Deployment of Lung Cancer Detection and CT scan in Real-Time for Clinical Use

Abstract

Lung cancer is one of the leading causes of death from cancer, and detection at an early stage would mean increased survival rates. Even though CT scans are very common in lung cancer screening, the procedure itself is time-consuming. It is susceptible to human error since a human being must interpret images. This paper discusses using a MATLAB-based CAD system in real-time lung cancer detection. The tool is designed to automatically identify the possible presence of lung nodules that may indicate malignancy. This research used a structured methodology involving image preprocessing, segmentation, feature extraction, and rule-based classification. Results in the system, with a synthetic dataset of 1000 CT images, achieved a detection sensitivity of 95% and a false-positive rate of 10%. The average processing time per image was 0.5 seconds. Results show promise for CAD systems to improve the efficiency of diagnosis. Further development is needed to reduce false positives sufficiently so they can be implemented into the clinical workflow.