the square–root slope function. These functions preserve the L 2 norm under warping and are ideally suited for simultaneous estimation of regression and warping parameters. Furthermore, using both simulated and real–world data sets, we demonstrate our approach and evaluate its prediction performance relative to current models. In addition, we propose an extension to functional logistic and multinomial logistic regression.

97 MATHEMATICS AND COMPUTING↗

Overhead longwave infrared hyperspectral material identification using radiometric models

Material detection algorithms used in hyperspectral data processing are computationally efficient but can produce relatively high numbers of false positives. Material identification performed as a secondary processing step on detected pixels can help mitigate false positives. A material identification processing chain for longwave infrared hyperspectral data of solid materials collected from airborne platforms is presented. The algorithms utilize unwhitened radiance data and Nelder–Meade numerical optimization to estimate the temperature, humidity, and ozone levels of the atmospheric profile. Pixel unmixing is done using constrained linear regression and Bayesian information criteria for model selection. The resulting identification product includes an optimal atmospheric profile and a full radiance material model that includes material temperature, abundance values, and several fit statistics. A logistic regression method utilizing the model parameters to improve identification is also presented. Furthermore, several examples are provided using modeled data at several noise levels.

58 GEOSCIENCES↗

Pelvic Lymph Node Status Assessed by 18F-Fluorodeoxyglucose Positron Emission Tomography Predicts Low-Risk Group for Distant Recurrence in Locally Advanced Cervical Cancer: A Prospective Study

Purpose: To develop a prediction model to identify a low-risk group for distant recurrence in patients with locally advanced cervical cancer treated by concurrent chemoradiation. Methods and Materials: Prospectively, 62 patients with locally advanced cervical cancer were recruited as a training cohort. Clinical variables and parameters obtained from positron emission tomography (PET) and magnetic resonance imaging were analyzed by logistic regression. For the test set, 54 patients were recruited independently. To identify the low-risk group, negative likelihood ratio (LR) less than 0.2 was set to be a cutoff. Results: Among the training cohort, multivariate logistic analysis revealed that advanced International Federation of Gynecology and Obstetrics (FIGO) stage and a high serum squamous cancer cell (SCC) antigen level were significant risk factors (p = 0.015 and 0.025, respectively). Using the two parameters, criteria to determine a low-risk subset for distant recurrence were postulated: (1) FIGO Stage IIB or less and (2) pretreatment SCC < 2.4 (Model A). Positive pelvic node on PET completely predicted all cases with distant recurrence and thus was considered as another prediction model (Model B). In the test cohort, although Model A did not showed diagnostic performance, Model B completely predicted all cases with distant recurrence and showed a sensitivity of 100% with negative LR of 0. Across the training and test cohort (n = 116), the false negative rate was 0 (95% confidence interval 0%-7.6%). Conclusions: Positive pelvic node on PET is a useful marker in prediction of distant recurrence in patients with locally advanced cervical cancer who are treated with concurrent chemoradiation.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Modeling and managing risk early in software development

In order to improve the quality of the software development process, we need to be able to build empirical multivariate models based on data collectable early in the software process. These models need to be both useful for prediction and easy to interpret, so that remedial actions may be taken in order to control and optimize the development process. We present an automated modeling technique which can be used as an alternative to regression techniques. We show how it can be used to facilitate the identification and aid the interpretation of the significant trends which characterize 'high risk' components in several Ada systems. Finally, we evaluate the effectiveness of our technique based on a comparison with logistic regression based models.

Briand, Lionel C.↗