Content uploaded by Shankargouda Patil
Author content
All content in this area was uploaded by Shankargouda Patil on Apr 23, 2022
Content may be subject to copyright.
Diagnostics 2022, 12, 1029. https://doi.org/10.3390/diagnostics12051029 www.mdpi.com/journal/diagnostics
Review
Artificial Intelligence in the Diagnosis of Oral Diseases:
Applications and Pitfalls
Shankargouda Patil 1,*, Sarah Albogami 2, Jagadish Hosmani 3, Sheetal Mujoo 4, Mona Awad Kamil 5,
Manawar Ahmad Mansour 6, Hina Naim Abdul 6, Shilpa Bhandi 7 and Shiek S. S. J. Ahmed 8
1 Department of Maxillofacial Surgery and Diagnostic Sciences, Division of Oral Pathology, College of
Dentistry, Jazan University, Jazan 45142, Saudi Arabia
2 Department of Biotechnology, College of Science, Taif University, Taif 21944, Saudi Arabia;
dr.sarah@tu.edu.sa
3 Department of Diagnostic Dental Sciences, Oral Pathology Division, Faculty of Dentistry, College of
Dentistry, King Khalid University, Abha 61411, Saudi Arabia; jhosmani@kku.edu.sa
4 Division of Oral Medicine & Radiology College of Dentistry, Jazan University, Jazan 45142, Saudi Arabia;
sheetalmujoo@yahoo.co.uk
5 Department of Preventive Dental Science, College of Dentistry, Jazan University, Jazan 45142, Saudi Arabia;
munakamil@yahoo.com
6 Department of Prosthetic Dental Sciences, College of Dentistry, Jazan University, Jazan 45142, Saudi Arabia;
ahmad955mls@gmail.com (M.A.M.); drhinaprostho@gmail.com (H.N.A.)
7 Department of Restorative Dental Sciences, Division of Operative Dentistry, College of Dentistry, Jazan
University, Jazan 45142, Saudi Arabia; shilpa.bhandi@gmail.com
8 Multi-Omics and Drug Discovery Lab, Chettinad Academy of Research and Education,
Chennai 600130, India; shiekssjahmed@gmail.com
* Correspondence: dr.ravipatil@gmail.com
Abstract: Background: Machine learning (ML) is a key component of artificial intelligence (AI). The
terms machine learning, artificial intelligence, and deep learning are erroneously used interchange-
ably as they appear as monolithic nebulous entities. This technology offers immense possibilities
and opportunities to advance diagnostics in the field of medicine and dentistry. This necessitates a
deep understanding of AI and its essential components, such as machine learning (ML), artificial
neural networks (ANN), and deep learning (DP). Aim: This review aims to enlighten clinicians re-
garding AI and its applications in the diagnosis of oral diseases, along with the prospects and chal-
lenges involved. Review results: AI has been used in the diagnosis of various oral diseases, such as
dental caries, maxillary sinus diseases, periodontal diseases, salivary gland diseases, TMJ disorders,
and oral cancer through clinical data and diagnostic images. Larger data sets would enable AI to
predict the occurrence of precancerous conditions. They can aid in population-wide surveillance
and decide on referrals to specialists. AI can efficiently detect microfeatures beyond the human eye
and augment its predictive power in critical diagnosis. Conclusion: Although studies have recog-
nized the benefit of AI, the use of artificial intelligence and machine learning has not been integrated
into routine dentistry. AI is still in the research phase. The coming decade will see immense changes
in diagnosis and healthcare built on the back of this research. Clinical significance: This paper reviews
the various applications of AI in dentistry and illuminates the shortcomings faced while dealing
with AI research and suggests ways to tackle them. Overcoming these pitfalls will aid in integrating
AI seamlessly into dentistry.
Keywords: artificial intelligence; artificial neural network; diagnosis; deep learning;
machine learning; oral diseases
Citation: Patil, S.; Albogami, S.;
Hosmani, J.; Mujoo, S.; Kamil, M.A.;
Mansour, M.A.; Abdul, H.N.;
Bhandi, S.; Ahmed, S.S.S.J. Artificial
Intelligence in the Diagnosis of Oral
Diseases: Applications and Pitfalls.
Diagnostics 2022, 12, 1029.
https://doi.org/10.3390/
diagnostics12051029
Academic Editor: Daniel Fried
Received: 29 March 2022
Accepted: 18 April 2022
Published: 19 April 2022
Publisher’s Note: MDPI stays neu-
tral with regard to jurisdictional
claims in published maps and institu-
tional affiliations.
Copyright: © 2022 by the authors. Li-
censee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and con-
ditions of the Creative Commons At-
tribution (CC BY) license (https://cre-
ativecommons.org/licenses/by/4.0/).
Diagnostics 2022, 12, 1029 2 of 15
1. Introduction
Artificial intelligence (AI) and machine learning (ML) are terms that are often used
in research that are interchangeable even though they have different meanings. John
McCarthy, called the father of artificial intelligence, coined the term ‘artificial intelligence’
to describe machines with the potential to perform actions that were considered intelligent
without any human intervention [1]. These machines are capable of solving problems
based on the data input. Artificial intelligence has long been the mainstay of popular sci-
ence fiction. It originally stemmed from Alan Turing’s “Imitation game” or the “Turing
test” [2]. Logic Theorist, developed by Allen Newell and Herbert Simon in the year 1955,
was the first-ever AI program [3].
Machine learning (ML) is a subset of artificial intelligence [4]. Simon Cowell coined
the term in 1959 [5]. ML predicts the outcome based on the dataset provided to it using
algorithms, such as artificial neural networks (ANN). These networks mimic the human
brain and have interconnected artificial neurons that receive and analyze data signals.
Warren McCulloch and Walter Pitts suggested this concept in a seminal paper published
in 1943. Later, Minsky and Dean Edmunds developed the first ANN, the stochastic neural
analog reinforcement calculator, in 1951 [6].
Convolutional neural network (CNN) or deep learning (DL) is an approach in ML
introduced in 2006 by Hinton et al. [7]. It utilizes multi-layer neural networks to compute
data. Deep learning algorithms have the potential to analyze patterns based on the data
and improve the outcome. The development of the backpropagation algorithm in 1969
paved the way for deep learning systems [8]. Figure 1 depicts the important milestones in
the advancement of AI through the years.
Figure 1. Important milestones in the advancement of AI.
An abundant supply of data sets is crucial for implementing machine learning. Data
can refer to a variety of inputs: it can be images in the form of clinical photographs, radi-
ographs, text in the form of patient data, patient symptoms information, and audio in the
form of voice, murmurs, bruits, auscultation, or percussion sound. Figure 2 shows the
working of AI in a schematic format. Adaptability to a variety of inputs in artificial intel-
ligence added advantage to revolutionizing medical, dental, and healthcare delivery. Re-
cently, artificial intelligence in dentistry alone has created immense attention in specialties
such as orthodontics [9–11], endodontics [12,13], prosthodontics [14,15], restorative den-
tistry [16,17], periodontics [18–20], oral and maxillofacial surgery [21–23]. Research re-
veals promising results, although most applications are in the developmental phase. It
becomes a necessity that dentists need to understand the foundational concepts and ap-
plications of AI in dentistry to adapt to a changing healthcare landscape [24].
Diagnostics 2022, 12, 1029 3 of 15
Figure 2. The working of AI in a schematic format.
Today, artificial intelligence (AI) has been suggested useful in disease diagnosis, pre-
dicting prognosis, or developing patient-specific treatment strategies [25]. Particularly, AI
can assist dentists in making time-sensitive critical decisions. It can remove the human
element of error in decision-making, providing a superior and uniform quality of health
care while reducing the stress load on the dentists. This paper reviews the available liter-
ature selected that are pertaining to the research and development of AI in the diagnosis
of various oral and maxillofacial diseases, such as dental caries, periodontal disease, max-
illary sinus diseases, salivary gland diseases, temporomandibular joint disorders, osteo-
porosis, and oral cancer.
2. Search Strategy
We used PubMed, Google Scholar, and ScienceDirect to conduct a systematic search
using a variety of key terms that included “Convolutional Neural Network” or “Deep
Learning” or “Natural Language Processing” OR “neural network” OR “Machine Learn-
ing” OR “unsupervised learning” OR “Artificial Intelligence” OR “supervised learning”
for the model. Similarly, for disease, the terms included “dental caries” OR “periodontal
disease” OR “maxillary sinus diseases” OR “salivary gland diseases” OR “Temporoman-
dibular joint disorders” OR “osteoporosis” OR oral cancer. We looked for articles that
were published between January 2016 and December 2021. In addition to the search, the
reference lists of the selected article were examined to add an article for this review.
3. Dental Caries
Dental caries is the most prevalent disease across the globe. Early diagnosis is key in
decreasing caries-related indisposition in patients. Caries diagnosis is exceedingly based
on visual cues and radiographic data. This visual data can be a form of input dataset for
machine learning (ML). Devito et al. (2008) evaluated the efficiency of a multi-layer per-
ceptron neural network in diagnosing proximal caries in bitewing radiographs and con-
cluded that the diagnostic improvement was 39.4% [26]. Lee et al. (2018) used 3000 peri-
apical radiographs to evaluate the efficacy of deep convolutional neural networks to iden-
tify dental caries. High accuracy of 89%, 88%, and 82% was observed in the premolar,
molar, and both the premolar-molar regions [27]. Hung et al. (2019) conducted a study
with the test and training set comprised of data obtained from the National Health and
Nutrition Examination Survey. Supervised learning methods were used to classify the
data based on the presence or absence of root caries. Among the various ML methods used
in their study, the support vector machine (SVM) showed the best performance in identi-
fying root caries [28].
Similarly, the clinical imaging data from various sources have been used in AI mod-
els for diagnosing dental caries with excellent results. In 2019, a study examined the use
Diagnostics 2022, 12, 1029 4 of 15
of convolutional neural networks (CNN) to identify dental caries in near-infrared transil-
lumination images. CNN increased the speed and accuracy of caries detection [29]. Cantu
et al. (2020) used bitewing radiographs to assess the performance of a deep learning (DL)
network in detecting carious lesions. A total of 3686 radiographs were used, out of which
3293 were used for training while 252 were used as test data. The deep neural network
showed higher accuracy compared to dentists and can be used to detect initial caries le-
sions on bitewing radiographs [30] Park et al. (2021) tested ML prediction models for the
detection of early childhood caries compared to traditional regression models. Data of
4195 children (1–5 yrs) were obtained from the Korea National Health and Nutrition Ex-
amination survey (2007–2018) and analyzed. ML-based prediction models were able to
detect ECC, predict high-risk groups, and suggest treatment, similar to traditional predic-
tion models [31].
4. Tooth Fracture
The third most common reason for tooth loss is traumatized or cracked teeth. Early
detection and treatment can save a cracked tooth and help retain it. However, cracked
teeth often present with discontinuous symptoms, making their detection problematic.
Conventional techniques, such as CBCT and intraoral radiographs, have low sensitivity
and clarity. Paniagua et al. (2018) developed a novel method capable of detecting, quanti-
fying, and localizing cracked teeth using high-resolution CBCT scans with steerable wave-
lets and machine learning methods. The performance of ML models was tested using Hr-
CBCT scans of healthy teeth with simulated cracks. ML models showed high specificity
and sensitivity [32]. Fukuda et al. (2020) used CNN to detect vertical root fractures using
300 panoramic radiographs with 330 vertically fractured teeth with visible fracture lines.
Moreover, 80% of the data was used for training while 20% was used as a test data set.
Results suggest that CNN can be used as a diagnostic tool for the detection of vertical root
fractures [33].
5. Periodontal Diseases
Periodontal disease affects more than a billion people globally, destroying alveolar
bone and leading to tooth loss. Early diagnosis of periodontal disease using AI can im-
prove the dental status of the patient and improve their overall health and quality of life.
Ozden et al. (2015) examined the use of a support vector machine (SVM), decision tree
(DT), and ANN to identify and classify periodontal disease. Data from a total of 150 pa-
tients were used, 100 as training data and 50 as test data. The three systems classified the
data into six types of periodontal conditions. SVM and DT were more accurate as diag-
nostic support tools compared to ANN [18]. Nakano et al. (2018) used deep learning (DL)
to detect oral malodor from microbiota. A total of 90 patients, 45 patients with weak or no
malodor, and 45 patients with marked malodor were selected using organoleptic tests and
gas chromatography. Gene analysis of the amplified 16s rRNA from the patient’s saliva
was carried out. DL was used to classify the samples into malodor and healthy breath. DL
showed a predictive accuracy of 97% compared to SVM, which showed 79% [19]. ANN
has been used to predict the occurrence of recurrent aphthous ulcers. Gender, serum B12,
hemoglobin, serum ferritin, folate levels, candida count in saliva, tooth brushing fre-
quency, the number of fruits and vegetables consumed daily, were related to the occur-
rence of ulcers [20] Danks et al. (2021) used a deep neural network to measure periodontal
bone loss with the help of periapical radiographs. Periapical radiographs of single, dou-
ble, and triple rooted teeth obtained from 63 patients were used. First, the DNN was
trained to detect dental landmarks on the radiographs, and then the periodontal bone loss
was measured using these landmarks by the DNN model. The system achieved a total
percentage of correct key points of 89.9%. The system showed promising results, which
can be further improved upon by experimentation and cross-validation with extended
data sets [34] Similarly, a DL model was used to detect and measure periodontal bone loss
from panoramic images, which was then used for staging periodontitis. The performance
Diagnostics 2022, 12, 1029 5 of 15
of the DL model was compared to that of three oral radiologists. The staging was done
according to the new classification of periodontal and peri-implant disease and conditions
[35]. A total of 340 panoramic radiographs were used out of which 90% were used for
training while 10% were used for testing. Data augmentation was carried out to increase
the data by 64%. The DL model had high accuracy and excellent reliability, suggesting
that it can be used for the automatic diagnosis of periodontal disease and as a routine
surveillance tool [36].
6. Maxillary Sinus Diseases
The maxillary sinuses are structures that are commonly visualized using extraoral
radiographs. Automated identification of the sinuses and detection of any pathology in
them by AI can lead to a manifold decrease in misdiagnoses. AI can be used as a tool to
assist inexperienced dentists. Murata et al. (2018) evaluated the performance of a DL sys-
tem in diagnosing maxillary sinusitis using panoramic radiographs. The AI performance
was compared to that of two radiologists and two residents. The diagnostic performance
of the system was similar to that of the radiologists. However, the AI was superior to
dental residents [37]. Kim et al. (2019) used radiographs of the maxillary sinus in Water’s
view to evaluate the diagnostic performance of the DL system. AI showed a statistically
significant improved sensitivity and specificity to radiologists [38]. Mucosal thickening
and mucosal retention cysts are often missed by radiologists. Kuwana et al. (2021) used
OPG to detect and classify lesions in the maxillary sinus using a DL object detection tech-
nique. Detection of the normal maxillary sinus and inflamed maxillary sinus showed
100% sensitivity, whereas the detection sensitivity of mucosal retention cysts was 98% and
89% in the two test data sets that were used. This DL model can be reliably used in a
clinical setup [39]. A recent study proposed a CNN model to assist radiologists. The CNN
model is capable of detecting and segmenting mucosal thickening and mucosal retention
cysts of the maxillary sinus using CBCT images. A total of 890 maxillary sinuses from 445
patients were used in the study. Low dose images were used for training and testing,
while full-dose images were used as test data sets. The CNN model performed effectively
in both dosage images with no significant difference [40].
7. Salivary Gland Diseases
Salivary gland diseases pose a diagnostic challenge to inexperienced dentists due to
their confusing and similar morphological resemblances. AI can be a valuable tool in sup-
porting diagnosing diseases of the salivary gland. DL models can, in some instances, be
superior to radiologists. In an early Japanese study, researchers used DL to detect fatty
degeneration of the salivary gland parenchyma on CT images, which is evident in the case
of Sjogren’s syndrome. Of the total 500 CT images, 400 CT images (200 CT images of the
control group, 200 CT images of the Sjogren’s syndrome patients) were used as a training
dataset while 100 CT images were used as the test data set to analyze the performance of
the ML system. The diagnostic performance of DL was equivalent to that of experienced
radiologists and significantly superior to inexperienced radiologists [41]. The low inci-
dence and overlapping morphologic features of salivary gland tumors make them chal-
lenging to diagnose for clinicians. ML was used to detect malignant salivary gland tumors
based on their cytologic appearance. A recursive partitioning algorithm was used to clas-
sify 115 malignant tumor samples into 12 morphologic variables. This performance was
compared to that of experienced clinicians. The decision tree system test was effective in
narrowing down the differential diagnoses, increasing the accuracy of pathological diag-
nosis [42]. AI has the potential to be used as a tool to predict the recurrence of salivary
gland malignancies [43]. Facial nerve injury after surgical treatment for a salivary gland
tumor is a severe complication. Chiesa-Estomba et al. (2021) used clinical, radiological,
histological, and cytological data to predict the occurrence of facial nerve palsy in patients
and reported that AI can be used as an assessment tool for the prediction of facial nerve
Diagnostics 2022, 12, 1029 6 of 15
injury so that both surgeons and patients are well aware of the complications in advance
[44].
8. Temporomandibular Joint Disorders
Diagnosing TMJ disorders is a challenging issue for inexperienced dentists. ANN
systems can simplify and assist in this diagnosis. The performance of an ANN model was
tested by recognizing non-reducing disks in patients. The frontal chewing data from 68
patients with normal disks, unilateral and bilateral non-reducing disks, were obtained.
Half the data was used to train the ANN system, while the other half was used for testing.
The system showed an acceptable level of error and showed potential as a supporting
diagnostic tool with an excellent cost/benefit ratio [45]. Bas et al. (2011) conducted a similar
study using clinical symptoms. The clinical symptoms and diagnoses of 219 patients were
obtained from experienced oral and maxillofacial surgeons. The data from the first 161
patients was used to train the ANN, while the rest of the data was used to test the ANN.
The neural network showed acceptable results in diagnosing internal derangements of the
TMJ. Additional patient data, clinical data, radiographs, and images could improve the
diagnostic capacity of ANN [46]. Iwasaki (2014) applied Bayesian belief network analysis
to MRI images to determine the progression of TMJ disorders. A total of 295 cases with
590 sides of TMJs were used with 11 algorithms. The results suggested that the osteoar-
thritic changes progressed from condyle to articular fossa, and then to the mandibular
bone contours. Age, disk form, bony space, and condylar translations were elements that
affected disk displacement and bony changes [47]. Choi et al. (2021) developed an AI
model to detect osteoarthritis from OPG images. This AI model can be used in clinical
setups where a CT facility or a maxillofacial radiologist is not readily available [48]. Orhan
et al. (2021) used magnetic resonance images of TMJs to detect TMJ pathologies, such as
condylar osseous changes and disk derangements using an AI model [49]. AI models can
use a variety of input data to learn. Researchers have even used infrared thermography
images of patients with masseter and lateral pterygoid muscles as the area of interest to
diagnose TMJ disorders in an AI model [50].
9. Osteoporosis
Osteoporosis can be detected on panoramic radiographs. Various indices, such as the
gonion index, mental index, mandibular cortical index, and panoramic mandibular index
have been used previously to detect osteoporosis [51–54]. AI could simplify the diagnosis
of osteoporosis and buttress the work of radiologists. Kim et al. (2019) evaluated the per-
formance of deep convolutional neural networks (DCNN) based on computer-aided di-
agnosis (CAD) in diagnosing osteoporosis from panoramic images against radiologists
with 10 years of experience. Out of the total 1268 images, 200 images were used as test
images. The DCNN- CAD showed results that were highly agreeable with the diagnostic
results of the radiologists. DCNN can be used to help dentists in early diagnosis, and re-
ferral to specialists [55]. Lee et al. (2020) compared different types of CNN models to as-
sess which model worked best for the diagnosis of osteoporosis and found that the CNN
model with transfer learning and fine-tuning was best able to diagnose osteoporosis au-
tomatedly [56].
10. Oral Cancer and Cervical Lymph Node Metastasis
Oral cancer is the sixth most common malignancy worldwide. Early detection can
lead to a better prognosis and a better survival rate [57]. AI can aid in early diagnosis and
decrease the mortality and morbidity associated with oral cancer. Nayak et al. (2005) used
ANN to discriminate between normal, premalignant, and tissues using laser-induced au-
tofluorescence spectra recordings. This was compared to a principal component analysis
of the same issues. The results showed an accuracy of 98.3%, specificity of 100%, and sen-
sitivity of 96.5%, suggesting that this method can have efficient real-time applications [58].
Diagnostics 2022, 12, 1029 7 of 15
Uthoff et al. (2017) used CNN to detect precancerous and cancerous lesions from auto-
fluorescence images and white light images. CNN was more effective than specialists in
diagnosing precancerous and cancerous lesions. The performance of the CNN model can
improve with larger data sets [59]. Aubreville et al. (2017) used DL to identify oral cancer
based on confocal laser endomicroscopy (CLE) images. This method had an accuracy of
88.3% and a specificity of 90% [60]. Shams et al. (2017) conducted a comparative study to
predict the development of oral cancer from oral potentially-malignant lesions using deep
neural networks (DNN). DNN was compared to support vector machines, regularized
least squares, and multi-layer perception. DNN had a higher accuracy rate of 96% com-
pared to the other systems [61]. These findings were confirmed by Jeyraj et al. (2019). CNN
was used to distinguish between cancerous and non-cancerous tissues based on hyper-
spectral images. Results suggest that CNN can be employed for image-based classification
and diagnosis of oral cancer without expert supervision [62]. Recently, a lot of research
has taken place in the field of oral cancer research. Many studies have successfully devel-
oped AI models that are capable of predicting the occurrence and recurrence of oral cancer
[63–67].
Several studies have compared deep learning (DL) systems against experienced ra-
diologists with varied results. Ariji et al. (2014) assessed the performance of DL in the
identification of cervical node metastasis using CT images. CT images of 137 positive his-
tologically proven cervical lymph nodes and 314 negative histological lymph nodes from
45 patients with oral squamous cell carcinoma were used. The results of the DL approach
were compared against two trained radiologists. The DL network was as accurate as
trained radiologists [68]. The researchers also used DL to detect the extra-nodal extension
of cervical lymph node metastases. A total of 703 CT images from 51 patients with and
without extra-nodal extension were collected and 80% were used as training data while
20% were used as test data. The performance of the DL system was significantly superior
to that of the radiologist, suggesting that it can be used as a diagnostic tool for detecting
extra-nodal metastasis [69].
Overall, this review on artificial intelligence for diagnosis points towards a positive
trend with encouraging results. Neural networks and machine learning appear as effec-
tive or better than trained radiologists and clinicians (Table 1) in detecting caries, sinusitis,
periodontal disease, and TMJ disorders. Cancer diagnosis by using artificial intelligence
models can curate diverse data streams to render judgments, assess risk and referral to
specialists (Table 2). Studies on premalignant lesions, lymph nodes, salivary gland tu-
mors, and squamous cell carcinoma show encouraging results for the diagnostic and
prognostic value of artificial intelligence. These efforts may reduce mortality rates through
early diagnosis and effective therapeutic interventions. These platforms will require large
data sets and resources to analyze data to provide a precise and cost-effective diagnosis.
In order to be securely integrated into daily clinical procedures, these models are needed
to be refined to reach the highest accuracy with specificity and sensitivity. Furthermore,
also required are regulatory frameworks for the deployment of these models in clinical
practice.
Table 1. Summary of studies examining the use of artificial intelligence in dental diagnosis.
Study
Algo-
rithm
Used
Study
Factor
Modality
Num-
ber of
Input
Data
Performance Comparison Outcome
Lee J et al.
(2018) [27] CNN
Dental
caries
Periapical ra-
diographs 600 Mean AUC—0.890 4 Dentists
Deep CNN showed a
considerably good
performance in de-
tecting dental caries
in periapical radio-
graphs.
Diagnostics 2022, 12, 1029 8 of 15
Casalegno et
al. (2019)
[29]
CNN
Dental
caries
Near-
infrared
transillumi-
nation imag-
ing
217
ROC of 83.6% for occlusal
caries; ROC of 84.6% for
proximal caries
Dentists with clinical experi-
ence
CNN showed in-
creased speed and
accuracy in detecting
dental caries
Cantu et al.
(2019) [30] CNN
Dental
caries
Bitewing ra-
diographs 141 Accuracy 0.80; sensitivity
0.75%; specificity 0.83%; 4 experienced dentists
AI model was more
accurate than den-
tists
Radke et al.
(2003) [45] ANN
Disk
dis-
place-
ment
Frontal plane
jaw record-
ings from
chewing
68 Accuracy 86.8%,
specificity
100%, sensitivity 91.8% None
The proposed model
has an acceptable
level of error and an
excellent cost/benefit
ratio.
Park YH et
al. (2021)
[31]
ML
Early
child-
hood
caries
Demographic
details, oral
hygiene man-
agement de-
tails, mater-
nal details
4195
AUROC between 0.774 and
0.785 Traditional regression model
Both ML-
based and
traditional regres-
sion models showed
favorable perfor-
mance and can be
used as a supporting
tool.
Kuwana et
al. (2021)
[39]
CNN
Maxil-
lary si-
nus le-
sions
Panoramic ra-
diographs 1174
Diagnostic accuracy, sensi-
tivity, and specificity were
90–91%, 81–85% and 91–
96% for maxillary sinusitis
and 97–100%, 80–
100% and
100% for maxillary sinus
cysts.
None
The proposed deep
learning model can
be reliably used for
detecting the maxil-
lary sinuses and
identifying lesions in
them.
Murata et al.
(2018) [37] CNN
Maxil-
lary si-
nusitis
Panoramic ra-
diographs 120
Accuracy 87.5%; sensitivity
86.7%; specificity 88.3%
2 experienced radiologists, 2
dental residents
The AI
model can be
a supporting tool for
inexperienced den-
tists
Kim et al.
(2019) [38] CNN
Maxil-
lary si-
nusitis
Water’s view
radiographs
200
AUC of 0.93 for temporal;
AUC of 0.88 for geographic
external
5 radiologists
the AI-
based model
showed
statistically
higher performance
than radiologists.
Hung KF et
al. (2022)
[40]
CNN
maxil-
lary si-
nusitis
Cone-beam
computed to-
mography
890
AUC for detection of mu-
cosal thickening and mu-
cous retention cyst was
0.91 and 0.84 in low dose,
and 0.89 and 0.93 for high
dose
None
The proposed model
can accurately detect
mucosal thickening
and mucous reten-
tion cysts in both
low and high-dose
protocol CBCT
scans.
Danks et al.
(2021) [34]
DNN
sym-
metric
hour-
glass ar-
chitec-
ture
Perio-
dontal
bone
loss
Periapical ra-
diographs 340
Percentage Correct Key-
points of 83.3% across all
root morphologies
Asymmetric hourglass archi-
tecture, Resnet
The proposed sys-
tem showed promis-
ing capability in lo-
calizing landmarks
and periodontal
bone loss and per-
formed 1.7% better
than the next best ar-
chitecture.
Chang et al.
(2020) [36] CNN
Perio-
dontal
bone
loss
Panoramic ra-
diographs 340
Pixel accuracy of 0.93; Jac-
card index of 0.92; dice co-
efficient values of 0.88 for
localization of
periodontal
bone.
None
The proposed model
showed high accu-
racy and excellent re-
liability in the detec-
tion of periodontal
Diagnostics 2022, 12, 1029 9 of 15
bone loss and classi-
fication of periodon-
titis
Ozden et al.
(2015) [18] ANN
Perio-
dontal
disease
Risk factors,
periodontal
data, and ra-
diographic
bone loss
150
Performance of SVM & DT
was 98%; ANN was 46% SVM &DT
SVM and DT
showed good perfor-
mance in the classifi-
cation of periodontal
disease while ANN
had the worst per-
formance
Devito et al.
(2008) [26] ANN
Proxi-
mal car-
ies
Bitewing ra-
diograph 160 ROC curve area of 0.884 25 examiners
ANN could improve
the performance of
diagnosing proximal
caries.
Dar-
Odeh et
al. (2010)
[20]
ANN
Recur-
rent
aph-
thous
ulcers
Predisposing
factor and
RAU status
96
Accuracy of
prediction for
network 3 & 8 is 90%; 4,6 &
9 is 80%; 1& 7 is 70%; 2 & 5
is 60%
None
the ANN model
seemed to use gen-
der, hematologic and
mycologic data,
tooth brushing, fruit,
and vegetable con-
sumption for the
prediction of RAU.
Hung M et
al. (2019)
[28]
CNN Root
caries Data set 5135
Accuracy 97.1%; Precision
95.1%; sensitivity 99.6%;
specificity 94.3%
Trained medical personnel
Shows good perfor-
mance and can be
clinically imple-
mented.
Iwasaki et
al. (2015)
[47]
BBN
Tem-
poro-
man-
dibular
disor-
ders
Magnetic res-
onance imag-
ing
590
Of the 11 BBN algorithms
used path conditions using
resubstitution validation
and 10—fold cross-valida-
tion showed an accuracy of
>99%
necessary path condition, path
condition, greedy search-and-
score with Bayesian infor-
mation criterion, Chow-Liu
tree, Rebane-Pearl poly tree,
tree augmented naïve Bayes
model, maximum log-likeli-
hood, Akaike information cri-
terion, minimum description
length, K2 and C4.5
The proposed model
can be used to pre-
dict the prognosis of
TMDs.
Orhan et a
l.
(2021) [49] ML
Tem-
poro-
man-
dibular
disor-
ders
Magnetic res-
onance imag-
ing
214
The performance accuracy
for condylar changes and
disk
displacement are 0.77
and 0.74
logistic regression (LR), ran-
dom forest (RF), decision tree
(DT), k-nearest neighbors
(KNN), XGBoost, and support
vector machine (SVM)
The proposed model
using KNN and RF
was found to be op-
timal for predicting
TMJ pathologies
Diniz de
lima et al.
(2021) [50]
ML
Tem-
poro-
man-
dibular
disor-
ders
Infrared ther-
mography 74
Semantic and radiomic-se-
mantic associated ML fea-
ture extraction methods
and MLP classifier showed
statistically good perfor-
mance in detecting TMDs
KNN, SVM, MLP
ML model associated
with infrared ther-
mography can be
used for the detec-
tion of TMJ patholo-
gies
Bas B et al.
(2012) [46] ANN
TMJ in-
ternal
de-
range-
ments
Clinical
symptoms
and diagno-
ses
219
Sensitivity and specificity
for unilateral
and anterior
disk displacement with
and without reduction
were 80% & 95% and 69%
& 91%; for bilateral and an-
terior disk displacement
Experienced surgeon
The developed
model can be used as
a supportive diag-
nostic tool for the di-
agnoses of subtypes
of TMJ internal de-
rangements
Diagnostics 2022, 12, 1029 10 of 15
with and without reduc-
tion were 37% &100% and
100% & 89% respectively.
Choi et al.
(2021) [48] CNN
TMJ os-
teoar-
thritis
Panoramic ra-
diographs 1189
Accuracy of 0.78, the sensi-
tivity of 0.73, and specific-
ity of 0.82
Oral and maxillofacial radiolo-
gist
The developed
model showed per-
formance equivalent
to experts and can be
used in general prac-
tices where OMFR
experts or CT is n
Fukuda et
al. (2019)
[33]
CNN
Vertical
root
fracture
Panoramic ra-
diograph 60 The precision of 0.93; Re-
call of 0.75
2 Radiologists and 1 Endodon-
tist
The CNN model was
a promising support-
ive tool for the detec-
tion of vertical root
fracture.
Table 2. Summary of studies examining the use of artificial intelligence in cancer diagnosis.
Study
Algo-
rithm
Used
Study
Factor Modality
Num-
ber of
Input
Data
Performance Comparison Outcome
Ariji et al.
(2019) [69]
CNN
Extra-
nodal ex-
tension of
cervical
lymph
node
CT images 703 Accuracy of 84% 4 radiologists
The diagnostic per-
formance of the DL
model was signifi-
cantly higher than
the radiologists
Lopez—
Janeiro et
al. (2022)
[42]
ML
Malignant
salivary
gland tu-
mor
Primary tumor resec-
tion specimens 115 84–89% of the samples were diag-
nosed correctly None
The developed
model can be used
as a guide for the
morphological ap-
proach to the diag-
nosis of malignant
salivary gland tu-
mors
Felice et al.
(2021) [43]
Deci-
sion
tree
Malignant
salivary
gland tu-
mor
Age at diagnosis, gen-
der, salivary gland
type, h
istologic type,
surgical margin, tu-
mor stage, n
ode stage,
lymphovascular inva-
sion/perineural inva-
sion, t
ype of adjuvant
treatment
54
5-year disease-free survival was
62.1%. Important variables to pre-
dict recurrence were pathological
tumor and node stage. Based on
the variables, 3 groups were parti-
tioned as pN0, pT1-2 pN+ and
PT3-4 pN+ with 26%, 38% and
75% of recurrence and 73.7%,
57.1% and 34.3% disease-free sur-
vival rate, respectively
None
The proposed
model can be used
to classify patients
with salivary gland
malignancy and
predict the recur-
rence rate.
Ariji et al.
(2019) [68]
CNN
Metasta-
sis of cer-
vical
lymph
nodes
CT images 441
Accuracy 78.2%; sensitivity 75.4%;
specificity 81.1% not clear
The diagnostic per-
formance of the
CNN model is simi-
lar to that of radiol-
ogists
Nayak et
al. (2005)
[58]
ANN
Normal,
premali-
gnant and
malignant
condi-
tions
Pulsed laser-
induced
autofluorescence
spectroscopic studies
Not
clear
Specificity and sensitivity were
100% and 96.5%
Principal compo-
nent analysis
ANN showed better
performance com-
pared to PCA in the
classification of nor-
mal, premalignant,
Diagnostics 2022, 12, 1029 11 of 15
and malignant con-
ditions
Shams et
al. (2017)
[61]
DNN
Oral can-
cer
Gene expression pro-
filing 86 Accuracy of 96%
support vector
machine (SVM),
Regularized
Least Squares
(RLS), multi-
layer perceptron
(MLP) with back-
propagation
The proposed sys-
tem showed signifi-
cantly higher per-
formance, which
can be easily imple-
mented
Jeyaraj et
al. (2019)
[62]
CNN
Oral can-
cer Hyperspectral images
600 Accuracy of 91.4% for benign tis-
sue and 94.5% for normal tissue
Support vector
machine and
Deep belief net-
work
The proposed
method can be de-
ployed for the auto-
matic classification
of
Aubreville
et al. (2017)
[60]
CNN
oral squa-
mous cell
carci-
noma
Confocal laser en-
domicroscopy (CLE)
images
7894 AUC 0.96; Mean accuracy sensi-
tivity 86.6%; specificity 90%; not clear
This method
seemed
better than
the state-of-the-
art
CLE recognition
system
Uthoff et
al. (2018)
[59]
CNN
Precan-
cerous
and can-
cerous le-
sions
Autofluorescence and
white light imaging 170
sensitivity, specificity, positive,
and negative predictive values
ranging from 81.25 to 94.94%
None
The proposed
model is a low-
cost,
portable, and easy-
to-use system.
11. Prospects and Challenges
AI in dentistry is mostly in the nascent stages. It has yet to enter the realm of day-to-
day dentistry. Numerous hurdles remain before it can seamlessly integrate into diagnosis
and healthcare. Machine learning requires large volumes of data that are held by private
dental setups and institutions. Data sharing and privacy are issues that need to be dealt
with through federated guidelines and laws. This can rectify a common drawback re-
ported in most studies: a shortage of data sets. European and American legislative bodies
have passed the General Data Protection Act (GDPRA) and the California Consumer Pro-
tection Act (CCPA) to limit the risks of data sharing and protect consumer confidentiality
[70,71]. Federated data systems similar to VANTAGE6, Personal Health Train (PHT), and
DataSHIELD need to be developed so that data can be shared without breaching data
security policies [71–73]. AI can also convert widely heterogeneous data into curated ho-
mogeneous data that is easy to use and interpret. Most of the studies in this review have
been supervised image-based studies for the identification of structures or associations.
This only provides partial information required for decision-making or treatment. AI ca-
pable of unsupervised diagnosis and prediction of diseases needs to be built to reduce
subjective errors and provide standardized decisions. A shortage of manpower and re-
sources is emblematic of rural communities. AI-based healthcare initiatives can connect
rural and far-flung places with quality health care, benefiting the local population. Pro-
spective randomized control trials and cohort studies have to be performed to evaluate
the impact of AI on treatment and to test the outcomes and cost-effectiveness of AI [74–
76].
12. Conclusions
The field of artificial intelligence (AI) is rapidly evolving to fill an ever-expanding
niche in medicine and dentistry. Most AI research is still in its nascent stage. Increased
availability of patient data can accelerate research into artificial intelligence, machine
learning, and neural networks. Today, there are few real-time AI applications integrated
into the internal operational process of dental clinics. Research has shown that data-driven
Diagnostics 2022, 12, 1029 12 of 15
AI is reliable, transparent, and in certain cases, better than humans in diagnosis. AI can
replicate human functions of reasoning, planning, and problem-solving. Its application
can save time and storage, reduce manpower and eliminate human errors in diagnosis.
The rise of artificial intelligence in dental care will revolutionize dentistry and usher in
wider access to dental health care with better patient outcomes.
Author Contributions: Conceptualization, S.P., S.A., and J.H.; methodology, S.S.S.J.A., M.A.K., and
S.M.; software, H.N.A. and M.A.M.; validation, M.A.M., S.B., and S.P.; formal analysis, S.M. and
S.A.; investigation, M.A.K. and H.N.A.; resources, M.A.M. and S.B.; data curation, S.A. and S.M.;
writing—original draft preparation, S.P., S.A., J.H., and S.M.; writing—review and editing, S.S.S.J.A.
and M.A.K.; visualization, H.N.A. and M.A.M.; supervision, S.B. and S.S.S.J.A. All authors have
read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Acknowledgments: The authors would like to acknowledge the inputs provided by Ahmed
Alamoudi, Bassam Zidane (King Abdulaziz University) in revising the manuscript.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Ensmenger, N.; Nilsson, N.J. The Quest for Artificial Intelligence: A History of Ideas and Achievements. Xv + 562 pp., Index; Cambridge
University Press: Cambridge, MA, USA; New York, NY, USA, 2010; Volume 102, pp. 588–589. https://doi.org/10.1086/663065.
2. Turing, A.M. On Computable Numbers, with an Application to the Entscheidungsproblem; London Mathematical Society: London,
UK, 1937; Volume s2–s42, pp. 230–265.
3. Newell, A.; Simon, H.A. Computer Science as Empirical Inquiry. Commun. ACM 1976, 19, 113–126.
https://doi.org/10.1145/360018.360022.
4. Khanagar, S.B.; Al-Ehaideb, A.; Maganur, P.C.; Vishwanathaiah, S.; Patil, S.; Baeshen, H.A.; Sarode, S.C.; Bhandi, S. Develop-
ments, Application, and Performance of Artificial Intelligence in Dentistry—A Systematic Review. J. Dent. Sci. 2021, 16, 508–
522. https://doi.org/10.1016/j.jds.2020.06.019.
5. Bowling, M.; Fürnkranz, J.; Graepel, T.; Musick, R. Machine Learning and Games. Mach. Learn. 2006, 63, 211–215.
https://doi.org/10.1007/s10994-006-8919-x.
6. Park, W.J.; Park, J.B. History and Application of Artificial Neural Networks in Dentistry. Eur. J. Dent. 2018, 12, 594–601.
https://doi.org/10.4103/ejd.ejd_325_18.
7. Hinton, G.E.; Osindero, S.; Teh, Y.W. A Fast Learning Algorithm for Deep Belief Nets. Neural Comput. 2006, 18, 1527–1554.
https://doi.org/10.1162/NECO.2006.18.7.1527.
8. Sarker, I.H. Deep Learning: A Comprehensive Overview on Techniques, Taxonomy, Applications, and Research Directions. SN
Comput. Sci. 2021, 2, 420. https://doi.org/10.1007/S42979-021-00815-1.
9. Jung, S.K.; Kim, T.W. New Approach for the Diagnosis of Extractions with Neural Network Machine Learning. Am. J. Orthod.
Dentofac. Orthop. 2016, 149, 127–133. https://doi.org/10.1016/J.AJODO.2015.07.030.
10. Niño-Sandoval, T.C.; Perez, S.V.G.; González, F.A.; Jaque, R.A.; Infante-Contreras, C. An Automatic Method for Skeletal Pat-
terns Classification Using Craniomaxillary Variables on a Colombian Population. Forensic Sci. Int. 2016, 261, 159.e1–159.e6.
https://doi.org/10.1016/J.FORSCIINT.2015.12.025.
11. Niño-Sandoval, T.C.; Pérez, S.V.G.; González, F.A.; Jaque, R.A.; Infante-Contreras, C. Use of Automated Learning Techniques
for Predicting Mandibular Morphology in Skeletal Class I, II and III. Forensic Sci. Int. 2017, 281, 187.e1–187.e7.
https://doi.org/10.1016/J.FORSCIINT.2017.10.004.
12. Saghiri, M.A.; Garcia-Godoy, F.; Gutmann, J.L.; Lotfi, M.; Asgar, K. The Reliability of Artificial Neural Network in Locating
Minor Apical Foramen: A Cadaver Study. J. Endod. 2012, 38, 1130–1134. https://doi.org/10.1016/J.JOEN.2012.05.004.
13. Saghiri, M.A.; Asgar, K.; Boukani, K.K.; Lotfi, M.; Aghili, H.; Delvarani, A.; Karamifar, K.; Saghiri, A.M.; Mehrvarzfar, P.; Garcia-
Godoy, F. A New Approach for Locating the Minor Apical Foramen Using an Artificial Neural Network. Int. Endod. J. 2012, 45,
257–265. https://doi.org/10.1111/J.1365-2591.2011.01970.X.
14. Chen, Q.; Wu, J.; Li, S.; Lyu, P.; Wang, Y.; Li, M. An Ontology-Driven, Case-Based Clinical Decision Support Model for Remov-
able Partial Denture Design. Sci. Rep. 2016, 6, 27855. https://doi.org/10.1038/SREP27855.
15. Li, H.; Lai, L.; Chen, L.; Lu, C.; Cai, Q. The Prediction in Computer Color Matching of Dentistry Based on GA+BP Neural Net-
work. Comput. Math. Methods Med. 2015, 2015, 816719. https://doi.org/10.1155/2015/816719.
16. Aliaga, I.J.; Vera, V.; de Paz, J.F.; García, A.E.; Mohamad, M.S. Modelling the Longevity of Dental Restorations by Means of a
CBR System. BioMed Res. Int. 2015, 2015, 540306. https://doi.org/10.1155/2015/540306.
Diagnostics 2022, 12, 1029 13 of 15
17. Thanathornwong, B.; Suebnukarn, S.; Ouivirach, K. Decision Support System for Predicting Color Change after Tooth Whiten-
ing. Comput. Methods Programs Biomed. 2016, 125, 88–93. https://doi.org/10.1016/J.CMPB.2015.11.004.
18. Ozden, F.O.; Özgönenel, O.; Özden, B.; Aydogdu, A. Diagnosis of Periodontal Diseases Using Different Classification Algo-
rithms: A Preliminary Study. Niger. J. Clin. Pract. 2015, 18, 416. https://doi.org/10.4103/1119-3077.151785.
19. Nakano, Y.; Suzuki, N.; Kuwata, F. Predicting Oral Malodour Based on the Microbiota in Saliva Samples Using a Deep Learning
Approach. BMC Oral Health 2018, 18, 128. https://doi.org/10.1186/s12903-018-0591-6.
20. Dar-Odeh, N.S.; Alsmadi, O.M.; Bakri, F.; Abu-Hammour, Z.; Shehabi, A.A.; Al-Omiri, M.K.; Abu-Hammad, S.M.K.; Al-Mashni,
H.; Saeed, M.B.; Muqbil, W.; et al. Predicting Recurrent Aphthous Ulceration Using Genetic Algorithms-Optimized Neural
Networks. Adv. Appl. Bioinform. Chem. 2010, 3, 7. https://doi.org/10.2147/AABC.S10177.
21. Kositbowornchai, S.; Plermkamon, S.; Tangkosol, T. Performance of an Artificial Neural Network for Vertical Root Fracture
Detection: An Ex Vivo Study. Dent. Traumatol. 2013, 29, 151–155. https://doi.org/10.1111/J.1600-9657.2012.01148.X.
22. de Bruijn, M.; ten Bosch, L.; Kuik, D.J.; Langendijk, J.A.; Leemans, C.R.; Leeuw, I.V. de Artificial Neural Network Analysis to
Assess Hypernasality in Patients Treated for Oral or Oropharyngeal Cancer. Logop. Phoniatr. Vocol. 2011, 36, 168–174.
https://doi.org/10.3109/14015439.2011.606227.
23. Chang, S.W.; Abdul-Kareem, S.; Merican, A.F.; Zain, R.B. Oral Cancer Prognosis Based on Clinicopathologic and Genomic
Markers Using a Hybrid of Feature Selection and Machine Learning Methods. BMC Bioinform. 2013, 14, 170.
https://doi.org/10.1186/1471-2105-14-170.
24. Pethani, F. Promises and Perils of Artificial Intelligence in Dentistry. Aust. Dent. J. 2021, 66, 124–135.
https://doi.org/10.1111/adj.12812.
25. Shan, T.; Tay, F.R.; Gu, L. Application of Artificial Intelligence in Dentistry. J. Dent. Res. 2021, 100, 232–244.
https://doi.org/10.1177/0022034520969115.
26. Devito, K.L.; de Souza Barbosa, F.; Filho, W.N.F. An Artificial Multilayer Perceptron Neural Network for Diagnosis of Proximal
Dental Caries. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endodontol. 2008, 106, 879–884. https://doi.org/10.1016/J.TRI-
PLEO.2008.03.002.
27. Lee, J.H.; Kim, D.H.; Jeong, S.N.; Choi, S.H. Detection and Diagnosis of Dental Caries Using a Deep Learning-Based Convolu-
tional Neural Network Algorithm. J. Dent. 2018, 77, 106–111. https://doi.org/10.1016/J.JDENT.2018.07.015.
28. Hung, M.; Voss, M.W.; Rosales, M.N.; Li, W.; Su, W.; Xu, J.; Bounsanga, J.; Ruiz-Negrón, B.; Lauren, E.; Licari, F.W. Application
of Machine Learning for Diagnostic Prediction of Root Caries. Gerodontology 2019, 36, 395–404.
https://doi.org/10.1111/GER.12432.
29. Casalegno, F.; Newton, T.; Daher, R.; Abdelaziz, M.; Lodi-Rizzini, A.; Schürmann, F.; Krejci, I.; Markram, H. Caries Detection
with Near-Infrared Transillumination Using Deep Learning. J. Dent. Res. 2019, 98, 1227–1233.
https://doi.org/10.1177/0022034519871884.
30. Cantu, A.G.; Gehrung, S.; Krois, J.; Chaurasia, A.; Rossi, J.G.; Gaudin, R.; Elhennawy, K.; Schwendicke, F. Detecting Caries
Lesions of Different Radiographic Extension on Bitewings Using Deep Learning. J. Dent. 2020, 100, 103425.
https://doi.org/10.1016/J.JDENT.2020.103425.
31. Park, Y.H.; Kim, S.H.; Choi, Y.Y. Prediction Models of Early Childhood Caries Based on Machine Learning Algorithms. Int. J.
Environ. Res. Public Health 2021, 18, 8613. https://doi.org/10.3390/IJERPH18168613/S1.
32. Paniagua, B.; Shah, H.; Hernandez-Cerdan, P.; Budin, F.; Chittajallu, D.; Walter, R.; Mol, A.; Khan, A.; Vimort, J.-B. Automatic
Quantification Framework to Detect Cracks in Teeth. Proc. SPIE Int. Soc. Opt. Eng. 2018, 10578, 105781K.
https://doi.org/10.1117/12.2293603.
33. Fukuda, M.; Inamoto, K.; Shibata, N.; Ariji, Y.; Yanashita, Y.; Kutsuna, S.; Nakata, K.; Katsumata, A.; Fujita, H.; Ariji, E. Evalu-
ation of an Artificial Intelligence System for Detecting Vertical Root Fracture on Panoramic Radiography. Oral Radiol. 2019, 36,
337–343. https://doi.org/10.1007/S11282-019-00409-X.
34. Danks, R.P.; Bano, S.; Orishko, A.; Tan, H.J.; Moreno Sancho, F.; D’Aiuto, F.; Stoyanov, D. Automating Periodontal Bone Loss
Measurement via Dental Landmark Localisation. Int. J. Comput. Assist. Radiol. Surg. 2021, 16, 1189–1199.
https://doi.org/10.1007/S11548-021-02431-Z.
35. Tonetti, M.S.; Greenwell, H.; Kornman, K.S. Staging and Grading of Periodontitis: Framework and Proposal of a New Classifi-
cation and Case Definition. J. Periodontol. 2018, 89 (Suppl. 1), S159–S172. https://doi.org/10.1002/JPER.18-0006.
36. Chang, H.J.; Lee, S.J.; Yong, T.H.; Shin, N.Y.; Jang, B.G.; Kim, J.E.; Huh, K.H.; Lee, S.S.; Heo, M.S.; Choi, S.C.; et al. Deep Learning
Hybrid Method to Automatically Diagnose Periodontal Bone Loss and Stage Periodontitis. Sci. Rep. 2020, 10, 7531.
https://doi.org/10.1038/S41598-020-64509-Z.
37. Murata, M.; Ariji, Y.; Ohashi, Y.; Kawai, T.; Fukuda, M.; Funakoshi, T.; Kise, Y.; Nozawa, M.; Katsumata, A.; Fujita, H.; et al.
Deep-Learning Classification Using Convolutional Neural Network for Evaluation of Maxillary Sinusitis on Panoramic Radi-
ography. Oral Radiol. 2018, 35, 301–307. https://doi.org/10.1007/S11282-018-0363-7.
38. Kim, Y.; Lee, K.J.; Sunwoo, L.; Choi, D.; Nam, C.M.; Cho, J.; Kim, J.; Bae, Y.J.; Yoo, R.E.; Choi, B.S.; et al. Deep Learning in
Diagnosis of Maxillary Sinusitis Using Conventional Radiography. Investig. Radiol. 2019, 54, 7–15.
https://doi.org/10.1097/RLI.0000000000000503.
39. Kuwana, R.; Ariji, Y.; Fukuda, M.; Kise, Y.; Nozawa, M.; Kuwada, C.; Muramatsu, C.; Katsumata, A.; Fujita, H.; Ariji, E. Perfor-
mance of Deep Learning Object Detection Technology in the Detection and Diagnosis of Maxillary Sinus Lesions on Panoramic
Radiographs. Dentomaxillofacial Radiol. 2021, 50, 20200171. https://doi.org/10.1259/DMFR.20200171.
Diagnostics 2022, 12, 1029 14 of 15
40. Hung, K.F.; Ai, Q.Y.H.; King, A.D.; Bornstein, M.M.; Wong, L.M.; Leung, Y.Y. Automatic Detection and Segmentation of Mor-
phological Changes of the Maxillary Sinus Mucosa on Cone-Beam Computed Tomography Images Using a Three-Dimensional
Convolutional Neural Network. Clin. Oral Investig. 2022, online ahead of print. https://doi.org/10.1007/S00784-021-04365-X.
41. Kise, Y.; Ikeda, H.; Fujii, T.; Fukuda, M.; Ariji, Y.; Fujita, H.; Katsumata, A.; Ariji, E. Preliminary Study on the Application of
Deep Learning System to Diagnosis of Sjögren’s Syndrome on CT Images. Dentomaxillofacial Radiol. 2019, 48, 48.
https://doi.org/10.1259/DMFR.20190019.
42. López-Janeiro, Á.; Cabañuz, C.; Blasco-Santana, L.; Ruiz-Bravo, E. A Tree-Based Machine Learning Model to Approach Mor-
phologic Assessment of Malignant Salivary Gland Tumors. Ann. Diagn. Pathol. 2022, 56, 151869.
https://doi.org/10.1016/J.ANNDIAGPATH.2021.151869.
43. De Felice, F.; Valentini, V.; de Vincentiis, M.; di Gioia, C.R.T.; Musio, D.; Tummulo, A.A.; Ricci, L.I.; Converti, V.; Mezi, S.;
Messineo, D.; et al. Prediction of Recurrence by Machine Learning in Salivary Gland Cancer Patients After Adjuvant
(Chemo)Radiotherapy. Vivo 2021, 35, 3355–3360. https://doi.org/10.21873/INVIVO.12633.
44. Chiesa-Estomba, C.M.; Echaniz, O.; Sistiaga Suarez, J.A.; González-García, J.A.; Larruscain, E.; Altuna, X.; Medela, A.; Graña,
M. Machine Learning Models for Predicting Facial Nerve Palsy in Parotid Gland Surgery for Benign Tumors. J. Surg. Res. 2021,
262, 57–64. https://doi.org/10.1016/J.JSS.2020.12.053.
45. Radke, J.C.; Ketcham, R.; Glassman, B.; Kull, R.S. Artificial Neural Network Learns to Differentiate Normal TMJs and Non-
reducing Displaced Disks after Training on Incisor-Point Chewing Movements. Cranio J. Craniomandib. Pract. 2003, 21, 259–264.
https://doi.org/10.1080/08869634.2003.11746260.
46. Bas, B.; Ozgonenel, O.; Ozden, B.; Bekcioglu, B.; Bulut, E.; Kurt, M. Use of Artificial Neural Network in Differentiation of Sub-
groups of Temporomandibular Internal Derangements: A Preliminary Study. J. Oral Maxillofac. Surg. 2012, 70, 51–59.
https://doi.org/10.1016/J.JOMS.2011.03.069.
47. Iwasaki, H. Bayesian Belief Network Analysis Applied to Determine the Progression of Temporomandibular Disorders Using
MRI. Dentomaxillofacial Radiol. 2015, 44, 20140279. https://doi.org/10.1259/DMFR.20140279.
48. Choi, E.; Kim, D.; Lee, J.Y.; Park, H.K. Artificial Intelligence in Detecting Temporomandibular Joint Osteoarthritis on Orthopan-
tomogram. Sci. Rep. 2021, 11, 10246. https://doi.org/10.1038/S41598-021-89742-Y.
49. Orhan, K.; Driesen, L.; Shujaat, S.; Jacobs, R.; Chai, X. Development and Validation of a Magnetic Resonance Imaging-Based
Machine Learning Model for TMJ Pathologies. BioMed Res. Int. 2021, 2021, 6656773. https://doi.org/10.1155/2021/6656773.
50. de Lima, E.D.; Paulino, J.A.S.; de Farias Freitas, A.P.L.; Ferreira, J.E.V.; da Silva Barbosa, J.; Silva, D.F.B.; Bento, P.M.; Araújo
Maia Amorim, A.M.; Melo, D.P. Artificial Intelligence and Infrared Thermography as Auxiliary Tools in the Diagnosis of Tem-
poromandibular Disorder. Dentomaxillofacial Radiol. 2022, 51, 20210318. https://doi.org/10.1259/DMFR.20210318.
51. Taguchi, A.; Ohtsuka, M.; Nakamoto, T.; Naito, K.; Tsuda, M.; Kudo, Y.; Motoyama, E.; Suei, Y.; Tanimoto, K. Identification of
Post-Menopausal Women at Risk of Osteoporosis by Trained General Dental Practitioners Using Panoramic Radiographs. Den-
tomaxillofacial Radiol. 2014, 36, 149–154. https://doi.org/10.1259/DMFR/31116116.
52. Okabe, S.; Morimoto, Y.; Ansai, T.; Yoshioka, I.; Tanaka, T.; Taguchi, A.; Kito, S.; Wakasugi-Sato, N.; Oda, M.; Kuroiwa, H.; et
al. Assessment of the Relationship between the Mandibular Cortex on Panoramic Radiographs and the Risk of Bone Fracture
and Vascular Disease in 80-Year-Olds. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endod. 2008, 106, 433–442.
https://doi.org/10.1016/J.TRIPLEO.2007.09.013.
53. Klemetti, E.; Kolmakov, S.; Kröger, H. Pantomography in Assessment of the Osteoporosis Risk Group. Eur. J. Oral Sci. 1994, 102,
68–72. https://doi.org/10.1111/J.1600-0722.1994.TB01156.X.
54. Taguchi, A.; Suei, Y.; Ohtsuka, M.; Otani, K.; Tanimoto, K.; Ohtaki, M. Usefulness of Panoramic Radiography in the Diagnosis
of Postmenopausal Osteoporosis in Women. Width and Morphology of Inferior Cortex of the Mandible. Dentomaxillofacial Ra-
diol. 2014, 25, 263–267. https://doi.org/10.1259/DMFR.25.5.9161180.
55. Lee, J.-S.; Adhikari, S.; Liu, L.; Jeong, H.-G.; Kim, H.; Yoon, S.-J. Osteoporosis Detection in Panoramic Radiographs Using a Deep
Convolutional Neural Network-Based Computer-Assisted Diagnosis System: A Preliminary Study. Dentomaxillofacial Radiol.
2019, 48, 20170344. https://doi.org/10.1259/dmfr.20170344.
56. Lee, K.S.; Jung, S.K.; Ryu, J.J.; Shin, S.W.; Choi, J. Evaluation of Transfer Learning with Deep Convolutional Neural Networks
for Screening Osteoporosis in Dental Panoramic Radiographs. J. Clin. Med. 2020, 9, 392. https://doi.org/10.3390/JCM9020392.
57. Cancer. Available online: https://www.who.int/news-room/fact-sheets/detail/cancer (accessed on 31 January 2022).
58. Nayak, G.S.; Kamath, S.; Pai, K.M.; Sarkar, A.; Ray, S.; Kurien, J.; D’Almeida, L.; Krishnanand, B.R.; Santhosh, C.; Kartha, V.B.;
et al. Principal Component Analysis and Artificial Neural Network Analysis of Oral Tissue Fluorescence Spectra: Classification
of Normal Premalignant and Malignant Pathological Conditions. Biopolymers 2006, 82, 152–166.
https://doi.org/10.1002/BIP.20473.
59. Uthoff, R.D.; Song, B.; Sunny, S.; Patrick, S.; Suresh, A.; Kolur, T.; Keerthi, G.; Spires, O.; Anbarani, A.; Wilder-Smith, P.; et al.
Point-of-Care, Smartphone-Based, Dual-Modality, Dual-View, Oral Cancer Screening Device with Neural Network Classifica-
tion for Low-Resource Communities. PLoS ONE 2018, 13, e0207493. https://doi.org/10.1371/JOURNAL.PONE.0207493.
60. Aubreville, M.; Knipfer, C.; Oetter, N.; Jaremenko, C.; Rodner, E.; Denzler, J.; Bohr, C.; Neumann, H.; Stelzle, F.; Maier, A.
Automatic Classification of Cancerous Tissue in Laserendomicroscopy Images of the Oral Cavity Using Deep Learning. Sci.
Rep. 2017, 7, 11979. https://doi.org/10.1038/s41598-017-12320-8.
61. Shams, W.K.; Htike, Z.Z. Oral Cancer Prediction Using Gene Expression Profiling and Machine Learning. Int. J. Appl. Eng. Res.
2017, 12, 4893–4898.
Diagnostics 2022, 12, 1029 15 of 15
62. Jeyaraj, P.R.; Samuel Nadar, E.R. Computer-Assisted Medical Image Classification for Early Diagnosis of Oral Cancer Employ-
ing Deep Learning Algorithm. J. Cancer Res. Clin. Oncol. 2019, 145, 829–837. https://doi.org/10.1007/S00432-018-02834-7.
63. Kim, D.W.; Lee, S.; Kwon, S.; Nam, W.; Cha, I.H.; Kim, H.J. Deep Learning-Based Survival Prediction of Oral Cancer Patients.
Sci. Rep. 2019, 9, 6994. https://doi.org/10.1038/S41598-019-43372-7.
64. Alabi, R.O.; Elmusrati, M.; Sawazaki-Calone, I.; Kowalski, L.P.; Haglund, C.; Coletta, R.D.; Mäkitie, A.A.; Salo, T.; Almangush,
A.; Leivo, I. Comparison of Supervised Machine Learning Classification Techniques in Prediction of Locoregional Recurrences
in Early Oral Tongue Cancer. Int. J. Med. Inform. 2020, 136, 104068. https://doi.org/10.1016/J.IJMEDINF.2019.104068.
65. Alhazmi, A.; Alhazmi, Y.; Makrami, A.; Masmali, A.; Salawi, N.; Masmali, K.; Patil, S. Application of Artificial Intelligence and
Machine Learning for Prediction of Oral Cancer Risk. J. Oral Pathol. Med. 2021, 50, 444–450. https://doi.org/10.1111/JOP.13157.
66. Chu, C.S.; Lee, N.P.; Adeoye, J.; Thomson, P.; Choi, S.W. Machine Learning and Treatment Outcome Prediction for Oral Cancer.
J. Oral Pathol. Med. 2020, 49, 977–985. https://doi.org/10.1111/JOP.13089.
67. Kirubabai, M.P.; Arumugam, G. Deep Learning Classification Method to Detect and Diagnose the Cancer Regions in Oral MRI
Images. Med. Leg. Update 2021, 21, 462–468. https://doi.org/10.37506/MLU.V21I1.2353.
68. Ariji, Y.; Fukuda, M.; Kise, Y.; Nozawa, M.; Yanashita, Y.; Fujita, H.; Katsumata, A.; Ariji, E. Contrast-Enhanced Computed
Tomography Image Assessment of Cervical Lymph Node Metastasis in Patients with Oral Cancer by Using a Deep Learning
System of Artificial Intelligence. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. 2019, 127, 458–463.
https://doi.org/10.1016/J.OOOO.2018.10.002.
69. Ariji, Y.; Sugita, Y.; Nagao, T.; Nakayama, A.; Fukuda, M.; Kise, Y.; Nozawa, M.; Nishiyama, M.; Katumata, A.; Ariji, E. CT
Evaluation of Extranodal Extension of Cervical Lymph Node Metastases in Patients with Oral Squamous Cell Carcinoma Using
Deep Learning Classification. Oral Radiol. 2019, 36, 148–155. https://doi.org/10.1007/S11282-019-00391-4.
70. General Data Protection Regulation (GDPR)—Official Legal Text. Available online: https://gdpr-info.eu/ (accessed on 28 March
2022).
71. Hulsen, T. Sharing Is Caring—Data Sharing Initiatives in Healthcare. Int. J. Environ. Res. Public Health 2020, 17, 3046.
https://doi.org/10.3390/IJERPH17093046.
72. Sun, C.; Ippel, L.; van Soest, J.; Wouters, B.; Malic, A.; Adekunle, O.; van den Berg, B.; Mussmann, O.; Koster, A.; van der Kallen,
C.; et al. A Privacy-Preserving Infrastructure for Analyzing Personal Health Data in a Vertically Partitioned Scenario. Stud.
Health Technol. Inform. 2019, 264, 373–377. https://doi.org/10.3233/SHTI190246.
73. Gaye, A.; Marcon, Y.; Isaeva, J.; Laflamme, P.; Turner, A.; Jones, E.M.; Minion, J.; Boyd, A.W.; Newby, C.J.; Nuotio, M.L.; et al.
DataSHIELD: Taking the Analysis to the Data, Not the Data to the Analysis. Int. J. Epidemiol. 2014, 43, 1929–1944.
https://doi.org/10.1093/IJE/DYU188.
74. Schwendicke, F.; Samek, W.; Krois, J. Artificial Intelligence in Dentistry: Chances and Challenges. J. Dent. Res. 2020, 99, 769–774.
https://doi.org/10.1177/0022034520915714.
75. Rodrigues, J.A.; Krois, J.; Schwendicke, F. Demystifying Artificial Intelligence and Deep Learning in Dentistry. Braz. Oral Res.
2021, 35, 1–7. https://doi.org/10.1590/1807-3107BOR-2021.VOL35.0094.
76. MacHoy, M.E.; Szyszka-Sommerfeld, L.; Vegh, A.; Gedrange, T.; Woźniak, K. The Ways of Using Machine Learning in Dentis-
try. Adv. Clin. Exp. Med. 2020, 29, 375–384. https://doi.org/10.17219/acem/115083.
