Data Before Models: Building a Surgical Data Layer

Abstract

Every AI system in surgery, from imaging models to workflow agents, depends on structured, high-quality data. Yet most surgical environments operate with scattered photographs, fragmented electronic records, and free-text notes. This article defines how to build a Surgical Data Layer: a unified, auditable foundation that allows models to learn safely, perform accurately, and improve continuously.

Keywords

surgical data layer, data governance, de-identification, machine learning, data quality