Opportunistic AI - detection of spinal anomalies indicative of osteoporosis on routine imaging

Principal Investigator

Name
Ronen Gordon

Degrees
MBBS BSc

Institution
Zebra Medical Vision Ltd.

Position Title
Clinical Director

Email
ronen@zebra-med.com

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-834

Initial CDAS Request Approval
Sep 14, 2021

Title
Opportunistic AI - detection of spinal anomalies indicative of osteoporosis on routine imaging

Summary
The project aims to test an automated AI-driven software device, designed to automatically segment and each vertebra and provide key measurements related to bone health including:
1) Vertebral height loss indicative of vertebral compression fracture
2) Vertebral trabecular bone attenuation, indicative of potential osteoporosis if low.
Such software has the potential to radically change the way that we approach primary prevention of major osteoporotic fractures but identifying patients at risk of future fracture from CT imaging already being performed for other clinical indications e.g. lung cancer screening. The epidemiological correlation between incidence of osteoporosis and COPD/Lung Cancer means the NLST dataset presents in ideal opportunity to test the performance and potential value of this AI-driven automated spinal measurement tool.

Aims

To test an AI-driven software device's ability to:
- Accurately segment vertebra on CT scans
- Accurately measure vertebra height loss
- Accurately measure vertebra trabecular bone attention, akin to bone mineral density
- Quantify the number of patients with a vertebra fracture and/or low bone density that could be detected via the NLST program and use existing literature to model the potential reduction in future fracture incidence if appropriate patients received bone strengthening agents.

Collaborators

N/a