FDA classifies cardiovascular machine-learning notification software as Class II

The U.S. Food and Drug Administration (FDA) has finalized the classification of cardiovascular machine learning-based notification software as Class II with special controls, a decision published on September 11, 2026, in the Federal Register under Docket No. FDA–2026–N–9907 (FR Doc. 2026–18612). The final order, which amends 21 CFR Part 870, takes effect on the date of publication and codifies the classification by adding new section 21 CFR 870.2380.

The action stems from a De Novo classification request submitted on January 10, 2023, by Viz.ai, Inc. for its Viz HCM device. The FDA issued its classification order to the requester on August 3, 2023, and this final rule now codifies that decision into regulation.

What the device does

The software employs machine learning techniques to suggest the likelihood of a cardiovascular disease or condition for further referral or diagnostic follow-up. It identifies a single condition based on one or more noninvasive physiological inputs gathered during routine medical care.

• It is intended as the basis for further testing, not to provide diagnostic-quality output.
• It is not intended to identify or detect arrhythmias.
• Because the FDA has not exempted the device type under section 510(m) of the FD&C Act, it remains subject to premarket notification (510(k)) requirements.

Special controls

The FDA determined that general controls alone are insufficient to assure safety and effectiveness, so it established special controls that, combined with general controls, provide reasonable assurance. The special controls include:

• Clinical performance testing, requiring validation on a test dataset of real-world data from a representative patient population, independent of training data, and drawn from at least three geographically diverse sites.
• Software verification, validation, and hazard analysis, including documentation of the model/algorithm, integration testing, and the impact of sensor acquisition hardware on performance.
• Human factors assessment of intended users to evaluate the risk of misinterpreting device output.
• Labeling requirements covering performance testing summaries, device limitations, warnings against relying on the absence of a finding to rule out follow-up, and guidance on interpretation and compatible sensors.

Why it matters

By moving the device out of the automatic Class III assignment that applies by operation of law to post-amendments devices, the FDA reduces regulatory burden and enhances patients’ access to beneficial innovation. The De Novo classification also allows the device to serve as a predicate, so other sponsors can market substantially equivalent devices through the less burdensome 510(k) pathway rather than a new De Novo request or premarket approval. The order was signed by Grace R. Graham, Deputy Commissioner for Policy, Legislation, and International Affairs.

Source: Federal Register, Vol. 91, No. 176, September 11, 2026, Rules and Regulations, p. 1 (official reference: FR Doc. 2026–18612, Docket No. FDA–2026–N–9907).