Dissertation > Excellent graduate degree dissertation topics show

Serum protein profiling combined with artificial neural network diagnosis of esophageal cancer pathological differentiation model of applied research

Author: WangBo
Tutor: DaiTianYang
School: Luzhou Medical College
Course: Surgery
Keywords: Esophageal Cancer Pathological differentiation Serum proteomic Artificial neural network diagnostic model SELDI-TOF-MS Gold Chip
CLC: R735.1
Type: Master's thesis
Year: 2010
Downloads: 28
Quote: 0
Read: Download Dissertation

Abstract

Objective: Esophageal cancer is malignancy of the high incidence of digestive system, after the five-year survival rate after surgery is still low, about 8% -30%, a serious threat to human health. In recent years, for cancer patients choose individualized treatment plan as demonstrated better short-term and medium-term efficacy far been more and more attention. Cancer patients individualized treatment options depend on accurate histological differentiation of clinical diagnosis. For patients with esophageal cancer esophageal pathological differentiation accurate clinical diagnosis of esophageal cancer patients can be individualized treatment options provide a useful basis. Protein profiling technology is a combination of mass spectrometry-based protein chip technology for a variety of diseases markers analysis platform, by measuring the quality of the protein to determine the type of protein that helps filter esophageal lesions associated protein changes. In this study, patients with esophageal cancer based on serum proteomic specific expression, the use of artificial neural network software (artificial neural networks, ANNs) to establish the degree of differentiation of esophageal cancer pathological diagnosis of artificial neural network model to study the degree of differentiation in esophageal clinical diagnosis of pathological serum Learning diagnostic significance. Methods: The experimental group (60 cases of esophageal cancer patients, including 23 cases of esophageal cancer poorly differentiated, moderately differentiated in 11 cases, scores of those 26 cases) and control group (30 cases of normal healthy people) serum samples of 90 cases, using surface enhanced laser desorption ionization - time of flight mass spectrometer (Surface-enhanced laser desorption ionization time of flight mass spectrometry, SELDI-TOF-MS) with a detection instrument gold chips to detect all serum samples; using Ciphergen Proteinehip software and Biomarker Wizard 3.1 software right mix of control sera obtained spectra for statistical analysis and calculate the differences, in order to filter out specific expression of protein mass to charge ratio; utilize specific expression of selected protein mass to charge ratio combining artificial neural network software, the establishment of esophageal cancer pathological differentiation artificial neural network diagnostic model; using the double-blind method to validate the established esophageal pathological differentiation model of artificial neural network diagnostic sensitivity and specificity. Results: After testing analysis, this study serum samples from 60 cases were screened out four specific expression of the protein has obvious mass peak (P lt; 0.05), their mass to charge ratio (mass / charge, M / Z) were 2332.2 , 4051.1,4260.0 and 4267.1, the use of screened four specific protein mass to charge ratio combining artificial neural network software to establish the degree of differentiation of esophageal pathology artificial neural network diagnostic model; then use the double-blind method for establishing the degree of differentiation of human esophageal pathology neural network diagnostic model for statistical testing to verify that the diagnostic sensitivity and specificity of the model. Statistical analysis showed that the experiment to establish the degree of differentiation of esophageal pathology diagnosis model of artificial neural networks in the cut-off of 0.3 to get the best diagnostic results, clinical diagnosis of pathological differentiation esophageal sensitivity of 87% and a specificity of 83.8% positive predictive value of 76.9%, negative predictive value of 91.2%, the diagnostic accuracy was 85%. Conclusion: This study successfully screened out four in the serum of patients with esophageal cancer-specific expression of the protein mass to charge ratio; 2. Successfully established esophageal tumor differentiation artificial neural network diagnostic model, but to be blind method validation and From an objective indicators to obtain satisfactory sensitivity and specificity; 3 esophageal pathology from the experiment to establish the degree of differentiation ANNs model predictions diagnostic efficiency analysis of the results can be seen, when the cut-off value of 0.3, ANNs model sensitivity and high specificity, it can be used for esophageal cancer clinical diagnosis of pathological differentiation.

Serum protein profiling combined with artificial neural network diagnosis of esophageal cancer pathological differentiation model of applied research

Abstract

Related Dissertations