Preview Shoulder

Preview Shoulder is a pre-operative planning software that utilizes artificial intelligence to process patient DICOM CT scans, generating interactive 3D models of shoulder anatomy for surgeons to simulate implant placement for total shoulder arthroplasty. Its methodology relies on AI-driven 3D reconstruction and digital templating to aid surgeons in developing detailed surgical plans for adult patients, with its performance validated through rigorous bench testing and surgeon assessment.

Preview Shoulder is a standalone software application designed to assist orthopedic surgeons in pre-operative planning for total shoulder arthroplasty (TSA), including both reverse and anatomic procedures, exclusively for adult patients. Its primary function is to process patient-specific DICOM CT scan series, which users import from local file sources. The software then utilizes an artificial intelligence algorithm to reconstruct a 3D model of the patient's shoulder anatomy from this CT data. Surgeons can interact with this 3D model, as well as the original CT images, to digitally place 3D representations of various implant models, thereby planning the surgical approach. The system is not intended for diagnostic purposes, and the surgeon remains responsible for all clinical decisions, including implant selection and placement. The core technology of Preview Shoulder involves AI-driven 3D reconstruction of bone from CT scans and providing tools for interactive surgical simulation. The software allows surgeons to visualize the patient's anatomy and the proposed implant positioning, ultimately generating a surgical report detailing the plan. Performance was established through comprehensive software verification and validation, including code reviews, automated and manual testing, and validation of its measurement capabilities against a benchmark tool (Osirix MD) using anonymized adult patient shoulder CT images. This bench testing confirmed the system's ability to accurately and repeatably produce 3D reconstructed bones and provide reliable surgical coordinates. Clinical testing was not conducted, as the bench testing and validation by surgeons were considered sufficient to demonstrate its intended performance for pre-surgical planning.

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 AI algorithms are employed for the 3D bone model reconstruction, and what is their documented accuracy compared to manual segmentation or other reconstruction methods?; How does the software account for variations in CT image quality or different scanner protocols, and what are the minimum image quality requirements for optimal performance?; Can the implant library be customized or updated with new implant models, and what is the process for validating these new additions within the software?; Beyond generating a planning report, does the software offer any quantitative outputs or measurements (e.g., predicted range of motion, bone-implant contact area) that aid in surgical decision-making?; What specific aspects of the user interface and workflow were highlighted by surgeons during the validation process as being particularly effective or requiring further refinement for clinical usability?