AUTHOR=Guizhen Zhang , Weiwei Zhu , Yun Wang , Guangying Cui , Yize Zhang , Zujiang Yu 

TITLE=An anoikis-based signature for predicting prognosis in hepatocellular carcinoma with machine learning

JOURNAL=Frontiers in Pharmacology

VOLUME=13

YEAR=2023

URL=https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2022.1096472

DOI=10.3389/fphar.2022.1096472

ISSN=1663-9812

ABSTRACT=<p><bold>Background:</bold> Hepatocellular carcinoma (HCC) is a common malignancy with high mortality worldwide. Despite advancements in diagnosis and treatment in recent years, there is still an urgent unmet need to explore the underlying mechanisms and novel prognostic markers. Anoikis has received considerable attention because of its involvement in the progression of human malignancies. However, the potential mechanism of anoikis-related genes (ANRGs) involvement in HCC progression remains unclear.</p><p><bold>Methods:</bold> We use comprehensive bioinformatics analyses to determine the expression profile of ANRGs and their prognostic implications in HCC. Next, a risk score model was established by least absolute shrinkage and selection operator (Lasso) Cox regression analysis. Then, the prognostic value of the risk score in HCC and its correlation with clinical characteristics of HCC patients were further explored. Additionally, machine learning was utilized to identify the outstanding ANRGs to the risk score. Finally, the protein expression of DAP3 was examined on a tissue microarray (TMA), and the potential mechanisms of DAP3 in HCC was explored.</p><p><bold>Results:</bold> ANRGs were dysregulated in HCC, with a low frequency of somatic mutations and associated with prognosis of HCC patients. Then, nine ANRGs were selected to construct a risk score signature based on the LASSO model. The signature presented a strong ability of risk stratification and prediction for overall survival in HCC patients.Additionally, high risk scores were closely correlated with unfavorable clinical features such as advanced pathological stage, poor histological differentiation and vascular invasion. Moreover, The XGBoost algorithm verified that DAP3 was an important risk score contributor. Further immunohistochemistry determined the elevated expression of DAP3 in HCC tissues compared with nontumor tissues. Finally, functional analyses showed that DAP3 may promote HCC progression through multiple cancer-related pathways and suppress immune infiltration.</p><p><bold>Conclusion:</bold> In conclusion, the anoikis-based signature can be utilized as a novel prognostic biomarker for HCC, and DAP3 may play an important role in the development and progression of HCC.</p>