Circle Cardiovascular Imaging vs Elucid
The sharpest architectural contrast in the lane sits in this pair. Elucid sends imaging out: trained analysts segment the CTA data before PlaqueIQ classifies plaque against carotid endarterectomy histopathology, a histology grounded method that earns an A on model transparency and comes with resolved US reimbursement. Circle keeps imaging in: cvi42 Plaque runs fully on premise, so scans never leave the institution, and a Johns Hopkins user describes immediate interactive control over the analysis rather than waiting on a report from a service. What separates them beyond architecture is evidence and disclosure, Elucid publishes its mechanism and holds FDA clearance with Medicare contractor coverage, while no validation study specific to cvi42 Plaque was located and Circle publishes no accuracy methodology, holding it at C on both evidence and transparency. If your failure mode is imaging that cannot leave the enterprise, or you already run cvi42, start with Circle. If your failure mode is validated, histology grounded plaque composition with resolved coverage, start with Elucid, noting its analyst step means it is not fully autonomous.
- The strongest data residency position in the lane: cvi42 Plaque runs fully on premise, so imaging never leaves the institution, graded A on PHI stewardship and deployment, where Elucid routes scans to vendor side processing whose analyst step is disclosed but unquantified.
- Scanner agnostic by heritage: cvi42 is a multi vendor cardiovascular reading platform compatible with all major systems, graded A on interoperability, a franchise spanning cardiac MRI and CT rather than a coronary only tool.
- Direct clinician control: a Johns Hopkins user describes immediate and interactive control over the analysis, the counterpoint to receiving a finished report from a send out service.
- Ground truth against tissue: PlaqueIQ's networks were trained on carotid endarterectomy histopathology co registered to CTA, graded A on model transparency, where Circle publishes no accuracy figures or validation methodology and grades C.
- Resolved US reimbursement: four of seven Medicare contractors extended coverage from November 2024 and United Healthcare aligned from October 2025, on an FDA cleared product graded A on regulatory status against Circle's single module B.
- AI is the whole product: Elucid sells the algorithmic reading itself and grades A on AI centrality, where Circle's plaque model is a module on a reading platform, graded B on the platform not algorithm principle.
Side-by-Side
| Axis | C Circle Cardiovascular Imaging |
E Elucid |
|---|---|---|
| AI Centrality | ||
| Autonomy and Oversight Model | ||
| Model and Technology Transparency | ||
| Clinical and Operational Evidence | ||
| AI Safety and PHI Stewardship | ||
| HIPAA and BAA Posture | ||
| Security Certifications and Trust Center | ||
| FDA and Regulatory Status | ||
| AI Governance and Bias Disclosure | ||
| EHR and Interoperability Depth | ||
| Deployment Model and Data Residency | ||
| Commercial Transparency | ||
| Setting and Specialty Coverage |
The grades split by vendor type and by disclosure. Elucid is an AI plaque company graded A on AI centrality, model transparency, and regulatory status, with the caveats that trained analysts segment the data before classification so it is not fully autonomous, and its FFRCT product is pending clearance and not available. Circle is an established reading platform with a cleared plaque module, graded B on AI centrality and C on both clinical evidence, since no validation study specific to cvi42 Plaque was located, and model transparency, since no accuracy or methodology was published. Circle's decisive strength is architectural: A on PHI stewardship, deployment, and interoperability for fully on premise, scanner agnostic analysis. The contrast is sharp, Elucid moves imaging out to analysts, Circle keeps it inside the institution. Neither publishes a licence price; Circle's on premise model resolves data residency structurally where Elucid's terms are undisclosed.