Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm

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

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Purpose: The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Methods: Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. Results: The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%-91.2%) for premolars and 73.4% (95% CI, 59.9%-84.0%) for molars. Conclusions: We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

Original languageEnglish
Pages (from-to)114-123
Number of pages10
JournalJournal of Periodontal and Implant Science
Volume48
Issue number2
DOIs
Publication statusPublished - 2018 Apr 1

Fingerprint

Tooth
Learning
Bicuspid
Confidence Intervals
ROC Curve
Boidae
Confusion
Area Under Curve
Weights and Measures
Sensitivity and Specificity
Datasets

All Science Journal Classification (ASJC) codes

  • Oral Surgery
  • Periodontics

Cite this

@article{624c39aa26c64ff48d5b5045365a8cc6,
title = "Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm",
abstract = "Purpose: The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Methods: Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95{\%} confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. Results: The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0{\%} for premolars and 76.7{\%} for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8{\%} (95{\%} CI, 70.1{\%}-91.2{\%}) for premolars and 73.4{\%} (95{\%} CI, 59.9{\%}-84.0{\%}) for molars. Conclusions: We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.",
author = "Lee, {Jae Hong} and Kim, {Do Hyung} and Jeong, {Seong Nyum} and Seongho Choi",
year = "2018",
month = "4",
day = "1",
doi = "10.5051/jpis.2018.48.2.114",
language = "English",
volume = "48",
pages = "114--123",
journal = "Journal of Periodontal and Implant Science",
issn = "2093-2278",
publisher = "Korean Academy of Periodontology",
number = "2",

}

Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm. / Lee, Jae Hong; Kim, Do Hyung; Jeong, Seong Nyum; Choi, Seongho.

In: Journal of Periodontal and Implant Science, Vol. 48, No. 2, 01.04.2018, p. 114-123.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Diagnosis and prediction of periodontally compromised teeth using a deep learning-based convolutional neural network algorithm

AU - Lee, Jae Hong

AU - Kim, Do Hyung

AU - Jeong, Seong Nyum

AU - Choi, Seongho

PY - 2018/4/1

Y1 - 2018/4/1

N2 - Purpose: The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Methods: Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. Results: The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%-91.2%) for premolars and 73.4% (95% CI, 59.9%-84.0%) for molars. Conclusions: We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

AB - Purpose: The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Methods: Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. Results: The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%-91.2%) for premolars and 73.4% (95% CI, 59.9%-84.0%) for molars. Conclusions: We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

UR - http://www.scopus.com/inward/record.url?scp=85046764664&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046764664&partnerID=8YFLogxK

U2 - 10.5051/jpis.2018.48.2.114

DO - 10.5051/jpis.2018.48.2.114

M3 - Article

VL - 48

SP - 114

EP - 123

JO - Journal of Periodontal and Implant Science

JF - Journal of Periodontal and Implant Science

SN - 2093-2278

IS - 2

ER -