Every year, the American Society of Clinical Oncology (ASCO) receives thousands of clinical abstracts with promising oncology insights. The problem? Those insights are often buried within mountains of unstructured data, leading to a time-consuming and costly manual process. This also delays decision-making, which then delays strategic planning and resource allocation.
To put it plainly: It’s time we streamlined scientific literature review (SLR) – and that’s precisely what IMO Health set out to do.
Four internal artificial intelligence (AI) experts – Kyeryoung Lee, PhD, Hunki Paek, PhD, Nneka Ofoegbu, and Xiaoyan Wang PhD, FAMIA – authored a paper about ASCOmind, an AI agent developed by IMO Health that can reduce large-scale manual SLR workloads in oncology and accelerate real-world evidence synthesis. Lee presented these findings at ASCO 2025.
Read on for more details about the study and how similar models could transform the life sciences space, generating faster, more reliable insights – without draining budgets.
Reinventing scientific literature review
Revealed at the ASCO 2025 Annual Meeting, ASCOmind is a GPT-4o-based platform powered by six autonomous and collaborative AI agents:
- Pre-processor
- Categorizer
- Metadata Extractor
- Analyzer
- Visualizer
- Protocol Maker
Together, these agents can automate key phases of the SLR process, including data extraction, evidence synthesis, and visualization – turning dense, free-form clinical abstracts into structured insights for more informed decision-making, planning, and resource management.
Accelerating oncology insights
IMO Health scientists tested the agent’s efficacy by applying it to dozens of ASCO 2024 myeloma studies, homing in on elements like data accuracy, visualized charts, and more.
Overall, ASCOmind delivered measurable improvements:
- Categorized 60 abstracts into 26 clinical trials, 34 as real-world studies
- Processed 51 predefined data elements per abstract in under 5 minutes vs. 60+ minutes manually
- Generated visual summaries and analysis within 10 minutes
This level of automation can save organizations valuable time and resources, reducing overhead and increasing ROI.
Boosting the accuracy and scalability of scientific evidence synthesis
ASCOmind was evaluated against human reviewers and proved to be highly accurate, generating only one misclassification that was quickly fixed. These results confirm ASCOmind’s ability to:
- Significantly reduce the time and cost associated with manual review of clinical abstracts
- Enhance rapid decision-making through real-time analysis and data visualization, especially critical in the oncology field
- Accelerate clinical insights generation from extensive data sets, streamlining strategic planning and resource allocation
ASCOmind’s demonstrated accuracy and speed highlight its potential to streamline oncology research processes, benefiting clinical researchers and healthcare providers by enabling faster, data-driven decision-making.
Furthermore, while IMO Health’s study focused on multiple myeloma, the implications of such an AI agent are much broader. By cutting tedious hours of manual work down to mere minutes, and delivering real-time insights, this tool will not only benefit researchers and scientists, but those who matter most – patients.
Cut review time. Reduce costs. Maximize insights.
