Automated Early Detection of Potentially Malignant Lung Nodules in CT Scans Using Enhanced CNN Models

Lung cancer is still a serious health risk, especially in places like India where smoking is common and infections are common. Diagnostics are made more difficult by the fact that traditional imaging techniques, such as PET-CT scans, frequently fail to distinguish between benign and malignant nodules. While the mortality rate from lung cancer has decreased by $20 \%$ due to low-dose CT scans, accurate interpretation is still a challenge. With the help of the large LUNA16 CT image dataset, the U-Net deep learning algorithm is trained in this work with the goal of improving lung nodule detection. With a remarkable $\mathbf{9 9. 4 \%}$ accuracy rate, the U-Net model greatly increased the lung cancer screenings reliability. This development should lead to better survival rates, treatment results, and early diagnosis. The study highlights how using sophisticated computational techniques can improve lung cancer detection and diagnostic precision.