While significant progress has been made in the treatment and management of cancer, challenges remain because of the complexity and the heterogeneous nature of the disease. The improvement that has been seen in survival rates reflects advancements not only in treatment, but also in early stage detection and diagnostics for certain cancers. In particular, early stage detection and treatment of cancer before it has metastasized to other organs has resulted in a dramatic improvement in patient survival rates. One area of research that has shown considerable promise in further advancing diagnostics and early cancer detection is nanotechnology. Specifically, semiconductor and metal nanoparticles have great potential to provide advanced technology platforms for ultrasensitive and multiplexed detection of disease markers and probe disease on the molecular level. Because they are in the same size regime as biological molecules, these nanoparticles exhibit unique interactions with proteins, nucleic acids and other biomarkers of interest for detecting and diagnosing disease. However, high-quality nanoparticles are often unsuited for use in complex biological environments because of their coatings and surface chemistry. In this work, we describe the design and development of polymer-coated nanoparticle imaging agents for use in blood, cell and tissue diagnostic applications. Low-molecular weight, amphiphilic polymers capable of noncovalent interactions with nanoparticle surface ligands and the aqueous environment were synthesized and characterized for use in nanoparticle coating applications. We demonstrate that the hydrophobic and hydrophilic interactions between the nanoparticle surface, the amphiphilic polymer and the aqueous solvent were able to drive the coating and water solubilization of quantum dots. Novel nanoparticle synthetic techniques were also developed using the amphiphilic polymers in a one-pot method to make high quality semiconductor and gold nanoparticles and stabilize and encapsulate the particles for transfer into water. Using the polymer functional groups as multidentate ligands, nanoparticles were synthesized with a high degree of size control and increased stability. In addition, by performing the synthesis in a noncoordinating amphiphilic solvent such as polyethylene glycol, nanoparticles were immediately transferred to water with the excess polymer forming a water soluble coating. Next, nanoparticle surface charge and how it relates to the nonspecific binding of nanoparticles in cells, tissues and other complex biological samples was studied. We have found that highly charged (negative and positive) particles exhibit significant nonspecific binding to biomolecules and other cellular components in biological environments. By reducing the surface charge through the incorporation of hydroxyl functional groups, we have nearly eliminated the nonspecific binding of quantum dots in blood, cells and tissues. Moreover, through crosslinking and altering the surface chemistry of the polymer-coated quantum dots, we have increased the stability of the nanoparticles while maintaining a small hydrodynamic size. Finally, we have investigated the use of the low-binding, hydroxyl quantum dots in tissue staining applications, where nonspecific binding presents a considerable challenge to detection sensitivity and specificity. A number of biomolecule conjugation techniques were examined for the coupling of quantum dots to antibody targeting molecules and preliminary staining experiments were performed.

Protein microarrays are becoming powerful tools to screen and identify tumor markers for cancer diagnosis, because of the multiplex detection and minute volume of sample requirement. Due to the diversity and variation in different cancers, no single tumor marker is sensitive and specific enough to meet strict diagnostic criteria. Therefore, a combination of tumor markers is required to increase sensitivity and to establish distinct patterns to increase specificity. To obtain reliable tests, the development of reproducible surface chemistry and immobilization procedure are crucial steps in the elaboration of efficient protein microarrays. In this thesis, 3D micro-structured glass slides were functionalized with various surface chemistries like silane monolayer (amino, epoxy and carboxy), and polymer layers of Jeff amine, chitosan, carboxymethyl dextran (CMD), maleic anhydride-alt-methyl vinyl ether copolymer (MAMVE) for physical adsorption or covalent binding with proteins. Surface characterizations, such as X-ray photoelectron spectroscopy (XPS) and Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR), confirmed the monolayer/polymer grafting on the glass slides. Colorimetric assay for determining amine density of three aminated surfaces demonstrated that APDMES had more grafting density than Jeffamine and chitosan. Contact angle measurements show that polymer surfaces were more hydrophilic than monolayer surfaces due to the increasing dosages of polar functional groups. Moreover, the parameters such as additives and pH of spotting buffer, probe concentration, blocking procedures etc, were optimized for tumor marker detection. Under the optimized conditions, antibody microarrays were validated with purified tumor antigens. The best analytical performances obtained for each tumor antigen tested were strongly dependent on functionalized surfaces, e.g. MAMVE exhibited best analytical performances for CEA andHsp60 while NHS leads to best results for PDI and CA19-9. Besides, the implemented antibody microarrays were applied to tumor marker detection from colorectal cancer sera. This evaluation shows the interest to combine several tumor markers on the same surface and the combination of tumor markers on their specific surface lead to remarkably increase the positive responses of tested cancer sera (even up to 100 %). A second type of microarrays (tumor-associated antigens - TAA microarrays) was designed to discriminate breast cancer patients from healthy donors through the detection of tumor autoantibodies. This study included a cohort of 29 breast cancer patients' and 28 healthy donors' sera. A panel of fiveTAAs (Hsp60, p53, Her2, NY-ESO-1 and Hsp70) immobilized on their respective optimized surface chemistry allowed to specifically detect over 82% of breast cancer patients.

Bakgrund: Cancer är en av de vanligaste diagnoserna som drabbar människor världen över. De senaste årens effektivare behandlingar har lett till att allt fler överlever eller lever längre med sin cancerdiagnos. Detta medför att ett ökat antal individer nu och i framtiden kommer att vara i behov av fler och bättre hälsofrämjande interventioner. WHO definierar hälsa som ett tillstånd där frånvaro av sjukdom inte betyder god hälsa utan även måste innefatta ett individuellt fysiskt, psykiskt och socialt välbefinnande. Syfte: Syftet med studien var att belysa effekterna av musikinterventioner som en god omvårdnadsinsats för patienter med en cancerdiagnos. Metod: Allmän litteraturstudie med sex kvantitativa studier, fyra kvalitativa och tre med mixad metod. Resultat: Resultatet visade att patienter med en cancerdiagnos använde musik som en hälsofrämjande intervention för att minska oro och ångest, lindra smärta samt som en copingstrategi. Resultatet visade även att musikinterventioner kunde fungera som ett stöd i en identitetsskapande process. Vidare visade resultatet att en del individer inte var hjälpta av musik som hälsofrämjande intervention. Musiken kunde i vissa fall förknippas med diagnosen och fick deltagarna att känna olust och kunde ge motsatt effekt än att lindra och stödja. Slutsats: Musikinterventioner har, för många individer med en cancerdiagnos, förmåga att ge tillgång till olika känslor och kan fungera som en copingstrategi. Det är även en flexibel och tillgänglig hälsofrämjande intervention som kan nå ut till många. / Background: Cancer is one of the most common diagnoses to affect people all across the globe. The last years more effective treatments have lead to more people surviving or living longer with their cancer diagnosis; which implicates that an increasing number of individuals now and in the future, will be in need of more and better health-promoting interventions. WHO defines health as a state in which absence of disease does not signify good health, but a state which also includes individual physical, mental and social well-being. Aim: The aim of the study was to highlight the effect of musical interventions for patients with diagnosed cancer. Method: A literary study which consists of six quantitative studies, four qualitative studies, and three of mixed methods. Results: The results showed that patients diagnosed with cancer used music as a health-promoting intervention to ease pain, distress and anxiety, as well as a strategy to cope with the diagnosis. The results also showed that musical interventions could function as a way of maintaining the identity the individual had before the disease. Furthermore showed the result that some individuals were not helped by music as a health-promoting intervention. Music in some cases were associated with the diagnoses, causing the participants to feel unease and having the opposite effect to relief and support.  Conclusion: Musical intervention, for most individuals with diagnosed cancer, has the ability to give participants access to all different emotions and cope with the disease. Musical intervention is also a health-promoting intervention which is flexible, easily adjustable with the ability to reach many people.

Electronic medical record databases (e.g. the Clinical Practice Research Datalink, CPRD) are increasingly used in epidemiological research. The CPRD has two formats of data: coded, which is the sole format used in almost all research; and free-text (or ‘hidden’), which may contain much clinical information but is generally unavailable to researchers. This thesis examines the ramifications of omitting free-text records from research. Cases with bladder (n=4,915) or pancreatic (n=3,635) cancer were matched to controls (n=21,718, bladder; n=16,459, pancreas) on age, sex and GP practice. Coded and text-only records of attendance for haematuria, jaundice and abdominal pain in the year before cancer diagnosis were identified. The number of patients whose entire attendance record for a symptom/sign existed solely in the text was quantified. Associations between recording method (coded or text-only) and case/control status were estimated (χ2 test). For each symptom/sign, the positive predictive value (PPV, Bayes' Theorem) and odds ratio (OR, conditional logistic regression) for cancer were estimated before and after supplementation with text-only records. Text-only recording was considerable, with 7,951/20,958 (37%) of symptom records being in that format. For individual patients, text-only recording was more likely in controls (140/336=42%) than cases (556/3,147=18%) for visible haematuria in bladder cancer (χ2 test, p<0.001), and for jaundice (21/31=67% vs 463/1,565=30%, p<0.0001) and abdominal pain (323/1,126=29% vs 397/1,789=22%, p<0.001) in pancreatic cancer. Adding text records reduced PPVs of visible haematuria for bladder cancer from 4.0% (95% CI: 3.5–4.6%) to 2.9% (2.6–3.2%) and of jaundice for pancreatic cancer from 12.8% (7.3–21.6%) to 6.3% (4.5–8.7%). Coded records suggested that non-visible haematuria occurred in 127/4,915 (2.6%) cases, a figure below that generally used for study. Supplementation with text-only records increased this to 312/4,915 (6.4%), permitting the first estimation of its OR (28.0, 95% CI: 20.7–37.9, p<0.0001) and PPV (1.60%, 1.22–2.10%, p<0.0001) for bladder cancer. The results suggest that GPs make strong clinical judgements about the probable significance of symptoms – preferentially coding clinical features they consider significant to a diagnosis, while using text to record those that they think are not.

Objetivo: Avaliar a acurácia da colposcopia utilizando a Classificação Colposcópica Internacional de 2002. Métodos: 3040 pacientes de população geral foram rastreadas para patologia cervical através de exame citopatológico, captura híbrida para HPV de alto risco e inspeção cervical. As colposcopias que resultaram em biópsia (n=468) executadas no rastreamento e acompanhamento destas pacientes foram gravadas, revistas por dois colposcopistas cegados e incluídas para análise. Resultados: Os observadores apresentaram excelente concordância (Kappa=0.843) no relato dos achados pela nova nomenclatura. A colposcopia apresentou sensibilidade de 86% e especificidade de 30.3% em diferenciar colo normal de colo anormal (LSIL, HSIL ou carcinoma); quando a colposcopia objetivava diferenciar colo normal ou LSIL de HSIL ou carcinoma, apresentou sensibilidade de 61.1% e especificidade de 94.4%. Os achados colposcópicos classificados como “maiores” pela nova classificação apresentaram valores preditivos positivos elevados para HSIL. Presença do achado colposcópico na zona de transformação e tamanho da lesão estavam associados a HSIL. Bordas externas definidas, associação de múltiplos achados distintos e presença de zona iodo negativa não estavam relacionados à gravidade das lesões. Conclusão: A colposcopia utilizando a Classificação Internacional de 2002 mostra-se um bom método de rastreamento, mas como método diagnóstico apresenta falhas, não podendo substituir a avaliação histológica. A categorização em achados colposcópicos “maiores” e “menores” apresentada pela nova classificação é adequada. Na realização da colposcopia, é importante também que a lesão seja situada em relação à zona de transformação e que seu tamanho seja indicado, já que estes foram fatores associados a lesões de alto grau. / Objectives: To evaluate the colposcopic accuracy according to 2002 International Colposcopic Classification. Methods: 3,040 women from the general population were screened for cervical pathology by Pap smear, high risk HPV Hybrid Capture and naked eye visual inspection. All colposcopic exams that needed biopsy (n=468) performed during screening or follow-up were recorded, reviewed by two blinded colposcopists and included for analysis. Results: The two observers showed excellent agreement (Kappa=0.843) on reporting colposcopic findings according to the new classification. Colposcopy had sensitivity of 86% and specificity of 30.3% when the objective was to distinguish normal cervix from abnormal (LSIL, HSIL or carcinoma); for distinguishing normal cervix or LSIL from HSIL or carcinoma, colposcopy had sensitivity of 61.1% and specificity of 94.4%. Colposcopic findings graded as “major changes” had the highest predictive positive values for detecting HSIL or carcinoma. Colposcopic findings at transformation zone and size of lesion were related to HSIL. Sharp outer border, multiple colposcopic findings and iodine negativity were not statically related to severe lesions. Conclusion: Colposcopy using the new International Classification is a considerable screening method, but its value for diagnoses is restricted and it is not possible to substitute histological sampling. Categorization in “major changes” and “minor changes” according to the new classification is appropriate. When performing colposcopy, it is important to describe where is the colposcopic finding, in or outside the transformation zone, and its size, whereas these characteristics were related to high grade lesions.

Cervical cancer is the most common malignancy after breast cancer in women worldwide. It accounts for 83% of all new cases and 85% cancer death in developing countries. In South Africa, cervical cancer is the common cancer in women with an annual crude incidence rate of 30.2 per 100,000 women and the highest rate were found to be in women between the ages of 66-69 years. Approximately 6800 women in S.A face new case of cervical cancer while accounting for 3700 cancer death annually. Because of unequal access to the health facilities, socio-economic differences, HPV and HIV infection, the rate of cervical cancer in black women is higher (42.1%) compared to the low rate in white women. Because of the name “cancer bush’ given to Sutherlandia Frutescence(S.F) plant by the traditional healers as well as Xhosas, Zulu, Sotho and cape Dutch for its anti-cancer activity, the plant was in this study to confirm its cytotoxic effect on the cervical cancer cell lines. Aim of the study: to evaluate the methanolic extracts of Sutherlandia Frutescens on cervical cancer cell lines. Materials and Methods: The MTT assay was performed to evaluate SiHa cell lines treated with methanolic extract of S.F (50μg/ml, 100μg/ml, 150μg/ml and 200μg/ml). The three compounds (Canavanine, GABA and Pinitol) were also evaluated for its anti-tumour activity. The cell growth was then showed in real time using Xcilligence. Flow cytometry was employed to determine the mode of action. Caspase 3/7 assay was performed to confirm if cell death was via caspase-dependent or independent and ATP was done to assess the ATP level in S.F treated cells. Results: MTT shows a significant decline in cells treated with 50μg/ml, 100μg/ml and 200μg/ml of the extract and 50μg/ml was concluded to be the concentration at which 50% of the cells die. The ATP results are inconsistent with MTT result; the ATP level increased in S.F treated cells. Cell index which represents the quantitative measure of cell growth in real time decline upon treatment with 50μg/ml. Flow cytometry showed cells are dying by apoptosis and the cell cycle arrest is mostly in the S phase. The cell death was caspase-dependent as it shows an increased luminescence which is proportional to the number of caspase. The concentrations of the compounds used, Canavanine (1000μM, 1500μM and 2500μM), GABA (100μM, 300μM and 500μM) and Pinitol (30μM, 90μM and 120μM) induce cell death and cell death shows to decrease after the maximum concentration. Conclusion: Sutherlandia Frutescence has proven with number of research that it induces cell death via apoptosis. After evaluating its cytotoxicity, the plant shows to be a promising anti-tumor agent that needs to be clinically proven. / Life Sciences / M. Sc. (Life Science)

A presente dissertação propõe métodos para mineração de dados para diagnóstico de câncer de mama (CM) baseado na seleção de variáveis. Partindo-se de uma revisão sistemática, sugere-se um método para a seleção de variáveis para classificação das observações (pacientes) em duas classes de resultado, benigno ou maligno, baseado na análise citopatológica de amostras de célula da mama de pacientes. O método de seleção de variáveis para categorização das observações baseia-se em 4 passos operacionais: (i) dividir o banco de dados original em porções de treino e de teste, e aplicar a ACP (Análise de Componentes Principais) na porção de treino; (ii) gerar índices de importância das variáveis baseados nos pesos da ACP e na percentagem da variância explicada pelos componentes retidos; (iii) classificar a porção de treino utilizando as técnicas KVP (k-vizinhos mais próximos) ou AD (Análise Discriminante). Em seguida eliminar a variável com o menor índice de importância, classificar o banco de dados novamente e calcular a acurácia de classificação; continuar tal processo iterativo até restar uma variável; e (iv) selecionar o subgrupo de variáveis responsável pela máxima acurácia de classificação e classificar a porção de teste utilizando tais variáveis. Quando aplicado ao WBCD (Wisconsin Breast Cancer Database), o método proposto apresentou acurácia média de 97,77%, retendo uma média de 5,8 variáveis. Uma variação do método é proposta, utilizando quatro diferentes tipos de kernels polinomiais para remapear o banco de dados original; os passos (i) a (iv) acima descritos são então aplicados aos kernels propostos. Ao aplicar-se a variação do método ao WBCD, obteve-se acurácia média de 98,09%, retendo uma média de 17,24 variáveis de um total de 54 variáveis geradas pelo kernel polinomial recomendado. O método proposto pode auxiliar o médico na elaboração do diagnóstico, selecionando um menor número de variáveis (envolvidas na tomada de decisão) com a maior acurácia, obtendo assim o maior acerto possível. / This dissertation presents a data mining method for breast cancer (BC) diagnosis based on selected features. We first carried out a systematic literature review, and then suggested a method for feature selection and classification of observations, i.e., patients, into benign or malignant classes based on patients’ breast tissue measures. The proposed method relies on four operational steps: (i) split the original dataset into training and testing sets and apply PCA (Principal Component Analysis) on the training set; (ii) generate attribute importance indices based on PCA weights and percent of variance explained by the retained components; (iii) classify the training set using KNN (k-Nearest Neighbor) or DA (Discriminant Analysis) techniques, eliminate irrelevant features and compute the classification accuracy. Next, eliminate the feature with the lowest importance index, classify the dataset, and re-compute the accuracy. Continue such iterative process until one feature is left; and (iv) choose the subset of features yielding the maximum classification accuracy, and classify the testing set based on those features. When applied to the WBCD (Wisconsin Breast Cancer Database), the proposed method led to average 97.77% accurate classifications while retaining average 5.8 features. One variation of the proposed method is presented based on four different types of polynomial kernels aimed at remapping the original database; steps (i) to (iv) are then applied to such kernels. When applied to the WBCD, the proposed modification increased average accuracy to 98.09% while retaining average of 17.24 features from the 54 variables generated by the recommended kernel. The proposed method can assist the physician in making the diagnosis, selecting a smaller number of variables (involved in the decision-making) with greater accuracy, thereby obtaining the highest possible accuracy.

Objetivo: Avaliar a acurácia da colposcopia utilizando a Classificação Colposcópica Internacional de 2002. Métodos: 3040 pacientes de população geral foram rastreadas para patologia cervical através de exame citopatológico, captura híbrida para HPV de alto risco e inspeção cervical. As colposcopias que resultaram em biópsia (n=468) executadas no rastreamento e acompanhamento destas pacientes foram gravadas, revistas por dois colposcopistas cegados e incluídas para análise. Resultados: Os observadores apresentaram excelente concordância (Kappa=0.843) no relato dos achados pela nova nomenclatura. A colposcopia apresentou sensibilidade de 86% e especificidade de 30.3% em diferenciar colo normal de colo anormal (LSIL, HSIL ou carcinoma); quando a colposcopia objetivava diferenciar colo normal ou LSIL de HSIL ou carcinoma, apresentou sensibilidade de 61.1% e especificidade de 94.4%. Os achados colposcópicos classificados como “maiores” pela nova classificação apresentaram valores preditivos positivos elevados para HSIL. Presença do achado colposcópico na zona de transformação e tamanho da lesão estavam associados a HSIL. Bordas externas definidas, associação de múltiplos achados distintos e presença de zona iodo negativa não estavam relacionados à gravidade das lesões. Conclusão: A colposcopia utilizando a Classificação Internacional de 2002 mostra-se um bom método de rastreamento, mas como método diagnóstico apresenta falhas, não podendo substituir a avaliação histológica. A categorização em achados colposcópicos “maiores” e “menores” apresentada pela nova classificação é adequada. Na realização da colposcopia, é importante também que a lesão seja situada em relação à zona de transformação e que seu tamanho seja indicado, já que estes foram fatores associados a lesões de alto grau. / Objectives: To evaluate the colposcopic accuracy according to 2002 International Colposcopic Classification. Methods: 3,040 women from the general population were screened for cervical pathology by Pap smear, high risk HPV Hybrid Capture and naked eye visual inspection. All colposcopic exams that needed biopsy (n=468) performed during screening or follow-up were recorded, reviewed by two blinded colposcopists and included for analysis. Results: The two observers showed excellent agreement (Kappa=0.843) on reporting colposcopic findings according to the new classification. Colposcopy had sensitivity of 86% and specificity of 30.3% when the objective was to distinguish normal cervix from abnormal (LSIL, HSIL or carcinoma); for distinguishing normal cervix or LSIL from HSIL or carcinoma, colposcopy had sensitivity of 61.1% and specificity of 94.4%. Colposcopic findings graded as “major changes” had the highest predictive positive values for detecting HSIL or carcinoma. Colposcopic findings at transformation zone and size of lesion were related to HSIL. Sharp outer border, multiple colposcopic findings and iodine negativity were not statically related to severe lesions. Conclusion: Colposcopy using the new International Classification is a considerable screening method, but its value for diagnoses is restricted and it is not possible to substitute histological sampling. Categorization in “major changes” and “minor changes” according to the new classification is appropriate. When performing colposcopy, it is important to describe where is the colposcopic finding, in or outside the transformation zone, and its size, whereas these characteristics were related to high grade lesions.

abstract: Early detection and treatment of disease is paramount for improving human health and wellness. Micro-scale devices promote new opportunities for the rapid, cost-effective, and accurate identification of altered biological states indicative of disease early-onset; these devices function at a scale more sensitive to numerous biological processes. The application of Micro-Electro-Mechanical Systems (MEMS) in biomedical settings has recently emerged and flourished over course of the last two decades, requiring a deep understanding of material biocompatibility, biosensing sensitively/selectively, biological constraints for artificial tissue/organ replacement, and the regulations in place to ensure device safety. Capitalizing on the inherent physical differences between cancerous and healthy cells, our ultra-thin silicone membrane enables earlier identification of bladder cancer—with a 70% recurrence rate. Building on this breakthrough, we have devised an array to multiplex this sample-analysis in real-time as well as expanding beyond bladder cancer. The introduction of new materials—with novel properties—to augment current and create innovative medical implants requires the careful analysis of material impact on cellular toxicity, mutagenicity, reactivity, and stability. Finally, the achievement of replacing defective biological systems with implanted artificial equivalents that must function within the same biological constraints, have consistent reliability, and ultimately show the promise of improving human health as demonstrated by our hydrogel check valve. The ongoing proliferation, expanding prevalence, and persistent improvement in MEMS devices through greater sensitivity, specificity, and integration with biological processes will undoubtedly bolster medical science with novel MEMS-based diagnostics and therapeutics. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2018

Capes / A neoplasia é um grande desafio para os pesquisadores devido a sua alta complexidade. Apesar dos avanços em diagnósticos, os estudos apontam que, além da análise de dados, são necessários métodos que otimizem e auxiliem o processo de tomada de decisão. Neste sentido, a redução de dimensionalidade de dados tem contribuído significativamente, auxiliando nesse processo, devido à quantidade de genes (atributos), ser muito ampla comparada ao número de amostras (classes). Este trabalho, portanto, visa fornecer um estudo comparativo entre dois métodos de redução de dimensionalidade, aplicados em três bases de dados no domínio de expressão gênica: LungCancer-Michigan, LungCancer-Ontario e LungCancerHarvard, todas relacionadas ao câncer de pulmão. Os métodos aplicados foram: Seleção de Atributos e Análise de Componentes Principais (PCA), ambos usados como uma etapa de pré-processamento na Mineração de Dados. Os algoritmos de classificação escolhidos foram: Naive Bayes, SVM, J48, 1-NN, 3-NN, 5-NN e 7-NN. Foi utilizado o Weka como software para procedimentos de análise. Uma série de experimentos foi realizada para avaliar a acurácia e aplicabilidade dos algoritmos para ambos os métodos. Como resultado, foram evidenciados avanços significativos nas taxas de acerto (acurácia) dos classificadores envolvendo os métodos empregados, utilizando como critério de avaliação a Validação Cruzada. A abordagem Wrapper, do método de Seleção de Atributos, obteve os melhores resultados para as três bases de dados analisadas. O método de Análise de Componentes Principais, mesmo apresentando taxa de acerto inferior, não pode ser descartado. Os algoritmos Naive Bayes, SVM e 1-NN foram os que apresentaram melhor desempenho dentre as bases. Foram denotados os atributos (genes) que apresentaram maior frequência nas bases de dados. Portanto, a partir dos subconjuntos escolhidos, estes podem ser submetidos a análises específicas, no intuito de direcionar diagnósticos mais precisos. / Neoplasm is a major challenge for researchers because of its high complexity. Despite advances in diagnosis, studies point out that in addition to data analysis, methods to optimize and aid the decision-making process are necessary. In this sense, the dimensionality reduction of data has contributed significantly, helping in this process, due to the large number of genes (attributes) compared to the number of samples (classes). This work, therefore, aims to provide a comparative study between two methods of dimensionality reduction, applied to three databases in the field of gene expression: LungCancer-Michigan, LungCancer-Ontario and LungCancer-Harvard, all related to lung cancer. The methods applied were: Attribute Selection and Principal Component Analysis (PCA), both used as a pre-processing step in Data Mining. The classification algorithms chosen were Naive Bayes, SVM, J48, 1-NN, 3-NN, 5-NN and 7-NN. Weka was used as a software for analyses procedures. A series of experiments was performed to evaluate the accuracy and applicability of the algorithms for both methods. As a result, significant advances in the hit rate (accuracy) of the classifiers involving the methods were evidenced, using Cross-Validation as the assessment criterion. The Wrapper approach, from the Attribute Selection method, obtained the best results for the three analyzed databases. The Principal Component Analysis method, even presenting lower hit rate, could not be ruled out. The Naive Bayes, SVM and 1-NN algorithms presented the best performance within the databases. The attributes (genes) which presented the highest frequency in the databases were denoted. Therefore, from the chosen subsets, these can be submitted to specific analyzes in order to direct more precise diagnoses.