Predicting High-Risk Patients with Lung Adenocarcinoma: The Power of Plasma Cell-Related Genes.

Abstract

Introduction: The incidence of lung cancer remains high worldwide and is still the leading cause of cancer-related deaths globally. The primary reason for this is that the vast majority of patients are diagnosed only when the disease has progressed to an advanced stage or metastasized. Therefore, early diagnosis of lung cancer is crucial. Approximately 85% of lung cancers are non-small cell lung cancer (NSCLC). As a type of NSCLC, lung adenocarcinoma (LUAD) is more prone to distant metastasis and has a poorer prognosis. It is often primarily treated with immunotherapy. Currently, immunotherapy mainly focuses on T cells. However, with the deepening of research, plasma cells, which have long been considered non-essential in anti-tumor responses, have been increasingly recognized for their critical role.

Methods: This study integrates data from TCGA, Tumor Immune Single-Cell Hub 2 (TISCH), and 10X databases, focusing on plasma cells. Through clustering analysis and Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis, it aimed to establish a predictive model for high-risk LUAD patients and further explore the relationship between the risk model and immune cells, with the goal of providing potential predictions for the efficacy of immunotherapy for patients. Additionally, we conducted drug sensitivity analysis and immune checkpoint analysis to identify drugs with potential benefits for the clinical management of high-risk patients. At the same time, we performed further immune checkpoint analysis to identify potential therapeutic targets for LUAD.

Results: By integrating the TCGA, TISCH, and 10X databases and focusing on plasma cells through clustering analysis and LASSO regression analysis, we established a predictive model for high-risk LUAD patients involving four feature genes: BEX5, CASP10, EPSTI1, and LY9. The ROC and results demonstrate that our model has strong predictive performance. Additionally, we found that the risk model is closely related to immune cells, providing the potential for predicting the efficacy of immunotherapy for patients. Subsequently, we conducted drug sensitivity analysis and immune checkpoint analysis, revealing that the majority of drugs are more sensitive to low-risk patients, while ABT-888, AS601245, and CCT007093 may have greater potential clinical benefits for high-risk patients. Immune checkpoint analysis showed significant differences in the expression of ADORA2A, BTLA, CD276, CD27, CD28, CD40LG, CD48, and TNFRSF14 between high-risk and low-risk patient groups, suggesting their potential as therapeutic targets for LUAD.

Conclusion: We constructed a risk assessment model for LUAD patients based on these genes. This model achieved breakthroughs in predicting the prognosis of LUAD patients with different risk levels and identifying potential immune targets, which were validated in the TCGA-LUAD clinical samples.