Auto Positioning

The Auto Positioning system aims to enhance CT examination setup by using a 2D/3D camera and Deep Learning algorithms to automate patient positioning, scout scan planning, and safety checks. This methodology streamlines workflow, improves consistency for optimal image quality, and incorporates automated checks for patient orientation and gantry collision avoidance, all validated through non-clinical bench testing.

The Auto Positioning device is an optional feature for GE CT systems designed to streamline and enhance the setup process for CT examinations up to the initiation of the first scout scan. Its primary purpose is to improve workflow efficiency, ensure consistent patient positioning for optimal image quality, and perform critical quality and safety checks. The system utilizes a ceiling-mounted 2D/3D video camera to acquire 3D spatial information of the patient on the CT table. This information, combined with data from the user-selected examination protocol and the CT system's precise gantry and table installation geometry, is processed by the Auto Positioning software. The software, incorporating Deep Learning neural networks (DNNs), automatically calculates and visually displays the optimal start and end locations for the scout scan, determines correct table height for patient centering, checks for proper patient orientation, and assesses potential collision risks between the patient and the gantry. Upon technologist acceptance, the system automatically moves the patient to the calculated scout start position. The technology provides an automated alternative to manual setup, though manual operation remains available. Non-clinical bench testing, which included evaluations of landmark location accuracy and scout scan positioning, successfully validated the system's performance and safety, meeting all predefined acceptance criteria without requiring clinical trials.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the Deep Learning model's performance vary across different patient morphologies and positioning scenarios (e.g., pediatric, bariatric, unconventional positioning)?; What is the quantifiable impact of Auto Positioning on overall CT examination time and technologist workload compared to traditional manual setup?; What specific calibration procedures are required for the 2D/3D camera and CT system geometry, and how frequently must they be performed to maintain accuracy?; How does the system handle edge cases or ambiguities in patient landmark identification, and what are the fail-safe mechanisms if automated detection is uncertain?; Are there any specific environmental factors (e.g., lighting conditions, room clutter) that significantly affect the 2D/3D camera's accuracy and the reliability of the Auto Positioning feature?