CT CoPilot

CT CoPilot™ is a software application designed to automate the post-acquisition analysis of brain CT images, providing quantitative volumetric measurements, visualizations, and longitudinal comparison of anatomical structures. Its methodology involves proprietary algorithms for image segmentation, registration to a neuroanatomic atlas and prior patient scans, and the calculation of subtraction maps, with performance validated through laboratory testing demonstrating high accuracy in registration and segmentation for patients aged 18 and older.

CT CoPilot™ is a software tool designed for the automated analysis of CT brain images in patients aged 18 and older. Its primary purpose is to automate the traditionally manual tasks of identifying, labeling, visualizing, and quantifying the volume of segmentable brain structures by processing sets of CT images and providing automated registration and reformatting of the imaging data for use by trained personnel in Neuro CT imaging, though it is not intended for sole use in patient management decisions. The CT CoPilot™ system performs post-acquisition quantitative analysis by automatically reformatting images, labeling structures with color overlays, and calculating their numerical volumes, with outputs provided in standard DICOM format. For longitudinal studies, it co-registers current CT data to the patient's most recent prior processed exam, facilitating comparison through generation of voxel-by-voxel subtraction maps of Hounsfield Unit changes; this process relies on registration, alignment, and segmentation algorithms benchmarked against a reference neuroanatomic Atlas, and includes a 'measurement index' as a safety feature to ensure alignment quality. Laboratory testing demonstrated registration accuracy with inter-subject variability less than 15 degrees for yaw, roll, and pitch, and segmentation accuracy showing high correlation coefficients against expert manual segmentation for key structures: Lateral Ventricle Volume (98%), Total CSF Volume (98%), Intra-Cranial Volume (99%), and Midline Shift Index (95%), though performance can be affected by motion, artifacts, or significant anatomical deviations from the Atlas.

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 'measurement index' quantitatively define the limits of acceptable anatomic alignment variance, and what is the clinical impact of studies terminated due to exceeding this index?; What specific algorithms are used for segmentation and registration, and how do they compare in robustness to variations in CT scanner protocols or image quality?; The system generates voxel-by-voxel subtraction maps; how are these maps validated for clinical utility in tracking subtle pathological changes over time?; How does the performance of CT CoPilot™ vary across different types of brain pathologies not explicitly present in the reference neuroanatomic Atlas?; Given the reliance on a reference neuroanatomic Atlas, what is the process for updating or adapting the Atlas to account for diverse patient populations or newly identified anatomical variations?