A Novel Concept of IOT Healthcare Monitoring Framework with Hybrid Optimization Techniques for Early Diagnosis of Cancer

Lung cancer is one of the deadliest diseases worldwide. Early detection is crucial for improving survival rates. Lung nodules are an important sign for early diagnosis. They help reduce the workload for radiologists and improve diagnostic accuracy. Artificial intelligence-based neural networks have shown promise in automating the detection of lung nodules by utilizing patient data collected from Internet-of-Things (IoT)-enabled healthcare systems. However, traditional models often depend on manually created features, which limits their diagnostic performance. In this study, we propose a new IoT-enabled healthcare monitoring framework that includes a Multi-Strategy Improved Grey Wolf Optimization (MSI-GWO) algorithm and a Capsule Network ELM: Extreme Learning Machine (CAPSNET-ELM) hybrid model for early lung cancer detection. To address the challenge of high-dimensional medical data, we employ an Artificial Bee Colony-Harris Hawks Optimization (ABC-HHO) algorithm for optimal feature selection. This improves the model's sensitivity and precision. We implement the proposed system using NS-2 for network simulation and TensorFlow-based Python libraries for training and deploying the model. The optimized CAPSNET-ELM hybrid classifier is trained on features extracted from the IoT platform and is evaluated against leading lung cancer detection models. We securely store diagnostic results in the cloud for clinical review, showing better performance in terms of accuracy, sensitivity, specificity, and precision.