Triomics, a San Francisco-based healthcare startup, has raised $15 million to streamline workflows for cancer centres using generative artificial intelligence (GenAI). Their technology focuses on automating tasks currently handled manually, aiming to improve efficiency and patient care.
The funding was led by leading Silicon Valley investors, including Y Combinator, Lightspeed, General Catalyst, and Nexus Venture Partners. Notably, whenever possible, Triomics seeks to pass costs directly to pharmaceutical sponsors, minimising the financial burden on healthcare providers. Additionally, performance-based agreements ensure partners see a return on investment before incurring any out-of-pocket expenses.
In the relative note, we also reported yesterday about a female-led Opmed.ai that secured $15M to address healthcare labour shortages with AI.
What problem is the startup solving
Co-founded by Sarim Khan (CEO) and Hrituraj Singh (CTO), Triomics focuses on a critical challenge in healthcare, particularly within oncology: the vast amount of unstructured data locked away in free-text patient records. Traditionally, accessing this data relies on manual chart review, a laborious process that bottlenecks crucial tasks like clinical research and patient care.
Triomics’ solution is a powerful combination – an oncology-focused large language model (OncoLLM) paired with custom software designed for specific use cases. This allows for automated data processing at scale, significantly reducing processing times. Their software suite includes Harmony, which facilitates EHR data curation for research and registry needs, and Prism, which streamlines patient-trial matching. Triomics also fosters continuous collaboration with partners to develop a future-proof product roadmap.
Who are the company’s users and beneficiaries
Triomics’ primary target users are hospital staff directly involved in oncology research and care, including clinical research nurses, coordinators, registrars, and others focused on improving healthcare delivery within their institutions. They partner with leading academic medical centres, health systems, and cancer centres.
The benefits of Triomics’ technology extend to various stakeholders. Clinical research coordinators can expedite patient-trial matching, leading to faster enrollment in potentially life-saving trials. Patients benefit from quicker access to these trials, while pharmaceutical companies can complete study enrollment more efficiently, accelerating the development and market launch of new drugs. Additionally, Triomics’ solutions streamline EHR-to-EDC integration (connecting electronic health records to clinical trial data capture systems) and empower registrars to efficiently abstract data for tumour registries.
Challenges in oncology data management
Currently, oncology staff face a significant burden in manually processing patient data. Extracting insights from free-text health records, which make up roughly 80% of medical data, is a time-consuming process. This workload can lead to delays in crucial tasks like matching patients to clinical trials and can contribute to provider burnout.
Triomics’ approach: OncoLLM and specialised software
Triomics co-founders Sarim Khan (CEO) and Hrituraj Singh (CTO) recognized the potential of generative AI to address these challenges. They developed OncoLLM, a generative AI model specifically trained on oncology data. This model, unlike generic AI models, is designed to understand the nuances of cancer care terminology and improve the accuracy of tasks like clinical trial matching and data extraction from progress notes.
While the inner workings of generative AI models can be complex, Triomics sheds light on how OncoLLM achieves its results. The company emphasises that it prioritises explainability – the ability to understand the rationale behind the model’s outputs. This is particularly important in oncology, where decisions can have significant consequences for patient care.
Triomics achieves explainability by building interpretable models and employing techniques to identify the data points driving the model’s conclusions. This allows clinicians to understand how OncoLLM arrives at its recommendations and fosters trust in the technology.
What we think about the startup
Triomics differentiates itself from foundational model companies like OpenAI, Anthropic, Google, or Microsoft by specialising in oncology and developing software tailored for complex use cases and specific end-users within healthcare. While competitors exist in areas like patient matching, they often rely on legacy technologies that lack the scalability and cost-effectiveness Triomics offers.
Triomics’ innovative approach to automating oncology workflows with generative AI holds significant promise for the future of cancer care. Their focus on explainability, user-centric design, and collaboration positions them well to navigate the complexities of implementing AI in this critical domain.
