Truace
Subscribe

Review finds AI not yet safe to decide alone when orthodontic patients need a visit

Study title: Remote Monitoring or Autonomous Triage: Can Artificial Intelligence Safely Decide When Orthodontic Patients Need In-Person Care?

A narrative review of PubMed literature looked at how well AI-driven remote monitoring and tele-orthodontics can separate routine virtual review from urgent in-person care. It found that the tools perform well on obvious visual findings but are inconsistent on subtler problems.

Read the study abstract

Aim and objectives: To examine the diagnostic scope, accuracy and limitations of Artificial Intelligence (AI)-based remote orthodontic monitoring for automated triage and to assess whether such systems can safely decide when orthodontic patients need in-person care. Methods: Narrative review of PubMed literature on AI-driven remote monitoring and tele-orthodontics, focusing on diagnostic performance, triage applications and the limitations relevant to distinguishing routine virtual review from urgent in-person c…

By the Truvace desk1 min read1 sourceEvidence: Emerging · 30/100 — Commentary, opinion, or narrative review · 1 independent sourcePeer-reviewedTRV-2026-2261· cite & verify
Remote Monitoring or Autonomous Triage: Can Artificial Intelligence Safely Decide When Orthodontic Patients Need In-Person Care?"Medical Laboratory" by ben.dracup, CC BY 2.0.
Trace impact readingContested
P 47The P score combines the specificity and measured human impact of the grounded problem claim with the strength of this Trace’s cited sources.

Both sides are scored from claims and sources, not community votes.

G 47The G score combines the specificity and measured human impact of the grounded gain claim with the strength of this Trace’s cited sources.

In brief

The finding matters because a wrong triage call could leave a patient with a developing complication unseen, or send people for unnecessary visits. The authors conclude that clinician oversight, standard thresholds, external validation and medico-legal frameworks are still needed before autonomous use.

Main points

  1. AI detects visually obvious conditions with high specificity.
  2. Sensitivity is variable and often low for periodontal changes and precise tooth-level tracking.
  3. GO/NO-GO outputs can be inconsistent between assessments.
  4. Radiographic complications such as root resorption cannot be assessed from photographs.
  5. Single-center models lose accuracy when applied elsewhere.

The gain

AI detects visually obvious orthodontic conditions with high specificity, and early triage trials show faster screening.

The problem

AI sensitivity is variable and often low for periodontal changes and tooth-level tracking, so it cannot yet safely decide alone when patients need in-person care.

The rundown

The review is a narrative synthesis of PubMed literature on AI-driven remote monitoring and tele-orthodontics. It focused on diagnostic performance, triage applications and limitations relevant to deciding between routine virtual review and urgent in-person care.

Reported weaknesses include inconsistent GO/NO-GO outputs, an inability to assess radiographic complications like root resorption from photographs, and reduced accuracy when single-center models are used at other sites.

Early triage trials show faster screening, but borderline cases still require clinician review. The authors call for standardized triage thresholds, external validation and clear medico-legal frameworks.

What this doesn’t fix

The work is a narrative review of existing literature rather than a new trial, and the evidence it summarizes is limited by single-center models and early-stage triage trials.

Sources

  1. Peer-reviewedJournal of Dental Health and Oral Research2026-10-24

The debate