Identification machine learning-based model of programmed cell-death patterns for investigating the immune microenvironment and prognosis of ovarian cancer patients after surgery and providing therapeutic strategies

Abstract

AbstractBackground Ovarian cancer (OV) is a highly aggressive and often fatal disease that frequently goes undetected until it has already metastasized. The classic treatment for ovarian cancer involves surgery followed by chemotherapy. However, despite the effectiveness of surgery, relapse is still a common occurrence. Unfortunately, there is currently no ideal predictive model for the progression and drug sensitivity of postoperative ovarian cancer patients. Cell death patterns play an important role in tumor progression and have the potential to be used as indicators of postoperative OV prognosis and drug sensitivity. Materials and methods A total of 12 PCD patterns were employed to construct the model. Bulk transcriptome, single-cell transcriptome, genomics, and clinical information were collected from TCGA-OV, GSE9891, GSE26712, GSE49997 and GSE63885. In addition, single-cell transcriptome data GSE210347 was procured from the Gene Expression Omnibus (GEO) database for subsequent analysis. Results In this study, a programmed cell death index (PCDI) was established using an 8-gene signature with the help of a machine learning algorithm. The PCDI was validated in four independent datasets, and it was found that patients with high-PCDI had a worse prognosis after surgery in OV. The investigation also revealed that PCDI is associated with chemokines, interleukins, interferons, and checkpoint genes, as well as important components of the immune microenvironment, as determined through a comprehensive analysis of bulk and single-cell transcriptomes. In addition, we also found that patients with low-PCDI values may exhibit sensitivity to immunotherapy, while those with high PCDI values may display increased responsiveness to tyrosine kinase inhibitors. Conclusion This study provides new insights into the significance of programmed cell death (PCD) patterns in ovarian cancer patients following surgery. Through a comprehensive analysis of different cell death patterns, we have developed a novel PCD model that can effectively predict the clinical prognosis and drug sensitivity of OV patients post-surgery.