Comprehensive review of reinforcement learning in lung cancer diagnosis and treatment: Taxonomy, challenges and...

Lung cancer (LuC) is one of the leading causes of death in the world, and due to the complex mechanisms and widespread metastasis, diagnosis and treatment are challenging. In recent years, the application of reinforcement learning (RL) techniques as a new tool to improve LuC diagnosis and treatment has been dramatically expanded. These techniques can potentially increase the accuracy of diagnosis and optimize treatment processes by learning from limited data and improving clinical decisions. However, RL in LuC diagnosis and treatment faces challenges such as limited access to clinical data, the complexity of algorithms, and the need for technical expertise for proper implementation. Our systematic review article aims to evaluate the latest developments in applications and challenges of using RL techniques in LuC diagnosis and treatment. The findings showed that RL has increased the accuracy of identifying disease trends by 37 % and enhancing treatment decisions by 23 %. Also, using this approach reduces data processing time by 17 % and streamlining treatment processes by 12 %. Ultimately, analyzing the current challenges and offering recommendations to researchers could help develop new strategies for improving the diagnosis and treatment of LuC.