Early Diagnosis of Hepatocellular Carcinoma Using Machine Learning Method
Zi-Mei Zhang1, Jiu-Xin Tan1, Fang Wang1
1Key Laboratory for Neuro-Information of Ministry of Education, School of Life Sciences and Technology, Center for Informational Biology, University of Electronic Science and Technology of China, Chengdu, China.
Frontiers in Bioengineering and Biotechnology
|April 16, 2020
Summary
This study introduces an 11-gene signature for early hepatocellular carcinoma (HCC) detection. This machine learning model accurately identifies HCC from non-cancerous tissues, improving early diagnosis and patient survival rates.
Area of Science:
- Oncology
- Bioinformatics
- Computational Biology
Background:
- Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death globally.
- Current diagnostic methods for HCC lack sufficient accuracy, necessitating improved early detection strategies.
- Distinguishing HCC from cirrhosis without cancer (CwoHCC) remains challenging with conventional approaches.
Purpose of the Study:
- To develop a computational diagnostic model for early hepatocellular carcinoma (HCC) detection.
- To identify a gene signature for accurate HCC recognition using machine learning.
- To enhance clinical decision-making for HCC diagnosis and improve patient outcomes.
Main Methods:
- Applied a machine learning approach to microarray data from 1091 HCC and 242 CwoHCC samples.
- Utilized within-sample relative expression orderings (REOs) to extract gene expression descriptors.
- Employed maximum redundancy minimum relevance (mRMR) with incremental feature selection to identify an 11-gene-pair signature.
Main Results:
- An "11-gene-pair" signature was identified with outstanding discriminatory capability for HCC.
- The selected gene pairs demonstrated robust performance in recognizing HCC across independent datasets.
- The computational model successfully discriminated HCC and adjacent non-cancerous tissues from CwoHCC, even with minimal biopsy samples.
Conclusions:
- The identified 11-gene-pair serves as a potential molecular signature for hepatocellular carcinoma.
- The developed computational model offers a practical and effective tool for aiding early HCC diagnosis at an individual level.
- This approach holds promise for improving HCC treatment and patient survival through early detection.
