Clinical Scorecard: Artificial Intelligence in Interventional Trials Registered at ClinicalTrials.gov
At a Glance
Category
Detail
Condition
Planned use of artificial intelligence in interventional clinical trials
Key Mechanisms
AI applications included disease screening and predictive modeling, patient education, remote care, and intervention planning; images and video were the most common data modality.
Target Population
ClinicalTrials.gov interventional trial records with actual or estimated start dates on or after January 1, 2022, that described AI involvement; the unit of analysis was the trial record.
Care Setting
ClinicalTrials.gov registry analysis of planned interventional trials across clinical specialties; the study assessed registry descriptions, not confirmed AI use in practice.
Key Highlights
The final cross-sectional analysis included 1,212 interventional trials after screening registry records and excluding trials without confirmed AI involvement.
Generative AI was planned in 166 trials (13.7%); none of the trials initiated in 2022 planned its use, compared with 100 of 491 trials (20.4%) initiated in 2025.
Cancer was the leading clinical condition (20.0%), followed by neurological and mental health conditions (19.6%) and cardiovascular and metabolic disease (18.6%).
Screening and predictive modeling for disease risk or treatment selection were the most frequent AI applications (39.8%); images and video were the most common data modality (29.2%).
All records described AI or machine learning, but information about specific methods was frequently missing; reviewers relied on broad terms and free-text registry entries.
Guideline-Based Recommendations
Diagnosis
Not applicable: this study analyzed registry records and did not evaluate diagnostic accuracy or establish clinical diagnostic guidance.
Management
The authors state that structured ClinicalTrials.gov fields consistent with AI reporting standards would improve transparency and description of AI methods.
Monitoring & Follow-up
The authors state that structured registry fields aligned with AI reporting standards would support clearer description and understanding of AI methods.
Risks
Registry records describe intended rather than confirmed trial conduct, so the analysis could not determine whether AI was used as described.
The reported counts do not indicate what fraction of all registered trials involved AI.
Patient & Prescribing Data
No patient-level population was analyzed; the unit of analysis was the ClinicalTrials.gov interventional trial record.
The study did not assess treatment efficacy, safety, or prescribing. AI was planned for varied trial purposes, including screening, predictive modeling, patient education, remote care, and intervention planning.
Clinical Best Practices
The authors state that transparent documentation of AI methods in registry records is essential; specific methods were frequently missing from free-text descriptions.
Interpret registry findings as planned AI use, not confirmation that AI was implemented during trial conduct.
