AI in aligner treatment planning: what an FDA clearance does and does not tell you

On Sep 25, 2026, FDA cleared an optional AI-enabled prediction service inside a clear-aligner treatment-planning system. We read the clearance summary and set out what it does and does not show.

Editorial · not sponsored L2 · Regulatory record

Evidence used: One FDA 510(k) summary
Checked: One 510(k) summary PDF in full. Oct 31, 2026 (ET)

Short answer · editorial read

Too early to say. The clearance summary reports that pre-set acceptance criteria were met but publishes no figures.

Strongest source we used

1 Vendor claim
2 Regulatory record
3 Peer-reviewed study
4 Randomized trial

01 · Claims

What vendors and sponsors claim

Their own words or their own testing. Treated as claims until independently checked.

Sponsor: the AI service supports planning "by optimization of aligner activation design". The 510(k) summary describes an optional AI-enabled prediction service added to the backend 3D treatment planning software. Claim · sponsor summary
Sponsor: acceptance criteria were met. The summary states that "primary endpoint analysis demonstrated pre-specified acceptance criteria were met" and that bias across geographic, demographic and clinical-severity subgroups was evaluated. It does not publish the endpoint or the numbers. Claim · sponsor summary

02 · What we checked

What we independently checked

What we could confirm from public records and peer-reviewed abstracts or full texts in this pass. Not a clinical study.

The clearance is real and recent. K262196 was decided Substantially Equivalent on Sep 25, 2026, under regulation 872.5470 with product codes PNN and NXC. Checked · 510(k) summary
The doctor stays in the loop, and the model does not learn in the field. The summary says the generated plan "is reviewed and approved by the doctor prior to manufacturing", that there is "no direct user interaction with the AI-enabled functionality", and that the model "is static and non-adaptive". The intended use is unchanged from the predicate, K252380. Checked · 510(k) summary

03 · Limits

Limits and what we could not verify

Gaps, caveats and anything we did not read.

No public performance numbers for the AI service. The summary says detailed results "including subgroup analyses" are in the premarket notification. That is not public, so we cannot say what the endpoint was or by how much it was met. Not verifiable · no public data
No outside accuracy data in this draft. This draft cites no peer-reviewed study of predicted-versus-achieved tooth movement with or without the AI service. Gap · level 3
Availability not confirmed. A clearance does not tell you when, where or for which cases the AI service is switched on. We did not confirm rollout with the company. Unchecked · rollout
No randomized trial found. We did not find a randomized comparison of AI-assisted versus conventional aligner planning in this pass. Gap · level 4

Buyer takeaway: ask any planning vendor for predicted-versus-achieved accuracy by movement type, measured on cases like yours, and for how many refinements the comparison group needed.

04 · Sources

Sources

510(k) Summary K262196 (PDF) Sponsor-written summary, filed with FDA · Prepared Jun 25, 2026 Read Oct 31, 2026 Sponsor-written summary. Full text read.