Lead Investigator: Jing Yu, Sichuan university
Title of Proposal Research: Constructing a predictive model for lung cancer immunotherapy
Vivli Data Request: 8809
Funding Source: This study was supported by grants from the Key Technology Research and Development Program of the Sichuan Province (2022YFS0089).
Potential Conflicts of Interest: None
Summary of the Proposed Research:
Lung cancer is the most common malignancy in males and females, and the leading cause of cancer-related deaths. Its incidence continues to increase with the application of screening programs in the community, and the prognosis of patients is poor. Advances in therapies for lung cancer, especially immunotherapy, have improved survival and shifted the paradigm of treatment. Immunotherapy uses the immune system to fight cancer. It works by helping the immune system recognize and attack cancer cells. One type of immunotherapy is monoclonal antibody (MAB) treatment. This works by recognizing and finding specific proteins on cells. Each MAB recognizes one particular protein, and depending on the protein they are targeting, they work in different ways to kill cancer cells, or to stop them from growing. One type of MAB are immune checkpoint inhibitors (ICIs) which block proteins that stop the immune system from attacking cancer cells. Some ICIs have been found to have a durable response in clinical trials in lung cancer.
Various immune checkpoint inhibitors (ICIs), such as pembrolizumab, nivolumab, and ipilimumab, are widely used in the first-line treatment of patients with advanced/metastatic (where the cancer has spread to other parts of the body) lung cancer. However, only a limited fraction of patients benefit from ICIs, even when combined with other therapies such as chemotherapy.
There is an urgent need to explore potential biomarkers for optimizing patient selection to enhance the response rate. Biomarkers are molecules found in blood, other body fluids, or tissues that can indicate how a patient is responding to treatment. This project will analyze the differences in the data from patients who responded and who did not respond to immunotherapy treatment. This information will be used to create a model for predicting lung cancer immunotherapy response.
Statistical Analysis Plan:
We will integrate transcriptomic data collected from public databases for lung cancer immunotherapy with corresponding clinical information. Then, by analyzing the transcriptomics data, differential signature genes between patients responding to immunotherapy and patients not responding will be obtained. Next, we will further screen the differential signature genes to obtain the genes that have a clear correlation with the response to immunotherapy. Using machine learning, a multivariate model will be constructed to predict response to immunotherapy. The prediction accuracy of our method will be evaluated by ROC and survival curve analysis. At the same time, we will characterize the tumor immune characteristics of patients with high and low prediction scores, and explore the underlying mechanisms. For the TEMPUS cohort, we need transcriptomics, clinical information including treatment response. For samples with missing values, if we have a sufficient number of samples, we will also try to avoid imputation of missing values; if we do not have enough samples for statistical analysis, we will perform imputation.
Requested Studies:
Integration of tumor extrinsic and intrinsic features associates with immunotherapy response in non-small cell lung cancer
Data Contributor: Tempus Labs, Inc.
Study ID: T21.02
Sponsor ID: T21.02
Summary of Results:
We used the dataset in project #8809 to validate the model we built. The results showed that the Area Under the Curve (AUC) value of our model reached above 0.8 in the additional dataset, demonstrating the excellent performance of the model