Innovative Medical AI & FHIR-Native Cancer Registry Technology.
The AI layer that turns pathology, imaging, and clinical feeds into the values your registry actually needs.
NeuralFrame was founded in 2017 to rethink how cancer data moves through hospital systems. From day one, we built around a specific problem: the data a cancer registry needs isn’t sitting in one place, in one format, on one document. It’s spread across pathology reports, addendums, imaging studies, clinical notes, and treatment records — each written at different times, by different people, for different purposes. Legacy registry software asks registrars to reconcile all of that by hand. KACI does the reconciliation first, and asks registrars to validate the work.
The engine behind this is what we call the Cancer Episode Enginer — KACI’s AI layer that combines findings across specimens, addendums, and treatment timing to produce NAACCR-compliant values rather than raw extracted text. A single biomarker value that requires reading five pathology documents in sequence, understanding which one supersedes which, and applying NAACCR timing rules — that’s the kind of work KACI does before a registrar ever sees the case. The registrar validates, corrects, and signs off. Every value is logged with its source document, its reasoning, and the registrar who approved it.
The impact shows up in registrar productivity. When we analyze usage data from our industry-leading cancer registry product, KACI. The time savings on case abstraction are substantial — not because registrars are doing less careful work, but because they’re no longer starting from a blank case. KACI’s automatic abstraction frees Oncology Data Specialists to focus on complex analysis and reporting instead of hunting through synoptic reports and unstructured medical documents:
