Immune system development-related signature predicts prognosis and sorafenib-treatment resistance of hepatocellular carcinoma by intergrating machine learning and single-cell analyses
摘要
The development of the immune system (ISD) plays a pivotal role in both the genesis and progression of tumors, yet its specific functions in hepatocellular carcinoma (HCC) and the mechanisms behind sorafenib resistance remain elusive. In our investigation, we initially created 167 algorithms by amalgamating 10 distinct machine learning algorithms to identify an optimal and potent ISD-related signature (ISDRS), which proved to be superior to 73 previously published signatures. Moreover, we found that resistance to sorafenib was associated with both ISD levels and ISDRS. These, in turn, exhibited a positive correlation with the infiltration of T regulatory cells (Tregs) and a negative correlation with the levels of IL-6/IL-6R. Furthermore, through single-cell analysis, we uncovered a novel Tregs subgroup that contributes to sorafenib resistance and identified BATF as a potential biomarker for this subgroup.