Optimised CNN Framework with Clinical Decision Support for Enhanced Lung Cancer Screening and Early Diagnosis

Detection of lung cancer is still one of the key health problems in the world, which makes it important to adopt highly effective measures for early diagnosis. This paper aims to present the new Convolutional Neural Network model, carefully designed, for the highly accurate differentiation of lung cancer from the medical images. The network design was inspired by the outcomes obtained from the NLST and NELSON trials. Block-wise, each layer is fine-tuned to precisely perform feature extraction. The network employs the ReLU activation function and the Adam optimizer with binary cross-entropy loss. Therefore, the network reveals high accuracy in classifying lung cancer cases. In addition, the incorporation of the classification dense layers leads to better discrimination in the model. By strict testing, the CNN model manages to obtain a 98% accurate result and beats other algorithms. This remarkable line of thought not only makes possible the utilisation of artificial intelligence in medical imaging diagnostics, but additionally, has the potential to revolutionise lung cancer screening and early intervention.