<p>The first part of this thesis is concerned with trend modelling of breast cancer mortality rates. By using an age-period-cohort model, the relative contributions of period and cohort effects are evaluated once the unquestionable existence of the age effect is controlled for. The result of such a modelling gives indications in the search for explanatory factors. While this type of modelling is usually performed with 5-year period intervals, the use of 1-year period data, as in Paper I, may be more appropriate.</p><p>The main theme of the thesis is the evaluation of the ability to detect signals in x-ray images of breasts. Early detection is the most important tool to achieve a reduction in breast cancer mortality rates, and computer-aided detection systems can be an aid for the radiologist in the diagnosing process.</p><p>The evaluation of computer-aided detection systems includes the estimation of distributions. One way of obtaining estimates of distributions when no assumptions are at hand is kernel density estimation, or the adaptive version thereof that smoothes to a greater extent in the tails of the distribution, thereby reducing spurious effects caused by outliers. The technique is described in the context of econometrics in Paper II and then applied together with the bootstrap in the breast cancer research area in Papers III-V.</p><p>Here, estimates of the sampling distributions of different parameters are used in a new model for free-response receiver operating characteristic (FROC) curve analysis. Compared to earlier work in the field, this model benefits from the advantage of not assuming independence of detections in the images, and in particular, from the incorporation of the sampling distribution of the system's operating point.</p><p>Confidence intervals obtained from the proposed model with different approaches with respect to the estimation of the distributions and the confidence interval extraction methods are compared in terms of coverage and length of the intervals by simulations of lifelike data.</p>

Breast cancer is the most commonly diagnosed cancer among Canadian women; x-ray mammography is the leading screening technique for early detection. This work introduces a semi-automated technique for analyzing mammographic x-ray images to measure their degree of suspiciousness for containing abnormalities. The designed system applies the discrete wavelet transform to parse the images and extracts statistical features that characterize an images content, such as the mean intensity and the skewness of the intensity. A naïve Bayesian classifier uses these features to classify the images, achieving sensitivities as high as 99.5% for a data set containing 1714 images. To generate confidence levels, multiple classifiers are combined in three possible ways: a sequential series of classifiers, a vote-taking scheme of classifiers, and a network of classifiers tuned to detect particular types of abnormalities. The third method offers sensitivities of 99.85% or higher with specificities above 60%, making it an ideal candidate for pre-screening images. Two confidence level measures are developed: first, a real confidence level measures the true probability that an image was suspicious; and second, a normalized confidence level assumes that normal and suspicious images were equally likely to occur. The second confidence measure allows for more flexibility and could be combined with other factors, such as patient age and family history, to give a better true confidence level than assuming a uniform incidence rate. The system achieves sensitivities exceeding those in other current approaches while maintaining reasonable specificity, especially for the sequential series of classifiers and for the network of tuned classifiers.

Breast cancer is the most commonly diagnosed cancer among Canadian women; x-ray mammography is the leading screening technique for early detection. This work introduces a semi-automated technique for analyzing mammographic x-ray images to measure their degree of suspiciousness for containing abnormalities. The designed system applies the discrete wavelet transform to parse the images and extracts statistical features that characterize an images content, such as the mean intensity and the skewness of the intensity. A naïve Bayesian classifier uses these features to classify the images, achieving sensitivities as high as 99.5% for a data set containing 1714 images. To generate confidence levels, multiple classifiers are combined in three possible ways: a sequential series of classifiers, a vote-taking scheme of classifiers, and a network of classifiers tuned to detect particular types of abnormalities. The third method offers sensitivities of 99.85% or higher with specificities above 60%, making it an ideal candidate for pre-screening images. Two confidence level measures are developed: first, a real confidence level measures the true probability that an image was suspicious; and second, a normalized confidence level assumes that normal and suspicious images were equally likely to occur. The second confidence measure allows for more flexibility and could be combined with other factors, such as patient age and family history, to give a better true confidence level than assuming a uniform incidence rate. The system achieves sensitivities exceeding those in other current approaches while maintaining reasonable specificity, especially for the sequential series of classifiers and for the network of tuned classifiers.

The first part of this thesis is concerned with trend modelling of breast cancer mortality rates. By using an age-period-cohort model, the relative contributions of period and cohort effects are evaluated once the unquestionable existence of the age effect is controlled for. The result of such a modelling gives indications in the search for explanatory factors. While this type of modelling is usually performed with 5-year period intervals, the use of 1-year period data, as in Paper I, may be more appropriate. The main theme of the thesis is the evaluation of the ability to detect signals in x-ray images of breasts. Early detection is the most important tool to achieve a reduction in breast cancer mortality rates, and computer-aided detection systems can be an aid for the radiologist in the diagnosing process. The evaluation of computer-aided detection systems includes the estimation of distributions. One way of obtaining estimates of distributions when no assumptions are at hand is kernel density estimation, or the adaptive version thereof that smoothes to a greater extent in the tails of the distribution, thereby reducing spurious effects caused by outliers. The technique is described in the context of econometrics in Paper II and then applied together with the bootstrap in the breast cancer research area in Papers III-V. Here, estimates of the sampling distributions of different parameters are used in a new model for free-response receiver operating characteristic (FROC) curve analysis. Compared to earlier work in the field, this model benefits from the advantage of not assuming independence of detections in the images, and in particular, from the incorporation of the sampling distribution of the system's operating point. Confidence intervals obtained from the proposed model with different approaches with respect to the estimation of the distributions and the confidence interval extraction methods are compared in terms of coverage and length of the intervals by simulations of lifelike data.

No description available.

A proposta deste trabalho foi analisar diferentes metodologias de treinamento de uma rede neural convolucional profunda (CNN) para a detecção de distorção arquitetural mamária (DA) em imagens de mamografia digital. A DA é uma contração sutil do tecido mamário que pode representar o sinal mais precoce de um câncer de mama em formação. Os sistemas computacionais de auxílio ao diagnóstico (CAD) existentes ainda apresentam desempenho insatisfatório para a detecção da DA. Sistemas baseados em CNN têm atraído a atenção da comunidade científica, inclusive na área médica para a otimização dos sistemas CAD. No entanto, as CNNs necessitam de um grande volume de dados para serem treinadas adequadamente, o que é particularmente difícil na área médica. Dessa forma, foi realizada neste trabalho, uma comparação de diferentes abordagens de treinamento para uma arquitetura CNN avaliando-se o efeito de técnicas de geração de novas amostras (data augmentation) sobre o desempenho da rede. Para isso, foram utilizadas 240 mamografias digitais clínicas. Uma das redes (CNN-SW) foi treinada com recortes extraídos por varredura em janela sobre a área interna da mama (aprox. 21600 em média) e a outra rede (CNN-SW+) contou com o mesmo conjunto ampliado por data augmentation (aprox. 345000 em média). Para avaliar o método, foi utilizada validação cruzada por k-fold, gerando-se em rodízio, 10 modelos de cada rede. Os testes analisaram todas as ROIs extraídas da mama, sendo testados 14 mamogramas por fold, e obtendo-se uma diferença estatisticamente significativa entre os resultados (AUC de 0,81 para a CNN-SW e 0,83 para a CNN-SW+). Mapas de calor ilustraram as predições da rede, permitindo uma análise visual e quantitativa do comportamento de ambos os modelos. / The purpose of this work was to analyze different training methodologies of a deep convolutional neural network (CNN) to detect breast architectural distortion (AD) in digital mammography images. AD is a subtle contraction of the breast tissue that may represent the earliest sign of a breast cancer in formation. Current Computer-Aided Detection (CAD) systems still have an unsatisfactory performance on AD detection. CNN-based systems have attracted the attention of the scientific community, including in the medical field for CAD optimization. However, CNNs require a large amount of data to be properly trained, which is particularly difficult in the medical field. Thus, in this work, different training approaches for a CNN architecture are compared evaluating the effect of data augmentation techniques on the data set. For this, 240 clinical digital mammography were used. One of the networks (CNN-SW) was trained with regions of interest (ROI) extracted by a sliding window over the inner breast area (approx 21600 on average) and the other network (CNN-SW+) had the same set enlarged by data augmentation (about 345000 on average). To evaluate the method, k-fold cross-validation was used, generating 10 instances of each model. The tests looked at all the ROIs extracted from the breast (14 mammograms per fold), and results showed a statistically significant difference between both networks (AUC of 0.81 for CNN-SW and 0.83 for CNN-SW+). Heat maps illustrated the predictions of the networks, allowing a visual and quantitative analysis of the behavior of both models.

Use of multi-detector CT in lung cancer screening has the potential to detect smaller lung nodules with improved sensitivity. In this study the development of a Computer Aided Detection (CAD) system for lung nodules is reported. A combination of two segmentation approaches is used, to segment lung regions. Following segmentation, a selective enhancement filter is applied for ''initial'' identification of nodule seed points in lung regions. Candidate lung nodule regions were delineated with the use of a region growing algorithm, with thresholds provided by minimum error thresholding. False positive regions were subsequently removed using two Support Vector Machines (SVM) classifiers in cascade, utilizing a set of 6 morphological features extracted from corresponding nodule candidate regions of the enhanced and the original images. The proposed automated scheme was tested on a reference dataset of 21 cases provided by the Lung Imaging Database Consortium. System performance on a case and slice basis provided sensitivities of 91% and 81% respectively, both with an average of 5 FPs per slice. Further analysis of the slice dataset with respect to size, contrast and location of nodules provided sensitivities of 81%, 83% and 85% for nodules of small size, low contrast and near pleura. This CAD scheme may be a useful tool in assisting radiologists in lung nodule detection. / Χρήση υπολογιστικής αξονικής τομογραφίας με πολλαπλών ανιχνευτών στον πληθυσμιακό έλεγχο καρκίνου το πνεύμονα αναμένεται να συμβάλει θετικά λόγω της ικανότητας της να ανιχνεύει οζίδια του πνεύμονα μικρού μεγέθους με αυξημένη ευαισθησία. Σε αυτή την μελέτη περιγράφεται η ανάπτυξη συστήματος αυτόματης ανίχνευσης οζιδίων του πνεύμονα, με στόχο την αύξηση της ευαισθησίας σε πολυτομική αξονική τομογραφία. Το σύστημα ανίχνευσης οζιδίων αποτελείται από τρία στάδια, το στάδιο της τμηματοποίησης των πνευμονικών πεδίων, την αναγνώριση των αρχικών υποψηφίων περιοχών και τέλος την μείωση των ψευδώς θετικών ενδείξεων. Η τμηματοποίηση των πνευμονικών πεδίων πραγματοποιήθηκε με τον συνδυασμό δύο αυτόματων τεχνικών τμηματοποίησης. Στην συνέχεια ένα επιλεκτικό ενισχυτικό φίλτρο εφαρμόζεται στην περιοχή των πνευμονικών πεδίων, για την ανίχνευση τον αρχικών υποψηφίων οζιδίων και τον συντεταγμένων τους. Τα όρια των υποψήφιων οζιδίων καθορίστηκαν με την βοήθεια ενός αλγορίθμου οριοθέτησης περιοχής με τις σταθερές κατωφλιού να υπολογίζονται αυτόματα βάση τις τεχνικής που προτάθηκε από τον Kittler et al. Η μείωση των ψευδώς θετικών ενδείξεων πραγματοποιήθηκε με την εφαρμογή δύο ταξινομητών Support Vector Machines (SVM) σε σειρά, οι οποίοι χρησιμοποίησαν 6 μορφολογικά χαρακτηριστικά τα οποία υπολογίστηκαν από τις περιοχές των υποψηφίων οζιδίων στην ενισχυμένη αλλά και στην αρχική εικόνα. Το σύστημα το οποίο παρουσιάζεται σε αυτή την εργασία εφαρμόστηκε και δοκιμάστηκε σε βάση δεδομένων αναφοράς η οποία περιλαμβάνει 21 εξετάσεις, την οποία τις παρέχει το Lung Imaging Database Consortium ((LIDC). Η απόδοση του συστήματος σε επίπεδο εξέτασης και επίπεδο τομής ήταν αντίστοιχα 91% και 81% με 5 ψευδώς θετικές ενδείξεις αντίστοιχα. Περαιτέρω ανάλυση βάση του μεγέθους, αντίθεσης και θέσης των οζιδίων απέδωσε ευαισθησίες 81%, 83% και 85% για οζίδια μικρού μεγέθους, χαμηλής αντίθεσης και οζίδια που βρίσκονται στον υπεζοκότα. Το προτεινόμενο σύστημα μπορεί να αποδειχθεί χρήσιμο εργαλείο υποβοήθησης ανάγνωσης οζιδίων σε πολυτομική αξονική τομογραφία για τους ακτινολόγους.

Made available in DSpace on 2015-05-14T12:36:35Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 4611703 bytes, checksum: c5e21f035f26ff6959248fbd243a0217 (MD5) Previous issue date: 2012-04-20 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Breast cancer is the most common cause of death by cancer in the female population and a serious world health problem. The mammographic exam allows an early detection which reduces the mortality rate of the disease. Its efficiency has made it the standard procedure for breast cancer diagnosis. These reasons have led to the development of Computer-Aided Detection and Diagnosis (CADDx) systems that assist the physician by working as a second opinion in the diagnostic. One of the algorithms studied during the development of this work, the mass detection algorithm created by Ozekes et al, has shown great potential reaching 99% of sensibility when applied in the test group images. However, its many parameters and the need to manual calibrate them make it impossible to use it in the constructions of practical CADDx systems. This work presents an automatic method for mass detection in mammography based on the algorithm of Ozekes et al. Multilayer Perceptron artificial neural networks (ANN) are used as functional approximators to automatically calibrate the necessary parameters of the proposed method. The computation of the neural networks produces the values used as parameters for thresholding and template application stages. Feature selection and network topologies were chosen by means of empirical tests. Results show in its best configuration point 82% of sensibility and 7,51 false positives per image. After a false positive reduction, 74% of sensibility and 3,56 false positives per image were achieved. Future works include the study of a wider set of image features and preprocessing algorithms. / O câncer de mama é a causa mais comum de morte por câncer na população feminina e um sério problema de saúde mundial. A mamografia permite uma detecção precoce do câncer, reduzindo a mortalidade da doença. Sua eficiência tornou-a procedimento padrão para diagnóstico do câncer de mama. Essas razões levaram ao desenvolvimento de sistemas computadorizados para o auxílio à detecção e ao diagnóstico - em inglês, Computer-Aided Detection and Diagnosis (CADDx) - que auxiliam os profissionais da saúde provendo uma segunda opinião ao diagnóstico. Um dos algoritmos estudados durante o desenvolvimento do trabalho, o algoritmo para detecção de massas criado por Ozekes et al, mostrou grande potencial atingindo 99% de sensibilidade quando aplicado nas imagens testadas. Entretanto, seus muitos parâmetros, e a calibração manual de cada um deles, tornam impossível a aplicação do algoritmo na construção de sistemas CADDx reais. Esse trabalho apresenta um método automático para detecção de massas em mamografias baseado no algoritmo de Ozekes et al. Redes neurais artificiais (RNA) Perceptron multicamadas são usadas como aproximadores universais para a calibração dos parâmetros necessários ao método. A computação dessas redes produz os valores que deverão ser usados como parâmetros para as etapas de binarização e aplicação dos templates. A seleção de atributos e topologias das redes neurais foi definida empiricamente. Resultados mostram, na melhor configuração do sistema, 82% de sensibilidade 7,51 falsos positivos por imagem e, após uma redução de falsos positivos, 74% de sensibilidade e 3,56 de falsos positivos por imagem. Trabalhos futuros incluem o estudo de mais atributos e descritores de imagens além da experimentação de outros algoritmos para pré-processamento.

Submitted by Maria Aparecida (cidazen@gmail.com) on 2017-08-21T14:59:23Z No. of bitstreams: 1 Andre Pereira.pdf: 3105574 bytes, checksum: 06e2fe68d48179a3c62a46e447b82513 (MD5) / Made available in DSpace on 2017-08-21T14:59:23Z (GMT). No. of bitstreams: 1 Andre Pereira.pdf: 3105574 bytes, checksum: 06e2fe68d48179a3c62a46e447b82513 (MD5) Previous issue date: 2009-02-20 / Breast cancer is one of the major causes of mortality among women throughout the world. Presently, the analysis of breast radiography is the most used method to early detection of this kind of cancer. It enables the identification of anomalies at their initial stage, which is a fundamental factor for success in the treatment. The sensitivity of this kind of exam, although, depends on several factors, such as the size and the location of the abnormalities, density of the breast tissue, quality of the technical resources and radiologist's ability. This work presents a methodology that uses the K-Means clustering algorithm and the Template Matching technique for segmentation of suspicious regions. Next, geometry and texture features are extracted from each of these regions, being the texture described by the Simpson's Diversity Index, a statistic used in Ecology to measure the biodiversity of an ecosystem. Finally, this information is submitted to a Support Vector Machine so that the suspicious regions are classified into masses and non-masses. The methodology was tested with 650 mammographic images from the DDSM database, achieving 83.94% of accuracy, 83.24% of sensibility and 84.14% of specificity in average. / O câncer de mama é uma das maiores causas de mortalidade entre as mulheres no mundo todo. Atualmente, a análise da radiografia da mama é o recurso mais utilizado na detecção precoce desse tipo de câncer, pois possibilita a identificação de anomalias em sua fase inicial, fator fundamental para o sucesso do tratamento. A sensibilidade desse tipo de exame, no entanto, depende de diversos fatores, tais como tamanho e localização das anomalias, densidade do tecido mamário, qualidade dos recursos técnicos e habilidade do radiologista. Este trabalho apresenta uma metodologia para detecção de massas em imagens digitais de mamografias que poderá auxiliar o especialista em sua análise. O método proposto utiliza o algoritmo de agrupamento K-Means e a técnica de Template Matching para segmentar as regiões suspeitas de conterem massas. Em seguida, medidas de geometria e textura são extraídas de cada uma dessas regiões, sendo a textura descrita através do Índice de Diversidade de Simpson, uma estatística usada na Ecologia para mensurar a biodiversidade de um ecossistema. Finalmente, essas informações são submetidas a uma Máquina de Vetores de Suporte para que as regiões suspeitas sejam classificadas em massas ou não massas. A metodologia foi testada com 650 imagens mamográficas obtidas da base de dados DDSM, atingindo 83,94% de acurácia, 83,24% de sensibilidade, e 84,14% de especificidade em média.

Made available in DSpace on 2016-08-17T14:53:07Z (GMT). No. of bitstreams: 1 Stelmo Magalhaes Barros Netto.pdf: 2768924 bytes, checksum: bf6f24780a03adb4f2940b818c95f293 (MD5) Previous issue date: 2010-02-10 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / Lung cancer is still one of the most frequent types throughout the world. Its diagnosis is very difficult because its initial morphological characteristics are not well defined, and also because of its location in relation to the lung. It is usually detected late, fact that causes a large lethality rate. Facing these difficulties, many researches are done, concerning both detection and diagnosis. The objective of this work is to propose a methodology for computer-aided automatic lung nodule detection. The return of the development of such methodology is that its application will aid the doctor in the simultaneous detection of several nodules present in computerized tomography images. The methodology developed for automatic detection of lung nodules involves the use of a method of competitive learning, called Growing Neural Gas (GNG). The methodology still consists in the reduction of the volume of interest, by the use of techniques largely used in thorax extraction, lung extraction and reconstruction. The next stage is the application of the GNG in the resulting volume of interest, that together with the separation of the nodules from the various structures present in the lung form the segmentation stage, and, finally, through texture and geometry measurements, the classification as either nodule or non-nodule is performed. The methodology guarantees that nodules of reasonable size are found with sensibility of 86%, specificity of 91%, what results in accuracy of 91%, in average, for ten training and test experiments, in a sample of 48 nodules occurring in 29 exams. The false-positive per exam rate was of 0.138, for the 29 analyzed exams. / O câncer de pulmão ainda é um dos mais incidentes em todo mundo. Seu diagnóstico é de difícil realização, devido as suas características morfológicas iniciais não estarem bem definidas e também por causa da sua localização em relação ao pulmão. É geralmente detectado tardiamente, que tem como conseqüência uma alta taxa de letalidade. Diante destas dificuldades muitas pesquisas são realizadas, tanto em relação a sua detecção, quanto a seu diagnóstico. O objetivo deste trabalho é propor uma metodologia de detecção automática do nódulo pulmonar com o auxílio do computador. O ganho com o desenvolvimento desta metodologia, é que sua implementação auxiliará ao médico na detecção simultânea dos diversos nódulos presentes nas imagens de tomografia computadorizada. A metodologia de detecção de nódulos pulmonares desenvolvida envolve a utilização de um método da aprendizagem competitiva, chamado de Growing Neural Gas (GNG). A metodologia ainda consiste na redução do volume de interesse, através de técnicas amplamente utilizadas na extração do tórax, extração do pulmão e reconstrução. A etapa seguinte é a aplicação do GNG no volume de interesse resultante, que em conjunto com a separação do nódulo das diversas estruturas presentes formam a etapa de segmentação, e por fim, é realizada a classificação das estruturas em nódulo e não-nódulo, por meio das medidas de geometria e textura. A metodologia garante que nódulos com tamanho razoável sejam encontrados com sensibilidade de 86%, especificidade de 91%, que resulta em uma acurácia de 91%, em média, para dez experimentos de treino e teste, em uma amostra de 48 nódulos ocorridos em 29 exames. A taxa de falsos positivos por exame foi de 0,138, para os 29 exames analisados.