CARA System

The CARA System is a medical device that plans and guides cardiac interventions by creating a 3D map of a patient's cardiac conduction system from CT scans and overlaying it onto live fluoroscopic images during procedures like transcatheter aortic valve replacement (TAVR) or cardiac conduction system pacing (CSP). It uses AI/ML algorithms for optical character recognition, anatomical segmentation, and catheter detection to assist physicians in avoiding or targeting the cardiac conduction system during structural heart procedures in adult patients (18+ years old).

Both the CARA System and predicate device (Cydar EV) use CT-to-fluoroscopy overlay technology with image registration based on anatomical landmarks. Key differences are anatomical specialization (CARA focuses on cardiac conduction system versus Cydar's general vasculature) and algorithm implementation (CARA uses AI/ML for OCR, segmentation, and device detection; Cydar uses deep learning for vascular segmentation and image registration/tracking). Both require physician interpretation and operate on standard workstation hardware with fluoroscopy splitters and monitors.

IEC 62304 (software validation), NIST SP 800-53 (cybersecurity), IEC 60601-1 and IEC 60601-1-2 (electrical safety and EMC). Non-clinical testing included CT-to-fluoroscopy registration accuracy (≤2.0 mm mean error, ≤3.0 mm maximum), system latency (≤133 ms at 95% confidence), and image fidelity (PSNR ≥35 dB, SSIM ≥0.95).

CARA is substantially equivalent to Cydar EV because both devices share the same regulatory classification (Class II, 21 CFR 892.1650), identical product code (OWB), and fundamental technology of CT-based anatomical modeling with fluoroscopic overlay under physician control. Differences in anatomical focus (cardiac conduction versus general vasculature) and algorithm implementation represent specialization that does not introduce new safety or effectiveness concerns, as both operate within the same framework of automated CT segmentation and fluoroscopic guidance. Clinical data demonstrated no device-related adverse events and showed clinical utility in guiding interventional procedures.