Improving lung cancer diagnoses: a machine learning approach for detection and prediction in CT-Scan image analysis

The purpose of this work is to meet the current challenge of early lung cancer detection, which is defined as the uncontrollable growth of cells in lung tissues. In particular, it focuses on the improvement of cancer cell identification and forecasting in the lungs that calls for early diagnostics and treatments. The first aim of the study is to assess the performance of DL algorithms in diagnosing lymph node stages from histopathological images, especially through convolution Neural Networks (CNN). To meet these objectives, the research considers Deep Learning approaches the CNN. The methodology suggested for the research includes the analysis of the CT scan images and histopathological images. To compare the proposed approach with other algorithms, this paper uses recognition accuracy, precision, recall, F-Score, and Area Under the Curve (AUC). The CNN model produced remarkable results with an accuracy of 96%, precision of 94%, recall of 90%, AUC of 96. The precision achieved was 81%, Recall was 73% and the F1-Score was 76%. From these metrics, it is clear that the proposed approach achieves a considerable enhancement in the identification and categorization of lung cancer as compared to the conventional methods particularly in the case of histopathological images. It can be concluded that CNN and other DL approaches have the ability to enhance the LNI diagnosis in lung cancer based on the histopathological image analysis. High diagnostic accuracy is confirmed, which underlines future studies and practice of DL methods in lung cancer detection and treatment for early and correct diagnostics and better patients’ prognosis.