Systems Modeling of Lung Cancer Screening Programs to Improve Quality

Thesis (Master's)--University of Washington, 2016-08

A Binary Classifier for Test Case Feasibility Applied to Automatically Generated Tests of Event-Driven Software

Modern software application testing, such as the testing of software driven by graphical user interfaces (GUIs) or leveraging event-driven architectures in general, requires paying careful attention to context. Model-based testing (MBT) approaches first acquire a model of an application, then use the model to construct test cases covering relevant contexts. A major shortcoming of state-of-the-art automated model-based testing is that many test cases proposed by the model are not actually executable. These \textit{infeasible} test cases threaten the integrity of the entire model-based suite, and any coverage of contexts the suite aims to provide. In this research, I develop and evaluate a novel approach for classifying the feasibility of test cases. I identify a set of pertinent features for the classifier, and develop novel methods for extracting these features from the outputs of MBT tools. I use a supervised logistic regression approach to obtain a model of test case feasibility from a randomly selected training suite of test cases. I evaluate this approach with a set of experiments. The outcomes of this investigation are as follows: I confirm that infeasibility is prevalent in MBT, even for test suites designed to cover a relatively small number of unique contexts. I confirm that the frequency of infeasibility varies widely across applications. I develop and train a binary classifier for feasibility with average overall error, false positive, and false negative rates under 5\%. I find that unique event IDs are key features of the feasibility classifier, while model-specific event types are not. I construct three types of features from the event IDs associated with test cases, and evaluate the relative effectiveness of each within the classifier. To support this study, I also develop a number of tools and infrastructure components for scalable execution of automated jobs, which use state-of-the-art container and continuous integration technologies to enable parallel test execution and the persistence of all experimental artifacts.

A more reliable PCR for detection of Mycobacterium tuberculosis in clinical samples

Diagnostic techniques based on PCR have two major problems: false-positive reactions due to contamination with DNA fragments from previous PCRs (amplicons) and false-negative reactions caused by inhibitors that interfere with the PCR. We have improved our previously reported PCR based on the amplification of a fragment of the Mycobacterium tuberculosis complex-specific insertion element IS6110 with respect to both problems. False-positive reactions caused by amplicon contamination were prevented by the use of uracil-N-glycosylase and dUTP instead of dTTP. We selected a new set of primers outside the region spanned by the formerly used primers to avoid false-positive reactions caused by dTTP-containing amplicons still present in the laboratory. With this new primer set, 16 copies of the IS6110 insertion element, the equivalent of two bacteria, could be amplified 10(10) times in 40 cycles, resulting in a mean efficiency of 77% per cycle. To detect the presence of inhibitors of the Taq polymerase, which may cause false-negative reactions, part of each sample was spiked with M. tuberculosis DNA. The DNA purification method using guanidinium thiocyanate and diatoms effectively removed most or all inhibitors of the PCR. However, this was not suitable for blood samples, for which we developed a proteinase K treatment followed by phenol-chloroform extraction. This method permitted detection of 20 M. tuberculosis bacteria per ml of whole blood. Various laboratory procedures were introduced to reduce failure or inhibition of PCR and avoid DNA cross contamination. We have tested 218 different clinical specimens obtained from patients suspected of having tuberculosis. The samples included sputum (n=145), tissue biopsy samples (n=25), cerebrospinal fluid (n=15), blood (n=14), pleural fluid (n=9), feces, (n=7), fluid from fistulae (n=2), and pus from a wound (n=1). The results obtained by PCR were consistent with those obtained with culture, which is the "gold standard." We demonstrate that PCR is a useful technique for the rapid diagnosis of tuberculosis at various sites.

Integrating Innate and Adaptive Immunity for Intrusion Detection

Network Intrusion Detection Systems (NIDS) monitor a net- work with the aim of discerning malicious from benign activity on that network. While a wide range of approaches have met varying levels of success, most IDS’s rely on having access to a database of known attack signatures which are written by security experts. Nowadays, in order to solve problems with false positive alerts, correlation algorithms are used to add additional structure to sequences of IDS alerts. However, such techniques are of no help in discovering novel attacks or variations of known attacks, something the human immune system (HIS) is capable of doing in its own specialised domain. This paper presents a novel immune algorithm for application to an intrusion detection problem. The goal is to discover packets containing novel variations of attacks covered by an existing signature base.

An Immune Network Intrusion Detection System Utilising Correlation Context

Network Intrusion Detection Systems (NIDS) are computer systems which monitor a network with the aim of discerning malicious from benign activity on that network. While a wide range of approaches have met varying levels of success, most IDSs rely on having access to a database of known attack signatures which are written by security experts. Nowadays, in order to solve problems with false positive alerts, correlation algorithms are used to add additional structure to sequences of IDS alerts. However, such techniques are of no help in discovering novel attacks or variations of known attacks, something the human immune system (HIS) is capable of doing in its own specialised domain. This paper presents a novel immune algorithm for application to the IDS problem. The goal is to discover packets containing novel variations of attacks covered by an existing signature base.

Anticorpos anti-trypanosoma cruzi como preditivos de infecção por leishmania infantum

Leishmania infantum and Trypanosoma cruzi are trypanosomatids of medical importance and are, respectively, the etiologic agents of visceral leishmaniasis (VL) and Chagas disease (CD) in Brazil. People infected with L. infantum or T. cruzi may develop asymptomatically, enabling the transmission of pathogens through blood transfusion and / or organs. The assessment of the infection by T. cruzi is included among the tests performed for screening blood donors in Brazil, however, there is no availability of tests for Leishmania. Serological tests for T. cruzi are very sensitive, but not specific, and may have cross-reactions with other microorganisms. Thus, the aim of this study was to determine the prevalence of Leishmania infection in blood donors and assess whether the serological test for T. cruzi detect L. infantum. Among the 300 blood samples from donors, discarded in 2011, 61 were T. cruzi positive, 203 were from donors with other infections and 36 were from handbags with low blood volume, but without infection. We also assessed 144 samples from donors without infections and able to donate blood, totaling 444 subjects. DNA was extracted from blood samples of all to perform quantitative PCR (qPCR) to detect Leishmania DNA. The buffy coat obtained from all samples was grown in Schneider medium supplemented and NNN. All samples were evaluated for the presence of anti-Leishmania antibody. The serological results indicate a percentage of 22% of Leishmania infection in blood samples obtained from discarded bags. A total of 60% of samples positive in ELISA for T. cruzi were negative by IFI, used as confirmatory test, ie 60% false positive for Chagas. Among these samples false positive for Chagas, 72% were positive by ELISA for Leishmania characterizing the occurrence of cross reaction between serologic assays. Of the 300 cultures performed, 18 grew parasites that were typed by qPCR and specific isoenzymes, found the species Leishmania infantum crops. Among the 18 cultures, 4 were purged from scholarships for low volume and all negative serology blood bank, thus demonstrating that there is a real risk of Leishmania transmission via transfusion. It is concluded that in an area endemic for leishmaniasis in Brazil, serological diagnosis performed to detect infection by T. cruzi among blood donors can identify infection by L. infantum and although cause false positive for Chagas, this cross-reactivity reduces the risk of Leishmania infection via blood transfusion, since tests are not applied specific detection of the parasite. In this way, there remains the need to discuss the implementation of a specific serological screening test for Leishmania in endemic countries such as Brazil

An immune inspired Network Intrusion Detection System utilising correlation

Network Intrusion Detection Systems (NIDS) are computer systems which monitor a network with the aim of discerning malicious from benign activity on that network. While a wide range of approaches have met varying levels of success, most IDSs rely on having access to a database of known attack signatures which are written by security experts. Nowadays, in order to solve problems with false positive alerts, correlation algorithms are used to add additional structure to sequences of IDS alerts. However, such techniques are of no help in discovering novel attacks or variations of known attacks, something the human immune system (HIS) is capable of doing in its own specialised domain. This paper presents a novel immune algorithm for application to the IDS problem. The goal is to discover packets containing novel variations of attacks covered by an existing signature base.

Risk Factors for Neonatal Sepsis and Method for Reduction of Blood Culture Contamination

Background False-positive blood cultures findings may lead to a falsely increased morbidity and increased hospital costs. Method The survey was conducted as retrospective - prospective study and included 239 preterm infants (born before 37 weeks of gestation) who were treated in Neonatal Intensive Care Unit (NICU) in Institute for Child and Youth Health Care of Vojvodina during one year (January 1st, 2012 to December 31st, 2012). The retrospective part of the study focused on examination of incidence of neonatal sepsis and determination of risk factors. In the prospective part of the study infants were sub-divided into two groups: Group 1- infants hospitalized in NICU during the first 6 months of the study; blood cultures were taken by the ‘’clean technique’’ and checklists for this procedure were not taken. Group 2- neonates hospitalized in NICU during last 6 months of the study; blood cultures were taken by ‘’sterile technique’’ and checklists for this procedure were taken. Results The main risk factors for sepsis were prelabor rupture of membranes, low gestational age, low birth weight, mechanical ventilation, umbilical venous catheter placement, and abdominal drainage. Staphylococcus aureus and coagulase negative Staphylococcus were the most frequently isolated microorganisms in false-positive blood samples. Conclusions Education of employees, use of checklists and sterile sets for blood sampling, permanent control of false positive blood cultures, as well as regular and routine monthly reports are crucial for successful reduction of contamination rates.

Object detection and recognition in complex scenes

Dissertação de Mestrado, Engenharia Informática, Faculdade de Ciências e Tecnologia, Universidade do Algarve, 2014

A COMPARISON OF PIXEL-BASED VERSUS OBJECT ORIENTED ANALYSIS OF LANDSLIDES USING HISTORICAL REMOTE SENSING DATA

With recent advances in remote sensing processing technology, it has become more feasible to begin analysis of the enormous historic archive of remotely sensed data. This historical data provides valuable information on a wide variety of topics which can influence the lives of millions of people if processed correctly and in a timely manner. One such field of benefit is that of landslide mapping and inventory. This data provides a historical reference to those who live near high risk areas so future disasters may be avoided. In order to properly map landslides remotely, an optimum method must first be determined. Historically, mapping has been attempted using pixel based methods such as unsupervised and supervised classification. These methods are limited by their ability to only characterize an image spectrally based on single pixel values. This creates a result prone to false positives and often without meaningful objects created. Recently, several reliable methods of Object Oriented Analysis (OOA) have been developed which utilize a full range of spectral, spatial, textural, and contextual parameters to delineate regions of interest. A comparison of these two methods on a historical dataset of the landslide affected city of San Juan La Laguna, Guatemala has proven the benefits of OOA methods over those of unsupervised classification. Overall accuracies of 96.5% and 94.3% and F-score of 84.3% and 77.9% were achieved for OOA and unsupervised classification methods respectively. The greater difference in F-score is a result of the low precision values of unsupervised classification caused by poor false positive removal, the greatest shortcoming of this method.

African swine fever virus transmission cycles in Central Europe: evaluation of wild boar-soft tick contacts through detection of antibodies against Ornithodoros erraticus saliva antigen.

BACKGROUND African swine fever (ASF) is one of the most complex viral diseases affecting both domestic and wild pigs. It is caused by ASF virus (ASFV), the only DNA virus which can be efficiently transmitted by an arthropod vector, soft ticks of the genus Ornithodoros. These ticks can be part of ASFV-transmission cycles, and in Europe, O. erraticus was shown to be responsible for long-term maintenance of ASFV in Spain and Portugal. In 2014, the disease has been reintroduced into the European Union, affecting domestic pigs and, importantly, also the Eurasian wild boar population. In a first attempt to assess the risk of a tick-wild boar transmission cycle in Central Europe that would further complicate eradication of the disease, over 700 pre-existing serum samples from wild boar hunted in four representative German Federal States were investigated for the presence of antibodies directed against salivary antigen of Ornithodoros erraticus ticks using an indirect ELISA format. RESULTS Out of these samples, 16 reacted with moderate to high optical densities that could be indicative of tick bites in sampled wild boar. However, these samples did not show a spatial clustering (they were collected from distant geographical regions) and were of bad quality (hemolysis/impurities). Furthermore, all positive samples came from areas with suboptimal climate for soft ticks. For this reason, false positive reactions are likely. CONCLUSION In conclusion, the study did not provide stringent evidence for soft tick-wild boar contact in the investigated German Federal States and thus, a relevant involvement in the epidemiology of ASF in German wild boar is unlikely. This fact would facilitate the eradication of ASF in the area, although other complex relations (wild boar biology and interactions with domestic pigs) need to be considered.

Genetic diversity of Mycobacterium avium isolates recovered from clinical samples and from the environment: molecular characterization for diagnostic purposes

Isolation of Mycobacterium avium complex (MAC) organisms from clinical samples may occur in patients without clinical disease, making the interpretation of results difficult. The clinical relevance of MAC isolates from different types of clinical samples (n = 47) from 39 patients in different sections of a hospital was assessed by comparison with environmental isolates (n = 17) from the hospital. Various methods for identification and typing (commercial probes, phenotypic characteristics, PCR for detection of IS1245 and IS901, sequencing of the hsp65 gene, and pulsed-field gel electrophoresis) were evaluated. The same strain was found in all the environmental isolates, 21 out of 23 (91.3%) of the isolates cultured from urine samples, and 5 out of 19 (26.3%) isolates from respiratory specimens. This strain did not cause disease in the patients. Testing best characterized the strain as M. avium subsp. hominissuis, with the unusual feature that 81.4% of these isolates lacked the IS1245 element. Contamination of certain clinical samples with an environmental strain was the most likely event; therefore, characterization of the environmental mycobacteria present in health care facilities should be performed to discard false-positive isolations in nonsterile samples, mainly urine samples. Molecular techniques applied in this study demonstrated their usefulness for this purpose.

Treatment trials for neonatal seizures: the effect of design on sample size

Neonatal seizures are common in the neonatal intensive care unit. Clinicians treat these seizures with several anti-epileptic drugs (AEDs) to reduce seizures in a neonate. Current AEDs exhibit sub-optimal efficacy and several randomized control trials (RCT) of novel AEDs are planned. The aim of this study was to measure the influence of trial design on the required sample size of a RCT. We used seizure time courses from 41 term neonates with hypoxic ischaemic encephalopathy to build seizure treatment trial simulations. We used five outcome measures, three AED protocols, eight treatment delays from seizure onset (Td) and four levels of trial AED efficacy to simulate different RCTs. We performed power calculations for each RCT design and analysed the resultant sample size. We also assessed the rate of false positives, or placebo effect, in typical uncontrolled studies. We found that the false positive rate ranged from 5 to 85% of patients depending on RCT design. For controlled trials, the choice of outcome measure had the largest effect on sample size with median differences of 30.7 fold (IQR: 13.7–40.0) across a range of AED protocols, Td and trial AED efficacy (p<0.001). RCTs that compared the trial AED with positive controls required sample sizes with a median fold increase of 3.2 (IQR: 1.9–11.9; p<0.001). Delays in AED administration from seizure onset also increased the required sample size 2.1 fold (IQR: 1.7–2.9; p<0.001). Subgroup analysis showed that RCTs in neonates treated with hypothermia required a median fold increase in sample size of 2.6 (IQR: 2.4–3.0) compared to trials in normothermic neonates (p<0.001). These results show that RCT design has a profound influence on the required sample size. Trials that use a control group, appropriate outcome measure, and control for differences in Td between groups in analysis will be valid and minimise sample size.

Evaluación del desempeño de pruebas diagnósticas de tuberculosis en pacientes VIH positivo en dos hospitales públicos de Bogotá

La tuberculosis TB es una de las principales causas de muerte en el mundo en individuos con infección por VIH. En Colombia esta coinfección soporta una carga importante en la población general convirtiéndose en un problema de salud pública. En estos pacientes las pruebas diagnósticas tienen sensibilidad inferior y la enfermedad evoluciona con mayor frecuencia hacia formas diseminadas y rápidamente progresivas y su diagnóstico oportuno representa un reto en Salud. El objetivo de este proyecto es evaluar el desempeño de las pruebas diagnósticas convencionales y moleculares, para la detección de TB latente y activa pacientes con VIH, en dos hospitales públicos de Bogotá. Para TB latente se evaluó la concordancia entre las pruebas QuantiFERON-TB (QTF) y Tuberculina (PPD), sugiriendo superioridad del QTF sobre la PPD. Se evaluaron tres pruebas diagnósticas por su sensibilidad y especificidad, baciloscopia (BK), GenoType®MTBDR plus (Genotype) y PCR IS6110 teniendo como estándar de oro el cultivo. Los resultados de sensibilidad (S) y especificidad (E) de cada prueba con una prevalencia del 19,4 % de TB pulmonar y extrapulmonar en los pacientes que participaron del estudio fue: BK S: 64% E: 99,1%; Genotype S: 77,8% E: 94,5%; PCRIS6110 S: 73% E: 95,5%, de la misma forma se determinaron los valores predictivos positivos y negativos (VPP y VPN) BK: 88,9% y 94,8%, Genotype S: 77,8% E: 94,5%; PCRIS6110 S: 90% y 95,7%. Se concluyó bajo análisis de curva ROC que las pruebas muestran un rendimiento diagnóstico similar por separado en el diagnóstico de TB en pacientes con VIH, aumentando su rendimiento diagnostico cuando se combinan

Fluorodeoxyglucose-positron emission tomography scan-positive recurrent papillary thyroid cancer and the prognosis and implications for surgical management.

BACKGROUND To compare outcomes for patients with recurrent or persistent papillary thyroid cancer (PTC) who had metastatic tumors that were fluorodeoxyglucose-positron emission tomography (FDG-PET) positive or negative, and to determine whether the FDG-PET scan findings changed the outcome of medical and surgical management. METHODS From a prospective thyroid cancer database, we retrospectively identified patients with recurrent or persistent PTC and reviewed data on demographics, initial stage, location and extent of persistent or recurrent disease, clinical management, disease-free survival and outcome. We further identified subsets of patients who had an FDG-PET scan or an FDG-PET/CT scan and whole-body radioactive iodine scans and categorized them by whether they had one or more FDG-PET-avid (PET-positive) lesions or PET-negative lesions. The medical and surgical treatments and outcome of these patients were compared. RESULTS Between 1984 and 2008, 41 of 141 patients who had recurrent or persistent PTC underwent FDG-PET (n = 11) or FDG-PET/CT scans (n = 30); 22 patients (54%) had one or more PET-positive lesion(s), 17 (41%) had PET-negative lesions, and two had indeterminate lesions. Most PET-positive lesions were located in the neck (55%). Patients who had a PET-positive lesion had a significantly higher TNM stage (P = 0.01), higher age (P = 0.03), and higher thyroglobulin (P = 0.024). Only patients who had PET-positive lesions died (5/22 vs. 0/17 for PET-negative lesions; P = 0.04). In two of the seven patients who underwent surgical resection of their PET-positive lesions, loco-regional control was obtained without evidence of residual disease. CONCLUSION Patients with recurrent or persistent PTC and FDG-PET-positive lesions have a worse prognosis. In some patients loco-regional control can be obtained without evidence of residual disease by reoperation if the lesion is localized in the neck or mediastinum.