JMIR Publications

ABSTRACT

Background:

Lung cancer has a significant high incidence rate all over the world nowadays. The mortality rate of lung cancer continues to rise, which is more common in the middle-aged and the elderly and poses a great threat to human health.

Objective:

It is important that we timely assess the risk for the elderly and establish an efficient pathway for the risk transfer of lung cancer.

Methods:

We proposed deep reinforcement learning model based on DQN (Deep Q-learning) to explore the risk transfer pathway for lung cancer of middle-aged and elderly people. Risk stratification of lung cancer occurrence was deduced through deep neural network. DQN model was developed using HRS cohort for model’s training and internal validation. We also employed CHARLS cohort for model’s external validation. Transfer simulation of multiple pathways in different cycles was calculated in stratified risk groups leveraged DQN model.

Results:

We developed and evaluated the DQN method for optimizing the risk transfer pathway in the middle-aged and the elderly, with accuracy ranging from 0.917 (95% CI 0.896-0.928; P=.002) to 0.949(95% CI 0.909-0.961; P=.002) and area under curve ranging from 0.906(95% CI 0.887-0.933) and 0.927(95% CI 0.893-0.938). External validation was conducted to assess the model’s effectiveness and availability. The results showed that DQN models illuminated the optimal risk transfer pathways for stratified risk groups. Lung cancer incidence of the high risk group had declined by 68.24% through risk transfer, which had declined by 56.92% in the medium risk group. According to 30 years’ simulated intervention and risk transition, lung cancer incidences of high risk, medium risk and low risk were obviously decreased.

Conclusions:

DQN-based deep reinforcement learning model was proposed and validated to develop risk transfer pathway of lung cancer for middle-aged and elderly people. Risk stratification supplied effective foundation of lung cancer risk transition.