Stanford Health Care, based in Palo Alto, Calif., is testing a new tool from Microsoft designed to ease the burden of preparing complex cancer cases for tumor board meetings.
The technology — a healthcare agent orchestrator — is being piloted by Stanford data scientists and developers to create autonomous AI agents that compile and synthesize key clinical data, according to a May 19 news release from Microsoft.
Tumor boards bring together a multidisciplinary team — often including oncologists, radiologists, pathologists and surgeons — to weigh treatment plans for high-risk cancer patients. Preparation involves pulling information from disparate sources, such as EHRs, imaging systems and research databases. With Microsoft’s tool, AI agents can complete many of those tasks in minutes, including creating chronological timelines, synthesizing literature, identifying clinical trials and assembling presentation reports, the company said.
The agents integrate with Microsoft 365 Copilot, allowing clinicians to interact via natural language prompts in apps like Microsoft Teams and Word.
One agent aggregates clinical notes, lab results, medications and genomic data into structured summaries with source citations. Others can read radiology scans or pathology slides, search medical literature for relevant research, or identify eligible clinical trials. A final report creation agent compiles the information into a shareable Word or PowerPoint document.
Timothy Keyes, MD, PhD, a data scientist at Stanford Health Care, emphasized in the release that the AI is not intended to replace medical decision-making.
Stanford Health Care is still testing the tool in a research setting and has not yet deployed it for real-time clinical use.
