ScanDiags Ortho L-Spine MR-Q

The ScanDiags Ortho L-Spine MR-Q software aims to provide quantitative measurements of lumbar spine anatomical structures from pre-acquired MR images. It achieves this through semi-automatic segmentation driven by deep convolutional neural networks, demonstrating robust accuracy and reliability for clinical review and analysis.

ScanDiags AG's Ortho L-Spine MR-Q software, a cleared medical image management and processing system, serves to provide quantitative spine measurements from pre-acquired lumbar spine Magnetic Resonance (MR) images. Its primary function is to assist users in hospitals and medical institutions by visualizing and documenting precise area and distance measurements of critical lumbar spine anatomical structures, including vertebral bodies, intervertebral discs, neuroforamina, and thecal sacs, for subsequent review, analysis, and interpretation. This software-as-a-medical-device (SaMD) utilizes semi-automatic segmentation powered by deep convolutional neural networks (DCNN) to process standard DICOM MRI images, subsequently generating a comprehensive PDF report of these measurements. While it functions as a valuable adjunct tool, it is explicitly designed not to produce medical diagnoses or supersede clinical judgment. Performance evaluations, notably a retrospective multicenter study, validated its robust accuracy in measurements and segmentations, demonstrating high Intraclass Correlation Coefficients (ICC) and Dice Similarity Coefficients (DSC) across various spinal structures, thereby affirming its consistency and reliability.

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. What specific DCNN architectures were employed and how were they trained and validated to ensure robustness across diverse patient populations and MRI scanner types?; How does the software handle image artifacts or variations in image quality, and what are the limitations or potential sources of error in its semi-automatic segmentation process?; Are there plans for prospective clinical trials to evaluate the software's impact on clinical workflow efficiency, diagnostic confidence, or patient outcomes?; How does the software's performance compare to manual measurements by expert clinicians, and what is the inter-observer and intra-observer variability when using the tool?; Beyond the stated structures, what is the potential for expanding the software's capabilities to measure other spinal pathologies or to integrate with other imaging modalities?