Triomics raised $15M for its oncology-focused generative AI solutions

Triomics has developed an AI-powered platform to help oncology healthcare professionals streamline workflows such as processing unstructured health record data at scale and pre-screening patients to establish eligibility for open clinical trials. These are processes currently being done manually in a time-consuming process certain to cause backlogs and delays that ultimately keep patients from being matched for clinical trials or biomarker-driven treatments, in addition to causing decreased quality and provider satisfaction. Founded by a former MIT biotech researcher and Adobe AI researcher, Sarim Khan (CEO) and Hrituraj Singh (CTO), Triomics recently raised $15 million from a selection of investors, including Lightspeed, Nexus Venture Partners, General Catalyst, and Y Combinator to refine its fine-tuned language models OncoLLM™ further and to develop additional use case-focused software.

Current research shows that OncoLLM can match 90% of patients with clinical trials within minutes, rather than the days or weeks it would take a qualified human team to review the same amount of health records. Additionally, OncoLLM can extract structured data points with equal or higher accuracy than commercial general-purpose solutions like Claude or GPT-4, costing 40x less. Regarding specialized oncology information retrieval, research shows that OncoLLM performs 1.5-2 times better than general-purpose, state-of-the-art alternatives. OncoLLM powers Trinomics specialist software solutions that connect to electronic health records (EHRs) to perform several tasks, such as Prism, which pre-screens patients for clinical trial matching, and Harmony, a solution supporting quality reporting, cohort analysis, and precision oncology.

Given the sensitivity of the use cases it handles, Triomics also works closely with leading research institutions including the Collaboration for Oncology-focused LLM Training (COLT), and the Cancer Informatics for Cancer Centers (CI4CC) Society to establish performance and safety benchmarks and best practices for generative AI in oncology. Another key differentiator, highlighted by Sarim Khan, is that Triomics has built its models from the ground up to tailor them for oncology-focused applications, rather than adapting existing solutions that may not scale as expected or bring a sufficient ROI. These features were also noted by Trinomics investors, who have great expectations for the positive impact Trinomics will have in the healthcare industry.