Triomics builds an AI-powered platform for oncologists and administrative staff to automate data-heavy workflows. Longer patient survival has increased the volume and complexity of medical records, often reaching thousands of pages with notes, imaging and pathology reports, and scanned faxes. The platform initially focused on matching patients to suitable clinical trials and later expanded as large language model capabilities improved. It adds verifiable patient summaries that surface key information inside existing clinician tools, reducing the time spent on appointment preparation. The platform also automates submitting tumor reports to government registries, a legal requirement for cancer centers. Models trained specifically on oncology data are used by major cancer institutions.
"Triomics, a startup building an AI-powered platform to help oncologists and administrative staff automate data-heavy tasks like clinical trial matching and appointment prep, has raised $22 million in Series B funding."
"A typical medical chart includes physician progress notes, imaging and pathology reports, and even scans of faxes. "We have seen medical records [with] thousands of pages of information," Triomics co-founder Sarim Khan (pictured left) told TechCrunch."
"Initially focused on helping doctors identify the most suitable clinical trials for their patients, Triomics expanded its platform as LLM capabilities grew. Over the last couple of years, Triomics added verifiable patient summaries to its platform, surfacing key information directly inside the tools clinicians already use, without requiring them to switch applications."
"Triomics is also used to automate the tedious task of submitting tumor reports to government registries, a legal mandate for cancer centers. While generic AI agents excel at basic summaries, prominent institutions like Memorial Sloan Kettering (MSK) and Yale Cancer Center use Triomics because its models are trained specifically on oncology data, Khan explained."
Read at TechCrunch
Unable to calculate read time
Collection
[
|
...
]