Radiotherapy is often used to treat prostate cancer. While using high dose of radiation does kill cancer cells, it can cause toxicity in healthy tissues for some patients. It would be best to apply this treatment only to patients who are likely to be immune from such toxicity. This requires a classifier that can predict, before treatment, which patients are likely to exhibit severe toxicity. Here, we explore ways to use certain genetic features, called Single Nucleotide Polymorphisms (SNPs), for this task. This thesis uses several machine learning methods for learning such classifiers for predicting toxicity. This problem is challenging as there are a large number of features (164,273 SNPs) but only 82 samples. We explore an ensemble classification method for this problem, called Mixture Using Variance (MUV), which first learns several different base probabilistic classifiers, then for each query combines the responses of the different base classifiers based on their respective variances. The original MUV learns the individual classifiers using bootstrap sampling of the training data; we modify this by considering different subsets of the features for each classifier. We derive a new combination rule for base classifiers in the proposed setting and obtain some new theoretical results. Based on characteristics of our task, we propose an approach that involves first clustering the features before selecting only a subset of features from each cluster for each base classifier. Unfortunately, we were unable to predict radiation toxicity in prostate cancer patients using just the SNP values. However, our further experimental results reveal strong relation between correctness of a classifier in its prediction and the variance of the response to the corresponding classification query, which show that the main idea is promising.

No description available.

Made available in DSpace on 2016-08-10T10:38:35Z (GMT). No. of bitstreams: 1 Hellen da Silva Cintra de Paula.pdf: 4823796 bytes, checksum: 7a1015ff712efd961fb8cb5dfd8ad2ef (MD5) Previous issue date: 2012-03-14 / The purpose of this study was to evaluate the association of single nucleotide polymorphisms of ATM and TP53 genes in prostate cancer patients with skin, urinary and lower gastrointestinal systems morbidity after radiotherapy. These two genes encode important proteins of the DNA repair pathways. It is believed that their polymorphisms are likely to modify the response of normal tissues to radiation. A group of 50 patients of the Radiotherapy Service at Araújo Jorge Hospital (Associação de Combate ao Câncer em Goiás) was selected. After signing the informed consent agreement, a sample of peripheral blood was collected for subsequent DNA extraction and polymerase chain reaction (PCR) for amplification of ATM and TP53 gene fragments. Finally the amplified fragments were sequenced to verify the presence of an exchange G> A in the codon 1853 of the ATM gene and polymorphisms of TP53 gene (C> G in the codon 72, C>T in the codon 47, C>A in the position 11299, C>T in the position 11322 and one insertion of 16 base pairs in intron 3). The side effects were classified according to the Radiation Therapy Oncology Group (RTOG) score. On univariate analysis, hypertension was strongly associated with a decreased risk of late urinary toxicities (OR= 0,048, 95% CI 0,004 - 0,620; p= 0,022). Patients receiving hormone therapy had a significantly higher incidence of acute skin toxicity (RR=1,286, 95% CI 0,907 - 1,823; p = 0,029). The exchange C>T in the position 11322 of the TP53 gene (intron 3) was significantly associated with the risk of acute skin toxicity (RR=22,0, 95%CI 5,680 - 85,207; p=0,006). There wasn t association between the other TP53 and ATM polymorphisms analysed and the frequency of side effects (p>0,05). In this study it has been shown that the presence of hypertension seemed to be protective for the development of urinary late effects after radiotherapy. Hormone therapy was apparently determinant for the development of acute skin toxicity. Our data also revealed that a TP53 intronic polymorfism (11322 C>T) was associated to increased acute skin radiosensitivity. This observation corroborates the importance of investigating the genetic profile to predict adverse side effects in patients undergoing radiotherapy. / O propósito deste estudo foi avaliar a associação entre polimorfismos de base única nos genes ATM e TP53 em pacientes com câncer prostático e a morbidade na pele e nos sistemas urinário e gastrointestinal inferior após a radioterapia. Estes dois genes codificam proteínas importantes nas vias de reparo do DNA. Acredita-se que seus polimorfismos possam modificar a resposta do tecido normal a radioterapia. Foi selecionado um grupo de 50 pacientes do serviço de radioterapia do Hospital Araújo Jorge (Associação de Combate ao Câncer em Goiás). Após a assinatura do termo de consentimento livre e esclarecido, foi coletada a amostra de sangue periférico com subseqüente extração de DNA e amplificação gênica por meio de reação em cadeia da polimerase (PCR) para amplificar os fragmentos gênicos de ATM e TP53. Finalmente, os amplicons foram seqüenciados para verificar a presença da troca de G>A no códon 1853 do gene ATM e polimorfismos do gene TP53 (C>G no códon 72, C>T no códon 47, C>A na posição 11299, C>T na posição 11322 e uma inserção de 16 pares de bases no intron 3). Os efeitos adversos foram classificados de acordo com o escore do Radiation Therapy Oncology Group (RTOG). Por meio de análise univariada, a hipertensão se associou fortemente ao baixo risco de desenvolvimento de toxicidade urinária crônica (OR=0,048, 95%IC 0,004 - 0,620; p=0,022). Pacientes que foram submetidos à hormonioterapia mostraram uma incidência significativa de toxicidade de pele aguda (RR = 1,286, 95%IC 0,907 1,823; p=0,029). A troca C>T na posição 11322 do gene TP53 (intron 3) mostrou uma associação significativa com o risco de toxicidade aguda de pele (RR = 22,0, 95%IC 5,68 85,207; p=0,006). Não houve associação entre os outros polimorfismos de TP53 e ATM analisados e a freqüência de efeitos adversos (p>0,05). Foi demonstrado que a presença de hipertensão parece ser protetora para o desenvolvimento de efeitos urinários tardios após a radioterapia. A hormonioterapia foi aparentemente determinante no surgimento de toxicidade aguda de pele. Nossos dados revelaram ainda que um polimorfismo intrônico de TP53 (11322 C>T) também estava associado ao aumento de radiossensibilidade aguda de pele. Estas observações mostram a importância de se investigar o perfil genético para futuramente predizer os efeitos adversos de pacientes em radioterapia.

Made available in DSpace on 2016-08-10T10:38:48Z (GMT). No. of bitstreams: 1 JOAO LINO FRANCO BORGES.pdf: 1554450 bytes, checksum: 2c0f4d5c22814fcc52394573bf4f06f6 (MD5) Previous issue date: 2014-03-28 / The purpose of this study wasto evaluate the association of single nucleotide polymorphisms of ATM and TP53 genes in breast cancer patients with skin and subcutaneous systems morbidity after radiotherapy. These two genes encode important proteins of the DNA repair pathways. It is believed that their polymorphisms are likely to modify the response of normal tissues to radiation. A group of 78 patients of the Radiotherapy Service at Araújo Jorge Hospital (Associação de Combate ao Câncer em Goiás) was selected. After signing the informed consent agreement, a sample of peripheral blood was collected for subsequent DNA extraction and polymerase chain reaction (PCR) for amplification of ATM and TP53 gene fragments. Finally the amplified fragments were sequenced to verify the presence of an exchange G> A in the codon 1853 of the ATM gene and polymorphisms of TP53 gene (C> G in the codon 72). The side effects were classified according to the Radiation Therapy Oncology Group (RTOG) score. On univariate analysis,the expected effects of treatment with RT events showed that there was a significant association between aboost RT and its suspension with the acute effects of high-grade skin (OR = 3.05, 95% CI 1.175 to 7.928, p = 0.035) (OR = 3.808, 95% CI 1.213 to 11.958, p = 0.033), respectively. Side effects of acute high-grade skin were associated with a longer duration of treatment. (p=0.01 IC95% -11.546 - (-1,638). The type of surgery was significantly associated with late subcutaneous tissue (OR = 1.326, 95% CI 1.143 to 1.537, p = 0.016) side effects There was no association between polymorphisms TP53 and ATM analysed with acute and late side effects of skin nor subcutaneous tissue (RTOG ≥ 2) (p> 0.05). / O propósito deste estudo foi avaliar a associação entre polimorfismos de base única nos genes ATM e TP53 em pacientes com câncer de mama e a morbidade na pele após a radioterapia. Estes dois genes codificam proteínas importantes nas vias de reparo do DNA. Acredita-se que seus polimorfismos possam modificar a resposta do tecido normal a radioterapia. Foi selecionado um grupo de 78 pacientes do serviço de radioterapia do Hospital Araújo Jorge (Associação de Combate ao Câncer em Goiás). Após a assinatura do termo de consentimento livre e esclarecido, foi coletada a amostra de sangue periférico com subsequente extração de DNA e reação em cadeia da polimerase (PCR) para amplificar os fragmentos gênicos de ATM e TP53. Finalmente, os fragmentos amplificados foram sequenciados para verificaçãoda presença da troca de G>A no códon 1853 do gene ATM e polimorfismos do gene TP53 (C>G no códon 72). Os efeitos adversos foram classificados de acordo com o escore do Radiation Therapy Oncology Group (RTOG). Por meio de uma análise univariada, os eventos esperados do tratamento com RT mostraram que houve uma associação significativa entre o reforço da RT e a sua suspensão com os efeitos agudos de alto grau da pele (OR=3.05, IC95% 1,175-7,928, p=0,035) (OR=3.808, IC95% 1,213-11,958, p=0,033), respectivamente. Os efeitos colaterais agudos de alto grau na pele se associaram a um tempo maior de duração de tratamento. (p=0.01 IC95% -11.546- (-1,638). O tipo de cirurgia se associou significativamente aos efeitos colaterais tardios do tecido subcutâneo (OR=1,326, IC95% 1,143-1,537, p=0,016). Não houve associação entre os polimorfismos de TP53 e ATM analisados com os efeitos colaterais agudos e tardios de pele, nem tampouco do tecido subcutâneo(RTOG≥2) (p>0,05).