Research on Risk Transfer Pathways for Lung Cancer Among Middle-Aged and Older Individuals Using Deep Reinforcement...

Area of Science:

• Artificial Intelligence
• Machine Learning
• Public Health

Background:

• Lung cancer poses a significant global health threat, particularly impacting middle-aged and older populations.
• Rising mortality rates underscore the urgent need for effective risk management strategies.

Purpose of the Study:

• To assess lung cancer risk in middle-aged and older individuals.
• To establish an efficient pathway for lung cancer risk transfer using advanced modeling.

Main Methods:

• A deep reinforcement learning model, specifically a deep Q-network (DQN), was developed for lung cancer risk stratification and pathway exploration.
• The model utilized the Health and Retirement Study and China Health and Retirement Longitudinal Study cohorts for training, internal validation, and external validation.
• Risk stratification was achieved through deep neural networks, enabling stratified risk group-leveraged DQN modeling.

Main Results:

• The DQN model demonstrated high accuracy (0.917-0.949) and area under the curve (0.906-0.927) in optimizing risk transfer pathways.
• External validation confirmed the model's effectiveness and availability across different cohorts.
• Simulated risk transfer significantly reduced lung cancer incidence: 68.2% in the high-risk group and 56.9% in the medium-risk group.

Conclusions:

• A validated DQN-based deep reinforcement learning model effectively simulates lung cancer risk transfer pathways in middle-aged and older adults.
• Risk stratification is crucial for enabling effective lung cancer risk transition and reduction.