Background
Clinical trials are essential in cancer research for introducing advanced therapies or devices into clinical practice, aiming to reduce cancer mortality and prolong patient survival. However, a significant challenge is the lack of access to screening and enrollment for many patients, and the failure of 20–40% of cancer clinical trials due to various reasons, including the lack of candidates. The development of health information technologies using AI, such as NLP and ML, offers a potential solution by automating the screening process, thereby enhancing efficiency and accuracy.
Methods
This study involved 1,053 inpatients diagnosed primarily with HCC at Nanfang Hospital, Southern Medical University, from January to December 2019. The CTMS was evaluated against two clinical trials for HCC, focusing on its ability to match patients based on eligibility criteria extracted from the trials. The system's performance was assessed in terms of accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).
Results
The CTMS demonstrated high specificity (99.0–99.1%) and good sensitivity (51.9–83.5%) in matching patients to clinical trials. It significantly reduced the time required for screening, from approximately 150 hours manually to just 2 hours using the CTMS. The system's accuracy ranged from 92.9% to 98.0%, with PPV and NPV values indicating reliable exclusion and consideration of patients for trials.
Discussion
The study highlights the potential of AI-based systems like CTMS in improving the efficiency and accuracy of clinical trial screening, particularly in the context of HCC. The system's ability to process and analyze large volumes of data quickly and accurately offers a promising solution to the challenges of manual screening. However, the study also acknowledges limitations, including the need for further optimization of the system's capabilities and the necessity of prospective studies to confirm its advantages in clinical practice.
Conclusion
The CTMS represents a significant advancement in the use of AI for clinical trial matching, offering a more efficient and accurate method for screening patients with HCC. Its integration with the Chinese EHR system and the use of advanced AI technologies like NLP and ML demonstrate the potential for such systems to enhance clinical research and patient care in oncology.
