Prognostic modelling of breast cancer patients: a benchmark of predictive models with external validation

Dissertação apresentada para obtenção do Grau de Doutor em Engenharia Electrotécnica e de Computadores – Sistemas Digitais e Percepcionais pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia

Análise de Desempenho e Recomendação para Soluções IVR

Mestrado em Engenharia Informática, Área de Especialização em Tecnologias do Conhecimento e da Decisão

Using data mining techniques to support breast cancer diagnosis

More than ever, there is an increase of the number of decision support methods and computer aided diagnostic systems applied to various areas of medicine. In breast cancer research, many works have been done in order to reduce false-positives when used as a double reading method. In this study, we aimed to present a set of data mining techniques that were applied to approach a decision support system in the area of breast cancer diagnosis. This method is geared to assist clinical practice in identifying mammographic findings such as microcalcifications, masses and even normal tissues, in order to avoid misdiagnosis. In this work a reliable database was used, with 410 images from about 115 patients, containing previous reviews performed by radiologists as microcalcifications, masses and also normal tissue findings. Throughout this work, two feature extraction techniques were used: the gray level co-occurrence matrix and the gray level run length matrix. For classification purposes, we considered various scenarios according to different distinct patterns of injuries and several classifiers in order to distinguish the best performance in each case described. The many classifiers used were Naïve Bayes, Support Vector Machines, k-nearest Neighbors and Decision Trees (J48 and Random Forests). The results in distinguishing mammographic findings revealed great percentages of PPV and very good accuracy values. Furthermore, it also presented other related results of classification of breast density and BI-RADS® scale. The best predictive method found for all tested groups was the Random Forest classifier, and the best performance has been achieved through the distinction of microcalcifications. The conclusions based on the several tested scenarios represent a new perspective in breast cancer diagnosis using data mining techniques.

Contribuição da produção em regime especial na garantia do abastecimento de energia elétrica e o impacto no planeamento da rede

Esta dissertação apresenta um estudo sobre a garantia de fornecimento de energia elétrica por parte dos produtores em regime especial com tecnologia cogeração e o impacto que estes traduzem na fase de planeamento da rede. Este trabalho foi realizado na Energias de Portugal - Distribuição (EDP-D) na direção de planeamento da rede (DPL). Para este estudo foi utilizado o caso de uma subestação com dezoito produtores em regime especial agregados à sua rede, em que dezasseis desses produtores são cogeração. A proposta de estudo para o caso concreto, passa pela análise das condições de funcionamento da subestação e apurar se a mesma necessita de alguma reformulação, tendo em vista as cargas a satisfazer atuais e possível incremento de carga futura. Considerando que a subestação está inserida num ambiente industrial e atendendo que existem diversos produtores de energia elétrica nas imediações da subestação. Para a resolução da garantia do fornecimento de energia por parte da cogeração, estudou-se a possibilidade de prever a energia produzida por estes produtores, através dos seguintes modelos de previsão: árvore de regressão, árvore de regressão com aplicação bagging e uma rede neuronal (unidirecional). Com a implementação destes modelos pretende-se estimar qual a potência que se pode esperar na garantia de abastecimento da carga, prevenindo maior solicitação de potência por parte da subestação. A metodologia utilizada baseia-se em simulações computacionais.

New records of Histoplasma capsulatum from wild animals in the Brazilian Amazon

Twenty-eight isolates of Histoplasma capsulation were obtained from eight species of forest mammals from the States of Amazonas, Pará and Rondônia in the Amazon Region of Brazil. Primary isolates were obtained by inoculating triturated liver and spleen tissue intradermally and intraperito-neally in hamsters. Mycological diagnosis in hamsters presenting lesions was confirmed by histopathology and culture on Sabouraud dextrose-agar. Infected hamsters developed signs of disease within two to nine months; all had disseminated visceral lesions and most also had skin lesions at the sites of inoculation. None of the hamsters inoculated with skin macerates of the original hosts developed histoplasmosis, and histopathological examination of the viscera of the wild hosts failed to reveal H. capsulation. Prevalence of infection was considerably higher in females than in males both for the opossum Didelphis marsupialis and for total wild animals (479) examined. It is proposed that canopy-dwelling mammals may acquire the infection from conidia borne on convective currents in hollow trees with openings at ground-level.

Interrupting Chagas disease transmission in Venezuela

The interruption of vectorial transmission of Chagas disease in Venezuela is attributed to the combined effects of ongoing entomoepidemiological surveillance, ongoing house spraying with residual insecticides and the concurrent building and modification of rural houses in endemic areas during almost five decades. The original endemic areas which totaled 750,000 km², have been reduced to 365,000 km². During 1958-1968, initial entomological evaluations carried out showed that the house infestation index ranged between 60-80%, the house infection index at 8-11% and a house density index of 30-50 triatomine bugs per house. By 1990-98, these indexes were further reduced to 1.6-4.0%, 0.01-0.6% and 3-4 bugs per house respectively. The overall rural population seroprevalence has declined from 44.5% (95% C.I.: 43.4-45.3%) to 9.2% (95% C.I.: 9.0-9.4%) for successive grouped periods from 1958 to 1998. The annual blood donor prevalence is firmly established below 1%. The population at risk of infection has been estimated to be less than four million. Given that prevalence rates are stable and appropriate for public health programmes, consideration has been given to potential biases that may distort results such as: a) geographical differences in illness or longevity of patients; b) variations in levels of ascertainment; c) variations in diagnostic criteria; and d) variations in population structure, mainly due to appreciable population migration. The endemic areas with continuous transmission are now mainly confined to piedmonts, as well as patchy foci in higher mountainous ranges, where the exclusive vector is Rhodnius prolixus. There is also an unstable area, of which landscapes are made up of grasslands with scattered broad-leaved evergreen trees and costal plains, where transmission is very low and occasional outbreaks are reported.

Characterization of Pinus spp. needles by gas chromatography and mass spectrometry: Application to plant-insect interactions

Dissertação apresentada para obtenção do Grau de Doutor em Ciências do Ambiente pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecn

Dynamic Analysis and Pattern Visualization of Forest Fires

This paper analyses forest fires in the perspective of dynamical systems. Forest fires exhibit complex correlations in size, space and time, revealing features often present in complex systems, such as the absence of a characteristic length-scale, or the emergence of long range correlations and persistent memory. This study addresses a public domain forest fires catalogue, containing information of events for Portugal, during the period from 1980 up to 2012. The data is analysed in an annual basis, modelling the occurrences as sequences of Dirac impulses with amplitude proportional to the burnt area. First, we consider mutual information to correlate annual patterns. We use visualization trees, generated by hierarchical clustering algorithms, in order to compare and to extract relationships among the data. Second, we adopt the Multidimensional Scaling (MDS) visualization tool. MDS generates maps where each object corresponds to a point. Objects that are perceived to be similar to each other are placed on the map forming clusters. The results are analysed in order to extract relationships among the data and to identify forest fire patterns.

Eco-epidemiological aspects of Trypanosoma cruzi, Trypanosoma rangeli and their vector (Rhodnius pallescens) in Panama

The eco-epidemiology of T. cruzi infection was investigated in the Eastern border of the Panama Canal in Central Panama. Between 1999 and 2000, 1110 triatomines were collected: 1050 triatomines (94.6%) from palm trees, 27 (2.4%) from periurban habitats and 33 (3.0%) inside houses. All specimens were identified as R. pallescens. There was no evidence of vector domiciliation. Salivary glands from 380 R. pallescens revealed a trypanosome natural infection rate of 7.6%, while rectal ampoule content from 373 triatomines was 45%. Isoenzyme profiles on isolated trypanosomes demonstrated that 85.4% (n = 88) were T. cruzi and 14.6% (n = 15) were T. rangeli. Blood meal analysis from 829 R. pallescens demonstrated a zoophilic vector behavior, with opossums as the preferential blood source. Seroprevalence in human samples from both study sites was less than 2%. Our results demonstrate that T. cruzi survives in the area in balanced association with R. pallescens, and with several different species of mammals in their natural niches. However, the area is an imminent risk of infection for its population, consequently it is important to implement a community educational program regarding disease knowledge and control measures.

Characterization of Cryptococcus neoformans isolated from urban environmental sources in Goiânia, Goiás State, Brazil

Cryptococcus neoformans is an opportunistic fungal pathogen that causes meningoencephalitis as the most frequent clinical presentation in immunocompromised patients, mainly in people infected by HIV. This fungus is an environmental encapsulated yeast, commonly found in soil enriched with avian droppings and plant material. A total of 290 samples of pigeon and the other avian droppings, soil, ornamental trees and vegetable material associated with Eucalyptus trees were collected to study environmental sources of Cryptococcus species in Goiânia, Goiás State. The determination of varieties, serotypes and the susceptibility in vitro to fluconazole, itraconazole and amphotericin B of C. neoformans isolates were performed. C. neoformans var. grubii (serotype A) was found in 20.3% (36/177) of pigeon dropping samples and in 14.3% (5/35) of samples of Eucalyptus. None of the environmental isolates of C. neoformans showed in vitro resistance to three antifungal agents. The knowledge of major route for human cryptococcal infection (inhalation of infectious particles from saprophytic sources) and a total of 60 C. neoformans isolates obtained from AIDS patients with cryptococcal meningitis between October 2001 and April 2002 justify the study of the habitats of these yeasts as probable sources of cryptococcosis in this city.

A Real-Time Optimization for 2R Manipulators

This work proposes a real-time algorithm to generate a trajectory for a 2 link planar robotic manipulator. The objective is to minimize the space/time ripple and the energy requirements or the time duration in the robot trajectories. The proposed method uses an off line genetic algorithm to calculate every possible trajectory between all cells of the workspace grid. The resultant trajectories are saved in several trees. Then any trajectory requested is constructed in real-time, from these trees. The article presents the results for several experiments.

Characterization of green zero-valent iron nanoparticles produced with tree leaf extracts

In the last decades nanotechnology has become increasingly important because it offers indisputable advantages to almost every area of expertise, including environmental remediation. In this area the synthesis of highly reactive nanomaterials (e.g. zero-valent iron nanoparticles, nZVI) is gaining the attention of the scientific community, service providers and other stakeholders. The synthesis of nZVI by the recently developed green bottom-up method is extremely promising. However, the lack of information about the characteristics of the synthetized particles hinders a wider and more extensive application. This work aims to evaluate the characteristics of nZVI synthesized through the green method using leaves from different trees. Considering the requirements of a product for environmental remediation the following characteristics were studied: size, shape, reactivity and agglomeration tendency. The mulberry and pomegranate leaf extracts produced the smallest nZVIs (5–10 nm), the peach, pear and vine leaf extracts produced the most reactive nZVIs while the ones produced with passion fruit, medlar and cherry extracts did not settle at high nZVI concentrations (931 and 266 ppm). Considering all tests, the nZVIs obtained from medlar and vine leaf extracts are the ones that could present better performances in the environmental remediation. The information gathered in this paper will be useful to choose the most appropriate leaf extracts and operational conditions for the application of the green nZVIs in environmental remediation.