SMART Bun-Yo-Matic CT

The SMART Bun-Yo-Matic CT software provides orthopedic professionals with an advanced tool for diagnosing and planning foot and ankle surgeries, particularly Lapidus Arthrodesis for hallux valgus, by leveraging machine learning to automatically segment anatomical structures from CT scans and generate detailed 3D models and quantitative measurements. Its methodology involves processing CT DICOM images to create patient-specific visualizations, including implant constructs and surgical instrument parameters, validated through rigorous testing on diverse datasets to ensure accuracy in segmentation and measurements for pre-operative decision-making.

The SMART Bun-Yo-Matic CT is a software tool designed for orthopedic healthcare professionals to aid in the diagnosis and surgical planning of foot and ankle conditions, specifically targeting hallux valgus correction via Lapidus Arthrodesis. It utilizes Computed Tomography (CT) scans as input to automatically segment bones of the foot and ankle. The software then generates a comprehensive case report which includes three-dimensional (3D) models of the patient's anatomy, pre-operative measurements, and visualizations of post-correction anatomy. Additionally, it provides parameters for orthopedic surgical instruments and an example of an implant construct, facilitating detailed pre-surgical planning. The core of the SMART Bun-Yo-Matic CT's functionality relies on machine learning algorithms for bone segmentation and metal identification. These algorithms were developed and validated using a diverse dataset of 357 CT image studies from various imaging systems, patient demographics, and clinical conditions, including a minimum of 50% US-sourced images. Ground truth for validation was established by a consensus of three orthopedic surgeons. Performance testing on 82 CT image series demonstrated 100% accuracy in identifying foot and ankle bones and 98.8% accuracy for metal identification (98% specificity, 100% sensitivity). Further software verification and validation confirmed model conformance within 1.0mm and angular measurement accuracy within specified tolerances for 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. How does the software's performance in segmenting bones and identifying metal vary across different CT scanner manufacturers and imaging protocols, beyond the reported aggregate statistics?; What specific machine learning model architectures are employed for bone segmentation and metal detection, and how were they optimized for complex anatomical variations or the presence of artifacts?; Can the surgical planning module adapt to different surgical techniques for Lapidus Arthrodesis or other foot/ankle procedures beyond what is explicitly stated for hallux valgus?; How does the accuracy of the software's automated measurements (e.g., intermetatarsal angle, hallux valgus angle) compare to manual measurements performed by experienced orthopedic surgeons, and what is the inter-observer variability of the software?; What is the typical processing time for a complete case, from CT image input to the generation of the final visualization report and surgical plan, in a clinical setting?