Truace
Subscribe

AI image matching beat manual methods for dental implant planning scans with metal distortion

Study title: AI-assisted versus point-based registration for digital planning in CBCT images with metallic artifacts

Researchers compared two ways of aligning cone-beam CT images with intraoral scans for digital implant planning in patients whose CT images contained metallic artifacts. One method used AI assistance and the other relied on manually chosen anatomical landmarks.

Read the study abstract

Accurate registration between cone-beam computed tomography (CBCT) images and intraoral scans is fundamental to digital implant planning because registration errors may propagate throughout the digital workflow and affect the accuracy of computer-guided implant surgery. Conventional point-based registration relies on manually selected anatomical landmarks and may be compromised by metallic artifacts that obscure dental structures. Artificial intelligence (AI)-assisted registration has recently emerged as a promi…

By the Truvace desk1 min read1 sourceEvidence: Moderate · 64/100 — Controlled or validated comparison · n = 45 · 1 independent sourcePeer-reviewedTRV-2026-2323· cite & verify
AI-assisted versus point-based registration for digital planning in CBCT images with metallic artifacts
Dental radiology by Le Roy Satterlee Jr., Francis. Public domain

In brief

The AI-assisted approach produced smaller alignment errors across all artifact severity levels while taking about as long. This matters because alignment errors can carry through the planning workflow into guided implant surgery, but the authors say multicenter, multi-software, prospective confirmation is still needed.

Main points

  1. The study included 45 patients and 78 paired arch observations analysed for accuracy.
  2. Median discrepancy was 0.190 mm with AI assistance versus 0.268 mm with point-based registration (p < 0.001).
  3. Errors rose with artifact burden for both methods, but AI-assisted registration stayed lower in mild, moderate and severe groups.
  4. Registration time was comparable, and intra-observer reliability was excellent (ICC 0.985).

The gain

In 45 patients with metal-distorted dental scans, AI-assisted registration aligned images with a median discrepancy of 0.190 mm versus 0.268 mm for manual point-based registration, with no added time.

The rundown

Accuracy was measured at the arch level as the mean discrepancy at three predefined anatomical landmarks using Geomagic Control X software. Arches were grouped by artifact burden into mild, moderate and severe categories.

Because the data were not normally distributed, the team used paired Wilcoxon signed-rank tests. A generalized estimating equations sensitivity analysis accounted for upper and lower arches within the same patient and confirmed the main result.

The authors conclude that AI-assisted registration may improve the accuracy and consistency of multimodal image registration in implant planning, particularly in difficult cases.

Sources

  1. Peer-reviewedBMC Oral Health2026-10-31

The debate