Optimization And Management Of Medical Laboratory Resources Based On Bioinformation Technology

BackgroundAlong with the completion of human genome project and initiation of proteome project, medical laboratory, as an emerging and comprehensive subject of clinical medicine has provided human beings with plenty of new technologies and monitoring indexes for the diagnosis, detection, treatment and prognosis of diseases. Huge data thus are produced, but this information on medical laboratory has not been used to the full. Additionally, the information management modes at present hospitals also cannot adapt to the rapid development pace of modern medical information. To optimize and manage medical resource information is an important aspect of current hospital management, which plays a central role in improving the comprehensive strength of hospitals.Colorectal cancer(CRC) is one of the common malignant tumors in digestive system. In recent years, people’s dietary structures and habits have changed, and the quality of their living standard has improved. The morbidity and mortality of the disease are on the increase. About 1/3 patients with colorectal cancer receive treatment at the advanced stage; and the relative survival rate is lower than 40%, which delay the treatment to a great extent. Therefore, establishing early diagnosis model based on serum markers can provide theoretical basis for the diagnosis and therapy of colorectal cancer.Congenital heart disease(CHD) is the most familiar congenital malformation at present. It is also the primary reason for the death of infants. The morbidity among newborns is 0.7% to 1%. Infants and young children’s health is being endangered seriously. Lots of factors are involved in heart development, which interact with each other temporally and spatially. The combined actions of genetic factors and environment factors in embryonic period lead to the dysplasia of the heart. Owing to the complex genetic mechanism of congenital heart disease, reasons resulting in heart malformation are not known. In addition, the complex and various types of the disease is also a big difficulty in the treatment and prevention of it. To this end, establishing the early diagnosis model based on related serum markers can provide a reference for the diagnosis and treatment of it.Based on the levels of clinical serum makers, this study built the early diagnosis models of colorectal cancer and congenital heart disease by logistic regression, support vector machine(SVM) and back propagation(BP) neural network, respectively. Then, serum markers with prediction and diagnosis values on related diseases were explored. This paper can not only support the diagnosis of the diseases, obtain serum markers closely associated with them, but provides technical assistance for the unified and effective management of medical laboratory resources in hospitals.PurposeTo establish the early diagnosis models of colorectal cancer and congenital heart disease and to evaluate their diagnostic value.MethodsLevels of related serum makers of colorectal cancer patients, congenital heart disease patients and healthy subjects were examined. Logistic regression, SVM and BP neural network were used respectively to establish the diagnosis models of colorectal cancer and congenital heart disease and to evaluate their diagnostic value. Through investigation, this study analyzed and assessed the application prospects of bioinformatic methods in congenital heart disease and colorectal cancer.Results1 Establishment of early diagnosis model of colorectal cancer serum tumor markers and analyzed by GO and KEGGExcept for AFP, levels of CEA, CA50, HSP60, CYFRA21-1, TPA, CA199, CA242, CA724, CA125 and other 11 serum markers of patients with colorectal cancer were significant higher than the benign controls(P<0.05). Logistic regression analysis indicated that, except for UGT1A8, other 11 indicators all had certain diagnostic value on colorectal cancer. The diagnosis accuracy of SVM model was 82.5%. The correct detection rate of BP neural network was 100% when 12 parameters were included. When indicators CEA, CA50, HSP60, CYFRA21-1, TPA, CA242 and UGT1A8(whose AUC were larger than 0.9) were included, the correct detection rate of BP neural network was 75%. The WWOX, P53, PTEN were analyzed by GO, and the regulation of apoptosis of colorectal cancer was drawn by KEGG.2 Establishment of early diagnosis model of congenital heart disease serum tumor markers and analyzed by GOSerum levels of c TnI, hs-CRP, BNP and Lp(a) of patients with congenital heart disease were obviously higher than the healthy controls. The difference was statistically significant(P<0.05). Logistic regression analysis showed they had a high diagnostic value except c Tn I and BNP. The diagnostic accuracy of SVM model was 77.5%. The correct detection rate of BP neural network was 72.5%. In this study, we analyzed the relationship between GATA4, FOG2 and Lp(a) BNP. The results showed that the relationship between gene expression and metabolic process, and Lp(a) was mainly related to lipid membrane transport and blood circulation.3 Application of bioinformatics in medical data analysisIn addition to other surveys and analyses, this study suggested the SVM diagnosis model, BP neural network diagnosis model and logistic regression analysis model established by related indicators all had favorable application prospects in data processing, diagnosis and prognosis evaluation of congenital heart disease and colorectal cancer.ConclusionsThe combined detection model based on logistic regression, the diagnosis models of colorectal cancer and congenital heart disease based on SVM and BP neural network all had high diagnostic value. They provided the basis for the early diagnosis of two diseases.