The Establishment and Clinical Application of a Prediction Model of Lung Cancer Distant Metastasis Based on the Genomic Characteristics of Circulating Tumor Cells

Lung cancer is the most common type of cancer in my country, but the 5-year survival time of lung cancer patients is only 17%. Among them, the biggest reason that affects the patient's prognosis is the metastasis of the tumor. There are very few clinical methods suitable for the treatment of metastatic lung cancer, and the curative effect is not good. Therefore, early monitoring and interventions to prevent distant colonization of metastases are the key to improving the survival of lung cancer. The preliminary research of this project found that circulating tumor cells in peripheral blood can be used as an effective means for clinical diagnosis and treatment of lung malignant tumors. Through the analysis of the difference in time and space metastasis of lung cancer patients, it is found that the genomes of different metastasis stages and metastatic organs of lung cancer are quite different , And is closely related to the patient's survival. For this reason, we propose the hypothesis that the genomic mutation characteristics of circulating tumor cells can detect tumor metastasis signals earlier than CT imaging diagnosis. To test this hypothesis, we will develop a cancer metastasis risk assessment system based on tumor genomics. First, we collect big data on the genome of primary and metastatic lung cancer from public databases, and use statistical methods to screen out genomic features that are significantly related to metastatic lung cancer and its metastatic colonization organs. Secondly, using these features to develop a set of machine learning models that can determine the risk of metastasis of a lung cancer based on its genome features. Finally, we applied the model to clinical practice. By detecting the circulating tumor cells of patients with primary lung cancer during the reexamination, we established a statistical noise reduction model to extract the genomic characteristics, and then substituted into the model to determine the circulating tumor cells carried by the patient Whether there is a risk of recurrence and metastasis. By comparing the imaging data in the review, we will verify whether the model detects early metastasis signals of lung cancer earlier than imaging methods. Ultimately, our model will aggregate genomic markers related to metastasis risk, explore their drug targeting, and provide powerful big data analysis support for early intervention in metastasis colonization and prolonging the survival of lung cancer patients. If the topic is demonstrated, it will help to clarify the use of tumor genome big data analysis to reveal the genomic driver mutations of metastatic lung cancer; demonstrate the feasibility of circulating tumor cell genome driver mutations to predict the risk of lung cancer metastasis; and finally clarify the PI3K/Akt/mTOR signal Can inhibitors of the pathway be used as a target for early intervention in lung cancer metastasis.