Metaheuristic-Driven Optimization of Deep Learning Models for Lung Cancer Detection in CT Imaging: Focus on Whale Optimization Algorithm and Its Variants – A Comprehensive ۲۰۲۵ Review

Lung cancer diagnosis via computed tomography (CT) imaging demands high-precision deep learning models to handle the intricacies of nodule detection, segmentation, and classification. Metaheuristic algorithms, particularly the Whale Optimization Algorithm (WOA) and its enhanced variants, have gained prominence for optimizing hyper-parameters in convolutional neural networks (CNNs) and graph neural networks (GNNs), addressing issues like local optima trapping and computational inefficiency. This comprehensive ۲۰۲۵ review synthesizes ۹۸ studies from ۲۰۱۶ to ۲۰۲۵, emphasizing WOA's bio-inspired mechanisms, hybrid integrations, and performance in lung cancer tasks. We delineate popular metaheuristics in healthcare, WOA principles, CNN/GNN synergies, and comparative benchmarks, while pinpointing limitations in convergence, scalability, and clinical validation. Recommendations for hybrid optimizers and federated learning integrations are proposed to advance real-world deployability.

کلیدواژه ها:

Whale Optimization Algorithm ، Metaheuristics ، Lung Cancer Detection ، CT Imaging ، Deep Learning Optimization ، CNN ، GNN ، Hyper-parameter Tuning