Immune - cell death index in hepatocellular carcinoma: a multi-omics and machine learning study for prognosis and
Yi Zhang1,2, Haiyu Zhao1,2, Yunpeng Zhai1,2
1The First Clinical Medical College, Zhengzhou University, Zhengzhou, Henan, China.
Abstract:
The heterogeneity of hepatocellular carcinoma (HCC) and individual disparities in immunotherapy response necessitate the urgent development of accurate evaluation tools. Programmed cell death (PCD) is implicated in the occurrence and development of HCC. Moreover, immune-related genes have a crucial role in cancer progression and patient prognosis. This study employed 10 clustering algorithms to conduct high-resolution molecular subtyping based on PCD-related genes, immune-related genes, microRNA, long non-coding RNA, and methylation data. Subsequently, we developed hepatocellular carcinoma consensus immune-cell death index (HICDI) by employing subtype-specific genes and merging 10 commonly used machine learning algorithms into 101 unique combination frameworks. Our HICDI score exhibited enhanced predictive ability compared to previously published HCC biomarkers. Patients with a low HICDI score exhibited higher overall survival and improved responses to immunotherapy. The high HICDI group exhibited a propensity for "cold" tumors marked by immune suppression and exclusion; however, drugs such as paclitaxel may present viable therapeutic options for these patients. We verified the model gene kinesin family member 2C through in vitro experiments, demonstrating its role as a potential oncogene affecting HCC progression and as a promising therapeutic target. Overall, HICDI possesses the potential for extensive applications in informing personalized treatment decisions and improving outcomes for patients with HCC.
