.png){kind=icon}
.png){kind=icon}
Return
AI-based augmentation of oncology clinical trials
.png){kind=icon}
PRE
AI
A
A
D
R
M
E
N
D
F
M
J
M
S
S
R
L
A
R
DOI：10.1038/s41571-026-01189-0.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Oncology clinical trials are often characterized by slow accrual, high failure rates and limited generalizability, reflecting both biological complexity and operational inefficiencies. Advances in artificial intelligence (AI) — enabled by large-scale electronic health record datasets and machine learning methods — offer new opportunities to address these challenges across the clinical trial lifecycle. In this Review, we discuss applications of AI across pre-trial design, trial conduct, and post-trial inference and generalization, highlighting how these tools can improve trial feasibility, support patient engagement and extend the relevance of trial findings. We also address cross-cutting challenges related to equity, data quality and drift, transparency, and regulatory oversight. The most immediate and evidence-supported role of AI in oncology trials lies in augmenting operational workflows under human oversight, particularly in the identification of candidate patients for enrollment, eligibility assessment, data extraction and trial monitoring (including remote patient and/or safety monitoring as well as monitoring of AI model performance and real-time trial data extraction) — applications that are now being implemented at select cancer centres. By contrast, AI applications designed to replace clinical evidence generation, such as synthetic control arms, outcome-prediction simulations and digital twins, remain at earlier stages of development, with limited prospective validation and unresolved methodological and regulatory challenges. Ultimately, we argue that achieving the potential of AI in oncology clinical trials will require rigorous prospective validation, harmonized regulatory standards, and coordination among clinicians, trialists, regulators, industry and patients. Oncology clinical trials are characterized by high failure rates — reflecting both the biological complexity of cancer and operational inefficiencies — and the generalizability of their results is often limited. In this Review, the authors discuss opportunities to address these limitations via application of artificial intelligence tools across the lifecycle of clinical trials as well as key challenges that must be overcome for routine clinical deployment of these tools.
Journal
IF:
82.2
Papers:
2.3K
Citations:
3.3W
Organization
Cited Papers