Triomics, a startup building an oncology-specific data pipeline for clinical workflows, closed a $22 million Series B led by Battery Ventures. Existing backers Nexus Venture Partners, Lightspeed, and Y Combinator joined the round. The funding arrives as cancer centers struggle with a data volume problem: longer patient survival means medical records balloon into thousands of pages—physician notes, imaging reports, fax scans—that staff must manually parse for trial matching, appointment prep, and regulatory reporting.

Founded in 2021, Triomics initially focused on clinical trial matching. As LLM capabilities matured, the team expanded into verifiable patient summaries that surface key data points inside existing EHR tools, eliminating context-switching. The architecture is purpose-built for oncology: generic AI scribes (e.g., Abridge, Nuance) handle ambient documentation, but Triomics trains its models on oncology-specific data to handle the domain’s complexity—multi-year histories, tumor registry submissions, and structured extraction from semi-structured records.

The engineering trade-off is clear: narrow specialization vs. horizontal generality. Triomics bets that oncology’s data density demands a tailored extraction layer rather than a one-size-fits-all summarizer. Early results support the bet—enterprise customers grew 4x year-over-year, with ARR up 10x. Institutions like Memorial Sloan Kettering and Yale Cancer Center are production users. For data engineers, the takeaway is that domain-tuned pipelines, not generic agents, win when the schema is messy and the stakes are high.

Source: TechCrunch