PSO-Based Evolutionary Approach to Optimize Head and Neck Biomedical Image to Detect Mesothelioma Cancer.

Area of Science:

• Oncology
• Medical Imaging
• Computational Biology

Background:

• Mesothelioma is an aggressive, fatal cancer affecting internal organ linings.
• Current treatments are limited, and a cure remains elusive for most patients.
• Accurate detection of mesothelioma is crucial for patient outcomes.

Purpose of the Study:

• To optimize biomedical image analysis for improved mesothelioma cancer detection.
• To propose a novel feature selection approach for pathological images.
• To address challenges posed by limited medical sample sizes.

Main Methods:

• A Relief-PSO (Particle Swarm Optimization) feature selection approach was developed.
• The method combines the Relief technique for feature weighting with Hybrid Binary Particle Swarm Optimization (HBPSO).
• This approach reduces multilevel dimensionality in pathological image datasets.

Main Results:

• The proposed Relief-PSO method effectively screens morphological features.
• Significant dimensionality reduction was achieved in mesothelioma image analysis.
• The technique demonstrated superior classification performance compared to seven other algorithms.

Conclusions:

• The Relief-PSO approach offers an effective strategy for mesothelioma pathological image analysis.
• This method enhances feature selection, aiding in more accurate cancer detection.
• The study contributes to advancing computational methods in oncology research.