AI–driven reconstruction of maxillary complete denture bases: a clinical feasibility study

Objectives Digital denture workflows remain limited by the inability of intraoral scanners (IOS) to capture functional soft-tissue dynamics and peripheral border molding. This clinical feasibility study evaluated whole-surface geometric reconstruction of maxillary complete denture bases from IOS data using a novel artificial intelligence (AI) model in which peripheral border features were emphasized during training. Materials and methods IOS datasets of the edentulous maxilla and corresponding clinically acceptable dentures from 79 patients were used to train an enhanced PointNet + + model incorporating a border-attention mechanism. The dataset was divided into 57 training, 7 validation, and 15 test cases. Whole-surface point-cloud performance was evaluated in the held-out test set using Density-Aware Chamfer Distance, maximum Hausdorff distance, and multi-threshold precision, recall, and F-scores. Whole-surface registration RMSE was additionally calculated for three randomly selected test-set cases that underwent additive manufacturing and an exploratory clinical try-in. Results On an independent test set, the mean Chamfer Distance was 2.6 ± 0.1 mm and the maximum Hausdorff distance was 2.4 mm. F-scores increased from 0.8 at a 0.5-mm tolerance to 1.0 at a 2.0-mm tolerance. Whole-surface RMSE values for the three pilot cases ranged from 0.4 to 0.6 mm. The additively manufactured AI denture bases showed promising adaptation and positional stability in this pilot evaluation; however, conventionally fabricated dentures provided superior retention. Conclusions Within the study’s limitations, the AI model showed potential for whole-surface reconstruction of maxillary denture bases from IOS data. However, in the descriptive try-in of three cases, the patients’ existing conventional dentures were subjectively more retentive than the additively manufactured AI-generated bases. Clinical relevance AI-based denture-base reconstruction from IOS data may contribute to future fully digital workflows. Before clinical implementation, border-specific geometric validation, larger external datasets, and objective measurements of retention, adaptation, and positional stability are required.

Authors

Publication Details

Journal
Clinical Oral Investigations
Published
2026-11-20
DOI
https://doi.org/10.1007/s00784-026-07279-8
Primary Topic
Dental materials and restorations
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

AI–driven reconstruction of maxillary complete denture bases: a clinical feasibility study

Benedikt Christopher Spies, Aiste Gintaute, Felix Burkhardt, Omnia Saleh et al.
Clinical Oral Investigations
Dental materials and restorations
article

AI–driven reconstruction of maxillary complete denture bases: a clinical feasibility study

Benedikt Christopher Spies, Aiste Gintaute, Felix Burkhardt, Omnia Saleh, Brian Goodacre, Marc C. Metzger
article en

Abstract

Objectives Digital denture workflows remain limited by the inability of intraoral scanners (IOS) to capture functional soft-tissue dynamics and peripheral border molding. This clinical feasibility study evaluated whole-surface geometric reconstruction of maxillary complete denture bases from IOS data using a novel artificial intelligence (AI) model in which peripheral border features were emphasized during training. Materials and methods IOS datasets of the edentulous maxilla and corresponding clinically acceptable dentures from 79 patients were used to train an enhanced PointNet + + model incorporating a border-attention mechanism. The dataset was divided into 57 training, 7 validation, and 15 test cases. Whole-surface point-cloud performance was evaluated in the held-out test set using Density-Aware Chamfer Distance, maximum Hausdorff distance, and multi-threshold precision, recall, and F-scores. Whole-surface registration RMSE was additionally calculated for three randomly selected test-set cases that underwent additive manufacturing and an exploratory clinical try-in. Results On an independent test set, the mean Chamfer Distance was 2.6 ± 0.1 mm and the maximum Hausdorff distance was 2.4 mm. F-scores increased from 0.8 at a 0.5-mm tolerance to 1.0 at a 2.0-mm tolerance. Whole-surface RMSE values for the three pilot cases ranged from 0.4 to 0.6 mm. The additively manufactured AI denture bases showed promising adaptation and positional stability in this pilot evaluation; however, conventionally fabricated dentures provided superior retention. Conclusions Within the study’s limitations, the AI model showed potential for whole-surface reconstruction of maxillary denture bases from IOS data. However, in the descriptive try-in of three cases, the patients’ existing conventional dentures were subjectively more retentive than the additively manufactured AI-generated bases. Clinical relevance AI-based denture-base reconstruction from IOS data may contribute to future fully digital workflows. Before clinical implementation, border-specific geometric validation, larger external datasets, and objective measurements of retention, adaptation, and positional stability are required.

Clinical Oral InvestigationsVol. 30(12)
Dental materials and restorations
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon