The rising complexity of cancer care has created significant challenges for oncologists, who must manage numerous cancer subtypes and keep pace with constantly updated clinical guidelines from multiple organizations. Frequent changes and lack of standardization across guidelines increase the risk of inconsistent patient care, while shortages of oncology specialists mean general practitioners often handle cases without full expertise or time to stay current.
To address these issues, the University of California at San Francisco partnered with Color to develop an AI-driven decision support system. This technology automates the aggregation and structuring of patient data from electronic health records and integrates the latest national and local clinical guidelines. The system flags missing diagnostic information, provides clear, evidence-based recommendations tailored to each patient, and explains the reasoning behind its suggestions, allowing clinicians to see exactly how each decision is made.
In practical use, the AI system streamlines the oncology workflow by:
– Compiling and organizing all relevant patient information before consultations
– Prompting for necessary tests if data is missing
– Delivering up-to-date treatment recommendations based on both national and institution-specific protocols
– Continuously updating itself with the latest research and clinical guideline changes
A study of 100 breast and colon cancer cases showed that the AI reduced the time needed to review records and guidelines from up to two hours to just 10-15 minutes per case. The system’s recommendations matched those of oncologists 95% of the time, and its early identification of missing workups helped reduce delays in starting treatment. This led to improved efficiency, more standardized care, and faster patient progression to therapy—critical factors in cancer outcomes.
For healthcare organizations considering similar solutions, success depends on ensuring the AI system has access to complete and accurate patient data through robust integration with electronic health records. Maintaining transparency in how the AI generates its recommendations is also vital, as this allows clinicians to trust, interpret, and, when necessary, override automated suggestions. By supporting rather than replacing clinical judgment, AI can make high-quality, evidence-based oncology care more accessible and efficient, especially where specialist resources are limited.
