Spine CAMP

Spine CAMP™ is a fully-automated software designed to analyze X-ray images of the lumbar and cervical spine, providing quantitative metrics on spinal morphology, alignment, and motion to support clinical decision-making for adult patients. Its methodology relies on non-adaptive artificial intelligence models, trained using data from the predicate device KIMAX QMA®, to automate image processing tasks, with performance validated through statistical equivalence to this predicate.

Spine CAMP™ is a fully-automated image processing software device designed to assist medical professionals by analyzing X-ray images of the spine. Its core function is to calculate various spinal parameters, including distances, angles, linear and angular displacements, and combinations thereof, to characterize spinal morphology, alignment, and motion. The system processes sagittal plane radiographs of the lumbar and cervical spine, employing non-adaptive artificial intelligence models, trained on data derived from its predicate device (KIMAX QMA®), to automate vertebral body detection, landmark specification, and registration, which were previously manual operations. The output of Spine CAMP™ includes comprehensive reports, annotated images, and visualizations of intervertebral motion using an image registration method called "stabilization." These outputs are intended to support qualified healthcare practitioners in clinical decision-making for skeletally mature patients (18 years and older). Common clinical applications involve assessing spinal stability, alignment, degeneration, fusion status, motion preservation, and implant performance. The device's performance was validated through non-clinical bench testing, which involved comparing its outputs (vertebral landmark coordinates, image calibration, radiographic metrics) against those of the KIMAX QMA® using the same image datasets; these tests demonstrated strong correlation and statistical equivalence between the two systems.

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 performance of Spine CAMP™'s non-adaptive AI models vary across different X-ray image qualities or patient anatomies not extensively represented in the training dataset derived from KIMAX QMA®?; What are the specific thresholds for statistical equivalence used in the comparison with KIMAX QMA®, and how do these translate to clinical significance for the various measured spinal parameters?; Given that the AI models are non-adaptive, what is the proposed process for updating or retraining these models if new imaging artifacts or spinal pathologies become prevalent, or if the predicate device's methodology evolves?; Can the "stabilization" feature for visualizing intervertebral motion provide quantitative data on the quality of image registration, and how sensitive is it to subtle but clinically relevant instability?; What are the typical processing times for analyzing lumbar versus cervical spine X-rays, and how does the system handle images with significant spinal deformities or surgical implants that might challenge automated detection algorithms?