Mohammad Monir Uddin

Ph.D.

North South University

Associate Professor

NLST (Learn more about this study)

NLST-1270

Jun 17, 2024

AI-Enhanced Lung Cancer Risk Prediction and Nodule Detection

Lung cancer is a major cause of mortality worldwide, making early detection and accurate risk prediction crucial. Low-dose computed tomography (LDCT) has proven effective in reducing lung cancer mortality among high-risk individuals. Yet, challenges remain in generalizing trial results to broader populations and minimizing inter-grader variability. This research project integrates AI and deep learning techniques to enhance lung cancer screening and risk assessment by leveraging large datasets, including the National Lung Screening Trial (NLST).

The project aims to develop sophisticated AI models for detecting and segmenting lung nodules in CT images, assessing the risk of developing and dying from lung cancer, and improving predictive accuracy through probabilistic analysis. By employing convolutional neural networks (CNNs) and advanced image segmentation techniques, the project seeks to create reliable and automated methods for lung nodule detection and risk stratification, ultimately contributing to more personalized and effective lung cancer screening protocols.

Risk Prediction: Use existing models and NLST data to determine lung cancer development and mortality risk. Analyze risk differences among various populations.

AI Detection and Segmentation: Develop CNNs for automatic detection and segmentation of lung nodules in CT images. Implement deep learning techniques for accurate lung segmentation and cancer detection.

Advanced Data Analysis: Introduce new algorithms for early disease prediction using neural networks. Compare new models with existing ones to ensure accuracy.

Educational Implementation: Provide resources for learning medical image analysis techniques. Implement new detection methods to prepare students for careers in medical imaging.

Evaluation: Assess the specificity and accuracy of detection models using comprehensive datasets. Continuously refine models through iterative data analysis and validation.