Banca de QUALIFICAÇÃO: LAURINDA FERNANDA SALDANHA SIQUEIRA

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE : LAURINDA FERNANDA SALDANHA SIQUEIRA
DATA : 24/11/2016
HORA: 15:00
LOCAL: Auditório do química 3
TÍTULO:

MULTIVARIATE CLASSIFICATION METHODS AND MID-INFRARED SPECTROSCOPY IN PROSTATE CANCER TISSUES

PALAVRAS-CHAVES:

Multivariate Classification Methods; FT-MIR; Prostate Cancer; Tissues

PÁGINAS: 160
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Química
SUBÁREA: Química Analítica
ESPECIALIDADE: Instrumentação Analítica
RESUMO:

This research is a contribution to the categorization of types of prostate cancer by chemometric classification methods applied in the mid-infrared spectra derived from human tissues. The multivariate classification methods have advantages as objectivity; speed; accuracy; easy procedure; it can be applied as exploratory or predictive methods; according to the method, it can be directly applied to the original data; it separate classes according to an optimal behavior found (linear, quadratic, nonlinear, etc). The application of these methods arises to reduce, select and classify useful information for diagnosis and categorization of cancer types, from the large amount information contained in the spectral region of the mid-infrared (400 to 4.000 cm^-1). The Mid-Infrared Spectroscopy (FT-MIR) on the other hand simplifies the sample preparation procedures and minimizes possible modifications, giving fast diagnosis; it is independent of intra variability and interobserver, and demonstrates high sensitivity and specificity, speed and results reliability. Based on the data from vibrational spectra of a given sample, it is possible to differentiate pathological conditions of tissues. With the general aim of applying multivariate classification methods (PCA, SPA, GA, LDA, QDA, SVM) and FT-MIR to differentiate degrees of prostate cancer via tissue analysis, it sought to identify spectral differences between the degrees of prostate cancer and to determine biochemical markers responsible for differentiation, besides to compare and evaluate the performance of classification methods. The  human tissue samples were provided by the Department of Pathology (DPAT / UFRN) and previously classified into three grades according to the Gleason grading system. 40 to 100 transmission spectra  were collected per sample (resolution 8 cm^-1, 32 scans). FT-MIR spectra were cut from 900–1800 cm^-1 (fingerprint region) and preprocessed by ESMC and 1st order Savitzky-Golay smoothing (15 points). Variable reduction methods (PCA) and variable selection methods (SPA and GA) was coupled to classification tecniques (LDA and QDA) to print robustness for the separation of prostate cancer stages and to identify biochemical markers responsible for differentiation. The models applied were compared according to classification rate, sensibility and specificity, and others quality metrics. All models achieved performance quality indicators satisfactory in the classification, especially the GA-LDA and GA-QDA models. The main spectral markers responsible for the classification, in different intensity orders, were: the region of secondary proteins (≈ 1591–1483 cm^-1), DNA and RNA (≈1000–1490 cm⁻¹) and protein phosphorylation ( ≈ 970 cm^-1). Changes in this regions can indicate metabolic changes caused by cancer stages. The results indicated that the multivariate classification methods coupled to FT-MIR enables differentiate pathological states of tissues even in the early stages of the disease with high sensibility and specificity. The techniques proposed may cause economic and social beneficial impacts provided by the precocius diagnosis and feasibility of the treatment in the early stages, which it will can allow gains in quality of life and survival of patients.

MEMBROS DA BANCA:
Externo ao Programa - 1913849 - EDGAR PERIN MORAES
Presidente - 1714946 - KASSIO MICHELL GOMES DE LIMA
Interno - 1523912 - TATIANA DE CAMPOS BICUDO