What You Should Know:
– Mount Sinai Tisch Cancer Center has deployed PRISM, an AI-powered platform designed by Triomics for systemwide clinical trial matching.
– Utilizing OncoLLM—a large language model pipeline built specifically for oncology—PRISM identifies eligible patients across the entire Mount Sinai Health System (including Queens and Brooklyn) in real-time.
– The strategic initiative aims to accelerate research enrollment and provide equitable access to cutting-edge treatments by replacing manual, time-intensive chart reviews with automated, oncology-specific intelligence.
Solving the “Fragmented Data” Dilemma in Oncology
Clinical trial matching has traditionally been a bottleneck in cancer research. It relies on a manual, time-intensive review of lengthy trial protocols against fragmented medical records, often leading to delays or missed opportunities.
PRISM (Powered by OncoLLM) addresses this by:
- Oncology-Specific Reasoning: Unlike generic LLMs, OncoLLM is trained to understand the nuances of cancer care, from specific biomarkers to complex treatment histories.
- Automated Eligibility Screening: The platform analyzes electronic health records (EHRs) to identify potentially eligible patients earlier and more consistently than manual review.
- Equitable Access: By deploying the tool across the entire Mount Sinai Health System, the initiative democratizes access to cutting-edge treatments for diverse patient populations across NYC.
Integration with Epic and Oracle Health (Cerner)
Unlike generic LLMs, Triomics’ OncoLLM is built to sit natively within the EHR’s clinical data warehouse.
- HL7 & FHIR Standards: PRISM utilizes Fast Healthcare Interoperability Resources (FHIR) to pull real-time patient demographics, pathology reports, and biomarker data directly from Epic or Oracle Health.
- Unstructured Data Parsing: The “magic” of OncoLLM is its ability to read the “notes” section—where 80% of clinical trial inclusion/exclusion criteria often hide—and convert that text into structured data points.
The market is currently flooded with “AI wrappers,” but PRISM represents a shift toward domain-specific intelligence. In 2026, general-purpose LLMs are often flagged for hallucinations in high-stakes clinical settings. Triomics’ focus on an oncology-specific pipeline (OncoLLM) reflects the 12% of hospitals that have moved toward rigorous, specialized AI governance to ensure patient safety and data accuracy.
“Clinical trials are essential to advancing cancer care, but too often patients and their treating physicians are not aware of studies that may be appropriate for them,” said Karyn Goodman, MD, MS, Professor and Vice Chair of Clinical Research in the Department of Radiation Oncology at the Icahn School of Medicine at Mount Sinai and Associate Director of Clinical Research at TCC. “By deploying an AI platform trained specifically for oncology, we can identify trial opportunities earlier, more consistently, and more equitably, allowing clinicians to focus on meaningful conversations with patients rather than manual chart review.”
