Artificial intelligence could play an expanding role across the lifecycle of oncology clinical trials, according to a narrative review published in ESMO Real World Data and Digital Oncology.
The authors describe how machine learning and other AI tools are increasingly being explored to improve trial design, patient recruitment, trial conduct, and data analysis.
One of the most promising applications is patient identification and recruitment. Machine learning models can analyse electronic health records, genomic data, and other clinical information to identify potentially eligible patients more efficiently, helping address one of the most persistent barriers to oncology trial enrolment.
CLINICAL SUMMARY
What was examined
A review explored potential applications of artificial intelligence across the lifecycle of oncology clinical trials.
Key findings
-
AI tools may support multiple stages of clinical trials, including trial design, patient recruitment, trial conduct, and data analysis.
-
Machine learning models can analyse electronic health records and genomic data to help identify patients eligible for clinical trials.
-
AI may also assist with automated data extraction, safety monitoring, and analysis of complex trial datasets such as multi-omics and imaging data.
Clinical implications
-
AI-enabled tools may improve efficiency and patient identification in oncology clinical trials.
-
Digital technologies could help address persistent barriers to trial recruitment and data management.
-
Further validation and regulatory oversight will be required before widespread implementation in clinical research settings.
AI may also support earlier stages of the trial process by analysing large datasets to identify therapeutic targets and generate hypotheses that inform trial design. During trial conduct, AI tools may support tasks such as automated data extraction, safety monitoring, and detection of adverse events.
The authors also highlight the potential role of AI in analysing complex datasets generated by modern oncology trials, including multi-omics data, imaging,g and real-world evidence.
Featured Event
However, they note that many applications remain at an early stage and will require further validation, regulatory oversight, ht and careful integration into clinical workflows.
The review suggests that while AI is unlikely to replace traditional clinical research methods, it could complement existing approaches and help improve the efficiency and quality of oncology trials.
Paper: Cerami, E. et al. AI for clinical trials in oncology. ESMO Real World Data and Digital Oncology, Volume 11, 100658. Access online here.
