A new study published in Nature Communications suggests that combining artificial intelligence with clinician review may improve the accuracy of identifying patients eligible for oncology clinical trials, highlighting a potential role for AI-assisted workflows in trial recruitment.
Clinical trial participation remains low in oncology, and identifying potentially eligible patients from electronic health records (EHRs) is a major operational challenge. Eligibility criteria are often complex and require detailed manual review, making pre-screening a time-intensive process. Researchers investigated whether an artificial intelligence–assisted approach could improve pre-screening performance compared with clinician review alone.
CLINICAL SUMMARY
What was examined
A randomised evaluation study assessed whether clinician–AI collaboration improved the accuracy and efficiency of oncology clinical trial eligibility pre-screening using retrospective electronic health records.
Key findings
-
AI-assisted pre-screening improved the accuracy of eligibility assessments compared with clinician review alone.
-
Accuracy improvements were greatest for complex criteria such as biomarkers, staging, and prior therapies.
-
Review time was similar between clinician-only and clinician–AI approaches
Clinical implications
-
AI-assisted workflows may improve the identification of patients eligible for oncology clinical trials.
-
Improved pre-screening may support increased access to clinical trials
-
Human oversight remains important when using AI-based clinical tools
In this randomised evaluation study, clinicians screened electronic health records from patients with lung and colorectal cancers either independently or with support from an AI system designed to interpret trial eligibility criteria and extract relevant clinical information.
Clinician–AI collaboration improved the accuracy of eligibility determinations compared with clinicians working alone. Improvements were most apparent for complex clinical variables such as tumour staging, biomarker status,s and prior treatment history. However, AI assistance did not significantly reduce the time required for chart review, suggesting that efficiency gains may depend on further workflow integration.
Featured Event
The findings support a collaborative model in which artificial intelligence tools augment clinician decision-making rather than replace it. The authors noted that human oversight remains essential to minimise errors and ensure appropriate interpretation of clinical data.
If validated in prospective clinical settings, AI-assisted pre-screening could help address one of the major barriers to oncology clinical trial recruitment and improve patient access to investigational therapies.
Paper: Parikh, R.B., et al. Human-AI teaming to improve accuracy and efficiency of eligibility criteria prescreening for oncology trials: a randomized evaluation trial using retrospective electronic health records. Nat Commun (2026). Access online here.
