Over the past twenty years, the digitization of medical records has opened doors for automation and data-driven clinical support in various routine clinical applications. We now find ourselves in the era of artificial intelligence (AI), which is revolutionizing many aspects of life. Despite this, AI-based tools have not yet been widely integrated into clinical trials. This is mainly due to the complexity of these studies and the challenges often encountered in patient recruitment, data collection and management.
The integration of AI into clinical research and oncology has the potential to transform the traditional approach to clinical trials, leading to improvements in efficiency, accuracy, and personalization of treatments. At MEDSIR, a leading company dedicated to independent oncological clinical research, we are committed to this transformation, with the ultimate aim of bringing more precise and effective cancer treatments to patients.
In keeping with this commitment , MEDSIR attended, for the first time, the 22nd International Conference on Artificial Intelligence in Medicine (AIME 2024) held in Salt Lake City, Utah, from July 9th to 12th. This conference was an incredible opportunity to hear the latest research and discuss the key topics in AI, such as predictive modeling, disease risk prediction, wearable devices, medical imaging analyses, and more. Additionally, it was a fantastic platform for MEDSIR to share with the scientific community our first experience with AI applied in clinical research, the MIRROR project.
MIRROR project: Validation of an Artificial Intelligence System in the extraction of data from Electronic Health Records (EHRs) compared to data obtained through traditional, manual data capture in Clinical Trials in Oncology.
The MIRROR project is a collaborative effort between MEDSIR, Hospital Universitario Virgen del Rocío, and Science4Tech. This study is a data-driven, retrospective, longitudinal, and observational study trial that compares AI-based data capture from electronic health records (EHRs) with clinical data obtained through manual methods. EHRs are digital versions of a patient's medical records. These real-time patient records make information available instantly and securely to authorized users. This study used secondary data from the electronic Case Report Forms 113 adult patients with early or advanced breast cancer who enrolled into clinical trials conducted at Virgen del Rocío University Hospital (Seville, Spain) between January 2012 and December 2021.
The results showed that this AI-based method had an effectiveness of ≥80% for most clinical variables after comparison with clinical data manually extracted. Clinical information from data with standardized format (e.g., birth and death dates, gender, or laboratory tests) were easier to extract by AI than clinical information from data with non-standardized format (e.g., clinical notes). Based on these results, the MIRROR study highlights the potential of this AI-based tool to automate the data extraction process and introduce its use in clinical trials.
Daniel Alcalá López, PhD. presenting the MIRROR project at AIME 2024
We are extremely proud to have been able to participate in AIME and we would like to especially thank Daniel Alcalá-López, PhD. Data Science Expert at MEDSIR, for presenting this study within the scientific community. We cannot wait to see you all at AIME again next year!
Contact us to learn more about the streamlined strategy in clinical development at MEDSIR.
Comments