Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm

Jae Hong Lee, Do Hyung Kim, Seong Nyum Jeong, Seong Ho Choi

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

Objectives: Deep convolutional neural networks (CNNs) are a rapidly emerging new area of medical research, and have yielded impressive results in diagnosis and prediction in the fields of radiology and pathology. The aim of the current study was to evaluate the efficacy of deep CNN algorithms for detection and diagnosis of dental caries on periapical radiographs. Materials and methods: A total of 3000 periapical radiographic images were divided into a training and validation dataset (n = 2400 [80%]) and a test dataset (n = 600 [20%]). A pre-trained GoogLeNet Inception v3 CNN network was used for preprocessing and transfer learning. The diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, and area under the curve (AUC) were calculated for detection and diagnostic performance of the deep CNN algorithm. Results: The diagnostic accuracies of premolar, molar, and both premolar and molar models were 89.0% (80.4–93.3), 88.0% (79.2–93.1), and 82.0% (75.5–87.1), respectively. The deep CNN algorithm achieved an AUC of 0.917 (95% CI 0.860–0.975) on premolar, an AUC of 0.890 (95% CI 0.819–0.961) on molar, and an AUC of 0.845 (95% CI 0.790–0.901) on both premolar and molar models. The premolar model provided the best AUC, which was significantly greater than those for other models (P < 0.001). Conclusions: This study highlighted the potential utility of deep CNN architecture for the detection and diagnosis of dental caries. A deep CNN algorithm provided considerably good performance in detecting dental caries in periapical radiographs. Clinical significance: Deep CNN algorithms are expected to be among the most effective and efficient methods for diagnosing dental caries.

Original languageEnglish
Pages (from-to)106-111
Number of pages6
JournalJournal of Dentistry
Volume77
DOIs
Publication statusPublished - 2018 Oct

Fingerprint

Bicuspid
Dental Caries
Area Under Curve
Learning
Radiology
ROC Curve
Biomedical Research
Pathology
Sensitivity and Specificity
Datasets

All Science Journal Classification (ASJC) codes

  • Dentistry(all)

Cite this

@article{bad7a49ff5724924a247d45ab4f09035,
title = "Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm",
abstract = "Objectives: Deep convolutional neural networks (CNNs) are a rapidly emerging new area of medical research, and have yielded impressive results in diagnosis and prediction in the fields of radiology and pathology. The aim of the current study was to evaluate the efficacy of deep CNN algorithms for detection and diagnosis of dental caries on periapical radiographs. Materials and methods: A total of 3000 periapical radiographic images were divided into a training and validation dataset (n = 2400 [80{\%}]) and a test dataset (n = 600 [20{\%}]). A pre-trained GoogLeNet Inception v3 CNN network was used for preprocessing and transfer learning. The diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, and area under the curve (AUC) were calculated for detection and diagnostic performance of the deep CNN algorithm. Results: The diagnostic accuracies of premolar, molar, and both premolar and molar models were 89.0{\%} (80.4–93.3), 88.0{\%} (79.2–93.1), and 82.0{\%} (75.5–87.1), respectively. The deep CNN algorithm achieved an AUC of 0.917 (95{\%} CI 0.860–0.975) on premolar, an AUC of 0.890 (95{\%} CI 0.819–0.961) on molar, and an AUC of 0.845 (95{\%} CI 0.790–0.901) on both premolar and molar models. The premolar model provided the best AUC, which was significantly greater than those for other models (P < 0.001). Conclusions: This study highlighted the potential utility of deep CNN architecture for the detection and diagnosis of dental caries. A deep CNN algorithm provided considerably good performance in detecting dental caries in periapical radiographs. Clinical significance: Deep CNN algorithms are expected to be among the most effective and efficient methods for diagnosing dental caries.",
author = "Lee, {Jae Hong} and Kim, {Do Hyung} and Jeong, {Seong Nyum} and Choi, {Seong Ho}",
year = "2018",
month = "10",
doi = "10.1016/j.jdent.2018.07.015",
language = "English",
volume = "77",
pages = "106--111",
journal = "Journal of Dentistry",
issn = "0300-5712",
publisher = "Elsevier BV",

}

Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm. / Lee, Jae Hong; Kim, Do Hyung; Jeong, Seong Nyum; Choi, Seong Ho.

In: Journal of Dentistry, Vol. 77, 10.2018, p. 106-111.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm

AU - Lee, Jae Hong

AU - Kim, Do Hyung

AU - Jeong, Seong Nyum

AU - Choi, Seong Ho

PY - 2018/10

Y1 - 2018/10

N2 - Objectives: Deep convolutional neural networks (CNNs) are a rapidly emerging new area of medical research, and have yielded impressive results in diagnosis and prediction in the fields of radiology and pathology. The aim of the current study was to evaluate the efficacy of deep CNN algorithms for detection and diagnosis of dental caries on periapical radiographs. Materials and methods: A total of 3000 periapical radiographic images were divided into a training and validation dataset (n = 2400 [80%]) and a test dataset (n = 600 [20%]). A pre-trained GoogLeNet Inception v3 CNN network was used for preprocessing and transfer learning. The diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, and area under the curve (AUC) were calculated for detection and diagnostic performance of the deep CNN algorithm. Results: The diagnostic accuracies of premolar, molar, and both premolar and molar models were 89.0% (80.4–93.3), 88.0% (79.2–93.1), and 82.0% (75.5–87.1), respectively. The deep CNN algorithm achieved an AUC of 0.917 (95% CI 0.860–0.975) on premolar, an AUC of 0.890 (95% CI 0.819–0.961) on molar, and an AUC of 0.845 (95% CI 0.790–0.901) on both premolar and molar models. The premolar model provided the best AUC, which was significantly greater than those for other models (P < 0.001). Conclusions: This study highlighted the potential utility of deep CNN architecture for the detection and diagnosis of dental caries. A deep CNN algorithm provided considerably good performance in detecting dental caries in periapical radiographs. Clinical significance: Deep CNN algorithms are expected to be among the most effective and efficient methods for diagnosing dental caries.

AB - Objectives: Deep convolutional neural networks (CNNs) are a rapidly emerging new area of medical research, and have yielded impressive results in diagnosis and prediction in the fields of radiology and pathology. The aim of the current study was to evaluate the efficacy of deep CNN algorithms for detection and diagnosis of dental caries on periapical radiographs. Materials and methods: A total of 3000 periapical radiographic images were divided into a training and validation dataset (n = 2400 [80%]) and a test dataset (n = 600 [20%]). A pre-trained GoogLeNet Inception v3 CNN network was used for preprocessing and transfer learning. The diagnostic accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, and area under the curve (AUC) were calculated for detection and diagnostic performance of the deep CNN algorithm. Results: The diagnostic accuracies of premolar, molar, and both premolar and molar models were 89.0% (80.4–93.3), 88.0% (79.2–93.1), and 82.0% (75.5–87.1), respectively. The deep CNN algorithm achieved an AUC of 0.917 (95% CI 0.860–0.975) on premolar, an AUC of 0.890 (95% CI 0.819–0.961) on molar, and an AUC of 0.845 (95% CI 0.790–0.901) on both premolar and molar models. The premolar model provided the best AUC, which was significantly greater than those for other models (P < 0.001). Conclusions: This study highlighted the potential utility of deep CNN architecture for the detection and diagnosis of dental caries. A deep CNN algorithm provided considerably good performance in detecting dental caries in periapical radiographs. Clinical significance: Deep CNN algorithms are expected to be among the most effective and efficient methods for diagnosing dental caries.

UR - http://www.scopus.com/inward/record.url?scp=85050642974&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050642974&partnerID=8YFLogxK

U2 - 10.1016/j.jdent.2018.07.015

DO - 10.1016/j.jdent.2018.07.015

M3 - Article

C2 - 30056118

AN - SCOPUS:85050642974

VL - 77

SP - 106

EP - 111

JO - Journal of Dentistry

JF - Journal of Dentistry

SN - 0300-5712

ER -