This paper presents the first comprehensive review of Cost-Sensitive Learning (CSL) for imbalanced medical data. A total of 173 papers published between 2010 and 2022 were analyzed, covering topics such as publication years, research types, medical disciplines, medical tasks, CSL approaches, strengths and weaknesses of CSL, frequently used datasets, evaluation metrics, and development tools. The results show a significant increase in publications since 2020, with a strong preference for CSL direct approaches. Medical images are the most common data type, while Python is the primary programming tool. The strengths and weaknesses of CSL were analyzed in three aspects: CSL strategy, CSL approaches, and relevant works. The study highlights the underutilization of cost-related metrics and the prevalence of Python as the primary programming tool. The paper also discusses the challenges of class imbalance in medical data, including biased learning, the significance of minority class instances, and the natural occurrence of imbalanced datasets. Various strategies to mitigate class imbalance were reviewed, including data-level, algorithm-level, cost-sensitive, and ensemble-based strategies. CSL is presented as a promising approach that preserves data distribution while ensuring computational efficiency. The paper provides an illustrative example of CSL applied to cervical cancer diagnosis, demonstrating its effectiveness in improving diagnostic accuracy. The study concludes that CSL is a valuable tool for addressing class imbalance in medical data, but further research is needed to validate its practical value in real-world medical settings. The paper also highlights the need for more synthesis and critical evaluation of existing CSL methods in medicine. The findings suggest a promising outlook for the future of CSL for medical data but also underscore the need for continued validation and rigorous evaluation of the developed techniques.This paper presents the first comprehensive review of Cost-Sensitive Learning (CSL) for imbalanced medical data. A total of 173 papers published between 2010 and 2022 were analyzed, covering topics such as publication years, research types, medical disciplines, medical tasks, CSL approaches, strengths and weaknesses of CSL, frequently used datasets, evaluation metrics, and development tools. The results show a significant increase in publications since 2020, with a strong preference for CSL direct approaches. Medical images are the most common data type, while Python is the primary programming tool. The strengths and weaknesses of CSL were analyzed in three aspects: CSL strategy, CSL approaches, and relevant works. The study highlights the underutilization of cost-related metrics and the prevalence of Python as the primary programming tool. The paper also discusses the challenges of class imbalance in medical data, including biased learning, the significance of minority class instances, and the natural occurrence of imbalanced datasets. Various strategies to mitigate class imbalance were reviewed, including data-level, algorithm-level, cost-sensitive, and ensemble-based strategies. CSL is presented as a promising approach that preserves data distribution while ensuring computational efficiency. The paper provides an illustrative example of CSL applied to cervical cancer diagnosis, demonstrating its effectiveness in improving diagnostic accuracy. The study concludes that CSL is a valuable tool for addressing class imbalance in medical data, but further research is needed to validate its practical value in real-world medical settings. The paper also highlights the need for more synthesis and critical evaluation of existing CSL methods in medicine. The findings suggest a promising outlook for the future of CSL for medical data but also underscore the need for continued validation and rigorous evaluation of the developed techniques.