Sample IRB language for Sybil

Sybil is a validated machine learning model that calculates an individual’s future risk of lung cancer based on axial images from a single chest computed tomography (CT) scan (Mikhael et al., JCO 2023, PMID 36634294). Specifically, Sybil analyzes Digital Imaging and Communications in Medicine (DICOM) images (input) and produces six numbers corresponding to the annual cumulative risk of lung cancer for the following six years (output). Sybil runs entirely within an institution’s firewall, so no personal health information is exported outside the institution’s secure server. Sybil’s code is downloaded onto the secure institutional server; once installed, it runs locally and does not require an internet connection. Sybil is not a commercial product; it is an open source model that is available for research purposes. This work is made possible thanks to the support of AstraZeneca.