Chest CT Segmentation - Case study

The challenge

• An industry-leading medical device company required the capability to automatically generate an accurate 3D model of complex patient-specific anatomies
• The anatomical model was needed for planning challenging surgical navigation procedures
• The pre-existing solution was slow and not robust, requiring extensive and time-consuming manual editing by the interventionist

Our approach

• The RSIP Vision team leveraged its deep-learning based 3D CT auto-segmentation technology to deliver a customized, fast, robust, production quality solution

The outcome

• Our core module has been integrated into our clients platform and is in clinical use
• Used in 1000s of cases in hospitals worldwide
• >99% segmentation success rate
• 1mm-level segmentation resolution
• 30 seconds running time on full-sized chest CT scan