Lung Cancer Early Diagnosis Model Combined With Exhaled Markers, Images And Risk Factors

Lung cancer, which is the public enemy of all mankind, is by far the most commonly diagnosed cancer as well as the leading cause of cancer death globally. In recent years, the incidence of lung cancer in our country presents a trend of rapid growth. However, the natural features of easy metastases and difficult identification of early symptoms, resulting in a short window of treatment and a low cure rate of lung cancer. Therefore, for extending therapeutic window and improving the survival rate of patients, the early diagnosis of lung cancer was particularly important.In this study, the volatile organic compounds (VOCs) in exhaled breath, which were related with lung cancer (LC), were combined with clinical imaging and LC risk factors to model the early diagnosis of LC. The main contents include:1) Design and implementation of detective hardware platforms and processes:the VOCs in exhaled breath of 149 cases of normal healthy people,27 patients with benign lung diseases and 60 patients with LC were collected by the gas extraction apparatus and analyzed by TD-GC-MS. The assessment questionnaire of lung cancer risk factors was designed, and a database which was based on the questionnaire information was established. Imaging information of participants involved in the study were analyzed, classified and recorded to the database. The paper also studied the exhaled collection process, the sample pretreatment process, the sample logistics process and the quality control process.2) Design and implementation of the software about data analysis and interactive applications:The software of VOCs analyses, that were mainly to deal with the background of all breath samples and realize batch processing and data import, were completed. Meanwhile, an application on mobile terminal was designed to facilitate the inquiry of test results for the subjects.3) Research of correlation between lung cancer risk factors and markers in exhaled breath:the correlation between the VOCs in breath, imaging, smoking behavior and risk factors such as gender, age were analyzed in the paper. And 13,10,3, 2 and 1 VOCs were correlated with imaging, smoking intensity, smoking status, years of smoking, depth of inhalation, respectively. In addition,3 and 8 VOCs were correlated with gender and age.4) Based on the results of correlation studies to build and optimize the model for early diagnosis of lung cancer:31 VOCs that were significantly correlated with LC were combined with imaging information, smoking behavior, gender, and age to model the early diagnosis of LC. The model adopted the Logistic regression model, BP neural network model, and the Fisher discriminant model, respectively. The results showed that the diagnostic model using multiple factors can effectively improve the diagnosis of lung cancer when compared with the judgment only using imaging information.The diagnostic model that was united the exhaled breath, imaging diagnosis, and the risk factors were preliminary studied in this paper. It provided a new technique for secondary prevention of lung cancer. With the deepening of the study, we hope that these models can be used in early diagnosis and screening of lung cancer in the near future.