Identifying potential clinical syndromes of hepatocellular carcinoma using PSO-based hierarchical feature selection...

Area of Science:

• Oncology
• Computational Medicine
• Traditional Chinese Medicine

Background:

• Hepatocellular Carcinoma (HCC) is a prevalent malignancy often diagnosed late due to asymptomatic presentation.
• Traditional Chinese Medicine (TCM) offers valuable diagnostic and therapeutic approaches for HCC.
• Early detection of HCC is crucial for timely intervention and improved patient outcomes.

Purpose of the Study:

• To develop a computational model for inferring potential syndromes for HCC diagnosis using TCM.
• To enhance the accuracy of HCC diagnosis by identifying relevant clinical features and syndromes.
• To explore causal relationships between symptoms and syndromes in HCC using Bayesian networks.

Main Methods:

• A particle swarm optimization-based hierarchical feature selection (PSOHFS) model was proposed.
• A three-layer tree structure was employed for hierarchical feature representation, linking symptoms to syndromes.
• An improved PSO algorithm was utilized for optimal syndrome subset selection within a reduced feature space.
• Bayesian networks were applied to infer causal relationships at both symptom and syndrome levels.

Main Results:

• The PSOHFS model aggregated 147 symptoms into 27 groups, extracting 27 syndrome features.
• The approach successfully identified 24 key syndromes, significantly improving diagnostic accuracy for HCC.
• Bayesian networks effectively represented the causal pathways between symptoms and identified syndromes.

Conclusions:

• The developed computational model demonstrates significant potential in facilitating the clinical diagnosis of Hepatocellular Carcinoma (HCC).
• The PSOHFS model offers a novel approach to syndrome identification in TCM for complex diseases like HCC.
• Integrating computational methods with TCM principles can enhance diagnostic capabilities and patient care for HCC.