Machine learning-based signature for prognosis and drug sensitivity in hepatocellular carcinoma using
Yufan Zhou1, Wei Liu1, Zhixue Fang1
1Department of General Surgery, Hunan Provincial People's Hospital, The First Affiliated Hospital of Hunan Normal University, Changsha, Hunan 410005, P.R. China.
Abstract:
Hepatocellular carcinoma (HCC) is one of the most prevalent tumors in the world and poses a considerable threat to global healthcare. The reprogramming of glucose metabolism in tumor cells has demonstrated a notable association with the genesis, advancement and resistance to the chemotherapy of malignancies. In the present study, 10 integrative machine learning algorithms were used to develop a glycolysis-related signature (GRS) using four datasets. Several predictive approaches were used to evaluate the performance of GRS in predicting the immunology response. In addition, in vitro experiments were performed to explore the biological functions of monocarboxylic acid transporter 1 (MCT1) in HCC. The optimal GRS developed by the Least Absolute Shrinkage and Selection Operator algorithm served as a risk factor for patients with HCC. Patients with HCC and a high-risk score experienced a poor prognosis, with the area under the curves of 1, 3 and 5-year receiver operating characteristic curves being 0.777, 0.787 and 0.766, respectively. A low-risk score indicated higher levels of CD8+ cytotoxic T cells and M1 macrophages, as well as an increased estimation of stromal and immune cells in malignant tumors score. Moreover, increased tumor mutational burden score and programmed cell death protein 1 and cytotoxic T-lymphocyte-associated protein 4 immunophenoscores, as well as decreased Tumor Immune Dysfunction and Exclusion and tumor escape scores were found in patients with HCC that had low-risk scores. The IC50 values of docetaxel, oxaliplatin, crizotinib and osimertinib were lower in HCC cases with a high-risk score. In addition, the gene set scores that were associated with angiogenesis and Notch signaling were higher in the high-risk score group. Downregulation of MCT1 inhibited the proliferation, migration and invasion of HCC cells and promoted the apoptosis of HCC cells. In conclusion, the present study developed a novel GRS for HCC, serving as an indicator for predicting clinical outcomes and responses to immunotherapy.
