Spine Auto Views

Spine Auto Views is an image analysis software that employs a deep learning algorithm to automatically generate anatomically focused multi-planar reformats and apply labels to CT images of the spine. Its methodology centers on the automated detection of vertebral bodies and intervertebral discs, enabling the creation of precise oblique axial, curved coronal, and curved sagittal views without manual interaction, aiming to improve consistency and efficiency in radiological examinations.

Spine Auto Views is a non-invasive image analysis software package designed to be used with CT images for the automatic generation of anatomically focused multi-planar reformats of the spine. Its primary purpose is to streamline the creation of clinically relevant batch reformat outputs, assisting clinicians by providing these reformats with anatomical labels for vertebral bodies and intervertebral disc spaces. The software utilizes a state-of-the-art deep learning algorithm to automatically generate patient-specific spine reformats, including oblique axial views precisely angled through each disc space, as well as 3D curved coronal and curved sagittal views, without requiring manual user interaction. This technology is applicable across multiple care areas, including trauma, oncology, and routine body examinations, and is not specific to any particular disease state. The core technology involves automated detection of vertebrae and, significantly, intervertebral discs, including their position and orientation, using a deep learning algorithm. This allows for the generation of advanced reformats like oblique axial views aligned with disc spaces and curved views traced through the posterior edge of disc centers, differentiating it from previous methods that relied on vertebral centers or manual adjustments. The system automatically applies anatomical labels and exports the resulting DICOM series to designated destinations, ready for clinical review. Non-clinical testing confirmed the algorithm's capability to detect intervertebral discs using a database of retrospective CT exams. Furthermore, a clinical reader study demonstrated high user acceptance (>95%) for the automatically generated oblique axial reformats, and positive assessment for curved coronal and sagittal views when compared to standard reformats.

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 algorithm's performance in detecting disc position and orientation vary across different spinal pathologies (e.g., severe scoliosis, spondylolisthesis, post-surgical changes) and CT image qualities?; What specific anatomical landmarks or features does the algorithm prioritize for tracing the centerline through the posterior edge of disc centers for curved reformats, and how does this compare to vertebrae-centerline methods in terms of diagnostic utility for specific conditions?; Can the software automatically adjust reformatting parameters based on suspected pathology indicated in the DICOM metadata or initial scan appearance to optimize views for specific diagnostic questions (e.g., foraminal stenosis vs. central canal stenosis)?; What is the quantitative impact of Spine Auto Views on radiologist reading time and inter-reader variability for spine CT interpretation compared to manual or semi-automated reformatting techniques?; Beyond anatomical labeling, does Spine Auto Views offer any preliminary quantitative measurements (e.g., disc height, foraminal diameter estimates) derived from its automated segmentation, even if not presented as a primary measurement tool?