Positional assessment of lower third molar and mandibular canal using deep learning
Keywords
Loading...
Authors
Issue Date
2022-06-20
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Objective
The aim of this study is to automatically assess the positional relationship between lower third
molars (M3i) and the mandibular canal (MC) based on panoramic radiograph(s) (PR(s)).
Material and methods
As a reference, 1444 PRs were manually annotated and labeled. A deep-learning approach
based on MobileNet-V2 combination with a skeletonization algorithm and a signed distance
method was trained and validated on 1227 PRs to classify the positional relationship between
M3i and MC. Subsequently, the trained algorithm was applied onto a test set consisting of 217
PRs. Accuracy, precision, sensitivity, specificity, negative predictive value, and F1-score were
calculated.
Results
The proposed method achieved a weighted accuracy of 0.951, a precision of 0.943, a
sensitivity of 0.941, a specificity of 0.800, a negative predictive value of 0.865 and a F1-score
of 0.938.
Conclusion
AI-enhanced assessment of PR(s) can determine the positional relationship between M3i and
MC in an objective, accurate and reproducible way.
Clinical significance
The use of such an explainable AI system can assist the clinician in the assessment of the risk
of inferior alveolar nerve injury and whether additional cone-beam computed tomography is
justified.
Keywords: deep learning, artificial intelligence, panoramic radiograph, third molar, inferior
alveolar nerve, cone-beam computed tomography
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen