Arti Ficial Intelligence Applications in Implant Dentistry: A Systematic Review
Arti Ficial Intelligence Applications in Implant Dentistry: A Systematic Review
Arti Ficial Intelligence Applications in Implant Dentistry: A Systematic Review
a
Assistant Professor and Assistant Program Director AEGD Residency, Department of Comprehensive Dentistry, College of Dentistry, Texas A&M University, Dallas, Texas;
and Affiliate Faculty Graduate Prosthodontics, Department of Restorative Dentistry, School of Dentistry, University of Washington, Seattle, Wash; and Researcher at Revilla
Research Center, Madrid, Spain.
b
Associate Professor Department of Conservative Dentistry and Prosthodontics, School of Dentistry, Complutense University of Madrid, Madrid, Spain.
c
Graduate Research Assistant, J. Mike Walker ’66 Department of Mechanical Engineering, Texas A&M University, College Station, Texas.
d
Assistant Professor Clinical Research and Biostatistics, Eastman Institute of Oral Health, University of Rochester Medical Center, Rochester, NY.
e
Raymond J. and Elva Pomfret Nagle Associate Professor of Restorative Dentistry and Biomaterials Sciences and Chair of the Department of Restorative Dentistry and
Biomaterials Science, Harvard School of Dental Medicine, Boston, Mass.
f
Professor and Chair Department of Prosthodontics, Tufts University School of Dental Medicine, Boston, Mass.
g
Assistant Professor, J. Mike Walker ’66 Department of Mechanical Engineering, Texas A&M University, College Station, Texas.
Table 2. Joanna Briggs Institute Critical Appraisal Checklist for Quasi-Experimental Studies (nonrandomized experimental studies)
Question Answer
1 Is it clear in the study what is the ‘cause’ and what Yes, No, Unclear, or Not
is the ‘effect’ (that is, there is no confusion about which variable comes first)? applicable
2 Were the participants included in any similar comparisons?
3 Were the participants included in any comparisons receiving similar treatment/care other than the exposure or intervention of
interest?
4 Was there a control group?
5 Were there multiple measurements of the outcome both before and after the intervention/exposure?
6 Was follow-up complete and, if not, were differences between groups in terms of their follow-up adequately described and
analyzed?
7 Were the outcomes of participants included in any comparisons measured in the same way?
8 Were outcomes measured in a reliable way?
9 Was appropriate statistical analysis used?
assessed by applying the Joanna Briggs Institute (JBI) risk factors and ontology criteria (Supplementary Table 2,
Critical Appraisal Checklist for Quasi-Experimental available online),33-39 and implant design optimization by
Studies (nonrandomized experimental studies) combining FEA calculations and AI models
(Table 2).26 Similarly, the third examiner (V.R.K.) was (Supplementary Table 2, available online).40-42
consulted to resolve lack of consensus. The overall accuracy outcome of the AI models
developed in the different reviewed studies ranged
RESULTS from 93.8% to 98%.13,27-32 The AI models to predict
osteointegration or implant success by using different
The search strategies yielded 207 studies. A total of 21 input data varied among the studies ranging from
duplicates were found. The 186 remaining articles were 62.4% to 80.5%.33-39 Finally, the studies that developed
evaluated by the titles and abstracts. Twenty-six articles AI models to optimize implant designs seem to agree
were identified for full-text revision. Nine articles were on the applicability of AI models to improve implant
excluded after full-text review, 2 excluded articles designs, minimizing the stress at the implant-bone
exposed a conceptual methodology, 1 applied AI models interface by 36.6% compared with the FEA model,40
for implant placement accuracy improvement evaluation optimizing the implant design porosity, length, and
by using robotics, 2 articles did not describe the AI diameter, improving the FEA calculations,41 or accu-
model, 1 study applied AI models to improve the data rately determining the elastic modulus of the implant-
search on systematic reviews, 2 studies used an AI model bone interface.42
to predict future developments by clustering patents and With respect to the selection of articles by reviewing
clinical implant studies, and 1 investigation was not their titles and abstracts, there was significant agreement
related to AI (Fig. 1). between the 2 investigators for the articles that were
Seventeen articles published between 2005 and 2020 selected (Cohen Kappa value=0.97, P<.001) and the ar-
were included in the present investigation (Fig. 2). The AI ticles that were not selected (Cohen Kappa value=0.97,
models used among the different studies are presented in P<.001). With respect to the selection of articles by
Table 3. The selected articles were distributed into 3 reviewing their full text, there was a significant agree-
groups depending on the application of the AI model: ment between the 2 investigators for the articles that
implant type recognition (Supplementary Table 1, avail- were selected (Cohen Kappa value=1, P<.001) and the
able online),13,27-32 models to determine osteointegration articles that were not selected (Cohen Kappa value=1,
success or implant success prediction by using patient P<.001).
Identification
207 studies identified data base searching
Pubmed: 43
Embase: 48
21 duplicated eliminated:
Web of science: 48
- Pubmed: 0
Scopus: 61
- Embase: 0
Cochrane: 5
- Web of science: 16
Hand search: 2
- Scopus: 5
- Cochrane: 0
- Hand search: 0
Screening
185 studies identified data base searching 160 studies excluded due to the following:
Pubmed: 43 - 11 AI studies related to another disciplines,
Embase: 48 not dentistry related.
Web of science: 32 - 49 AI studies related to another dental
Scopus: 56 disciplines, not following the inclusive criteria.
Cochrane: 5 Such as endodontics, maxillofacial surgery,
Hand search: 1 radiology, orthodontics, pediatric dentistry,
Periodontology, and restorative dentistry.
- 8 AI for enhancement of radiographic images
Eligibility
- 9 robotics in dentistry
26 studies of full text assessed for eligibility - 19 AI for segmentation
Pubmed: 6 - 2 age estimation models
Embase: 1 - 1 augmented reality
Web of science: 5 - 59 dental studies, not related with AI.
Scopus: 12 - 1 poster
Cochrane: 0 - 1 letter to editor
Hand search: 2
9 studies were excluded due to the following
criteria:
Included - 4 AI model not described
- 1 AI model for robotics
- 2 conceptual description of AI method
- 1 AI model to improve systematic reviews
17 studies included in systematic review search
- 1 AI not for implant recognition, prediction
of implant success or optimization of implant
designs
Figure 1. PRISMA flow diagram with information through phases of study selection.
Article Year
7
6
6
5
Num. Articles
3
2 2
2
1 1 1 1 1 1 1
1
0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Figure 2. Number of included articles by year and purpose of artificial intelligence model.
The JBI Critical Appraisal Checklist for Quasi- with the prosthetic part was visible on the radiographical
Experimental results showed a 100% low risk of bias images; therefore, comparisons among the different
in all included articles for question 8. For question 1, all studies are difficult.
the studies obtained a low risk of bias except for that of All the reviewed studies that developed AI models for
Morais et al13 that obtained a high risk of bias. For implant type recognition used 2-dimensional (2D) radi-
question 4, all the investigations attained a low risk of ography as the input data set. A diversity of deep CNN
bias except for those of Hadj Saïd et al,27 Papantono- architectures has been dedicated and trained by using 3-
poulos et al,36 Zaw et al,42 Sukegawa et al,29 Ha et al,34 dimensional (3D) computed tomographic images.43,44
Morais et al,13 Takahashi et al,32 and Kim et al30 that Two-dimensional images including periapical and
showed a high risk of bias. For question 9, all the panoramic radiographs are more distorted than 3D scans.
studies displayed a low risk of bias except for that of Even though clinicians normally obtain periapical radio-
Zaw et al42 that did not have statistical analysis. As no graphs for the radiographic evaluation of dental implants,
specific in vitro study quality assessment tool has been the inclusion of cone beam computed tomography
developed, questions 2 and 6 of the JBI were not (CBCT) images might aid in the AI development for the
applicable in this systematic review. Questions 3, 5, and recognition of dental implant types. All the included
7 were not applicable for any of the included studies studies used CBCT images to develop the AI model.
(Fig. 3). Considering the broad implant types available in
the market,10 limited implant types were analyzed in
different reviewed studies. Furthermore, implant de-
DISCUSSION
signs can be different from each other, facilitating AI
The number of publications that use AI models for recognition among the different implant types analyzed
implant dentistry applications has risen notably since in a study, while other implant designs are similar,
2018. The year ranged from 2005 to 2020, with very few which may require a data base large enough to train
publications before 2005. the AI model to differentiate them. However, the
A total of 7 included studies developed AI models for overall accuracy outcome of the AI models developed
implant type recognition. Except 1 study that used in the different reviewed studies ranged from 93.8% to
regression analysis k-nearest neighbors (k-NN), all the 98%.13,27-32
studies selected developed a convolutional neural Lee and Jeong28 used a data set of 10 770 radio-
network (CNN) d a deep neural network algorithm for graphic images from 3 different implant types to train a
image recognition and classification by using as an input deep CNN model. The authors compared the implant
radiographical data such as periapical27,28,30,31 and recognition capabilities of the examiners (board-certified
panoramic images -,27-29,31,32 or the type of radio- periodontists and the AI model) and of the radio-
graphical data was not provided.13 The efficacy compar- graphical image used: periapical, panoramic, or both
isons among the different AI models used are difficult images. Implant recognition accuracy varied among the
because of the data input or methods used on the studies 3 types of implants tested, but higher specificity and
reviewed. While each study attempted to standardize the sensitivity were found when both periapical and pano-
collection of the radiographical data set, differences ramic images were used for both the AI model and the
among the studies were identified, including projection periodontists.
geometry, exposure factors, film contrast, and film speed. While in other medicine specialties different regis-
Furthermore, variations on the radiographic information tering strategies have developed orthopedic records,45
differed among the reviewed studies where the implant one of the current limitations in implant dentistry is the
by itself (with a cover screw or a healing abutment) or absence of available data records which can facilitate AI
D1 D4 D8 D9 Overall
Hadj Saïd et al, 2020 + × + + –
Lee et al, 202028 + + + + +
Li et al, 2019 + + + + +
Papantonopoulos et al, 2017 + × + + –
Zaw et al, 2009 + × + ? –
Zhang et al, 2020 + + + + +
Q1
Sukegawa et al, 2020 + × + + – Q4
Q8
Ha et al, 2018 + × + + – Q9
Overall
Roy et al, 2018 + + + + + 0% 25% 50% 75% 100%
Morais et al, 2015 × × + + – Low Unclear High No information
Oliveira et al, 200538 + + + + +
Oliveira et al, 200539 + + + + +
Moayeri et al, 2016 + + + + +
Lee et al, 202031 + + + + + Judgement
D1: Q1 + Low
Takahashiet al, 2020 + × + + –
D4: Q4 – Unclear
Kim et al, 2020 + × + + – D8: Q8 × High
Liu et al, 2018 + + + + + D9: Q9
Figure 3. Joanna Briggs Institute JBI Critical Appraisal Checklist for Quasi-Experimental evaluation.
model development and training on implant recognition. not include the implant type used, a definition of implant
However, the clinical applicability of such an AI appli- success, implant prosthesis design, and genetic, immu-
cation would help clinicians trying to restore an unknown nological, or microbiological variables, which might have
implant. Furthermore, clinicians that have less clinical impacted the results. Because of methodological dis-
experience in implant dentistry may obtain assistance by crepancies, comparisons among the different studies
using an implant recognition software program. were not feasible.
A total of 7 included studies aimed to develop AI Papantonopoulos at al36 aimed to cluster de-
models to predict implant success by using a broader mographic, clinical, and radiographic data from 72 pa-
variety of AI models compared with the implant recog- tients with 237 implants and recognize potential implant
nition AI application. The main AI models used were “phenotypes” and forecasters of bone levels around
regression analysis (support vector machine classifica- implants. The AI model produced an implant map
tion), decision tree learning, logistic regression, and establishing the existence of 2 distinct implant clusters,
classifier neural network.33-39 However, because of a high which the authors identified as 2 possible types of
variation in the methodologies among the different implant “phenotypes,” namely implant phenotype with
studies, comparisons among the obtained results are susceptibility or resistance to peri-implantitis. The inter-
difficult. pretation of the data is interesting, as the AI model was
Prediction models are based on clustering data and developed by using the data obtained from 1 private
investigating the structural properties of the data network practice which might not represent the general popula-
generated by intricate relations of demographic, radio- tion. The limited data and measurements collected by 1
graphical, and clinical variables. Therefore, the prediction periodontist, the restricted patient follow-up period of 2
AI algorithm is assembled based on the input data pro- years, or implants placed with bone grafting procedures
vided. Most of the included studies used demographic were excluded. The difficulty of obtaining data to develop
data, physical and intraoral conditions, lifestyle, anatomic and train AI models is a challenge for researchers,
condition of the area receiving the implant, implant limiting the faster development of AI models in implant
placement with or without bone grafting procedures, dentistry.
bone levels around the implant measured by using per- Three included studies applied AI models for implant
iapical radiographs, or characteristics of the prosthesis as design optimization by using finite element analysis
an input. Furthermore, most of the reviewed studies did (FEA) methods.40-42 Li et al40 replaced the FEA model
with an AI algorithm to compute the stress at the 4. The studies that developed AI models to optimize
implant-bone interface by considering 3 implant design implant designs seem to agree on the applicability of
variables, namely the implant length, thread length, and AI models to improve implant designs, minimizing
thread pitch. The AI model sought to optimize the the stress at the implant-bone interface by 36.6%
implant design variables to minimize the stress at the compared with the FEA model, optimizing the
implant-bone interface. The results of this study showed implant design porosity, length, and diameter,
a reduction of 36.6% of the stress at the implant-bone improving the FEA calculations, or accurately
interface compared with the FEA model. Roy et al41 determining the elastic modulus of the implant-
aimed to optimize the implant design porosity, length, bone interface.
and diameter by using an artificial neural network
(ANN) combined with genetic algorithms by substitut- REFERENCES
ing the FEA calculations. Similarly, Zaw et al42 devel-
1. International Organization for Standardization ISO/IEC 2382:2015(en): In-
oped a reduced-basis method of creating the reactions formation technology e Vocabulary. Available at: https://www.iso.org/
of the dental implant-bone system to develop a neural standard/63598.html. (Accessed July 2020).
2. International Organization for Standardization ISO/IEC TR 24028:2020. In-
network architecture. The proposed AI algorithm was formation technology - Artificial intelligence - Overview of trustworthiness in
capable of accurately computing the elastic modulus of artificial intelligence. Available at: https://www.iso.org/standard/77608.html?
browse=tc. (Accessed July 2020).
the implant-bone interface. All the studies agreed on 3. Wooldridge MJ, Jennings NR. Intelligent agents: Theory and practice. Knowl
the applicability of AI models to optimize implant de- Eng Rev 1995;10:115-52.
4. Das K, Behera RN. A survey on machine learning: Concept, algorithms and
signs; however, further investigations are needed to applications. IJIRCCE 2017;5:1301-9.
improve the AI calculations on implant design and 5. Alpaydin E. Introduction to machine learning. 4th Edition. Massachusetts
Institute of Technology; 2020. p. 23-491.
evaluating the outcomes in in vitro, animal, and clinical 6. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature 2015;521:436-44.
studies. 7. Legg S, Hutter M. Universal intelligence: A definition of machine intelli-
gence. Minds & Machines 2007;17:391-444.
Future directions in implant dentistry could combine 8. Park WJ, Park JB. History and application of artificial neural networks in
CBCT scans with radiographic image data to aid in data dentistry. Eur J Dent 2018;12:594-601.
9. El-Hassoun O, Maruscakova L, Valaskova Z, Bucova M, Polak S, Hulin I.
analysis and increase the accuracy of implant type Artificial intelligence in service of medicine. Bratisl Lek Listy 2019;120:218-22.
recognition. The implementation of a special class of 10. Jokstad A, Braegger U, Brunski JB, Carr AB, Naert I, Wennerberg A. Quality
of dental implants. Int Dent J 2003;53:409-43.
deep learning methods such as 1-shot learning and less- 11. Sahiwal IG, Woody RD, Benson BW, Guillen GE. Radiographic identification
than-1-shot learning that require fewer data points than of nonthreaded endosseous dental implants. J Prosthet Dent 2002;87:
552-62.
neural network models might facilitate the implementa- 12. Michelinakis G, Sharrock A, Barclay CW. Identification of dental implants
tion and improvement of AI models for implant dentistry through the use of implant recognition software (IRS). Int Dent J 2006;56:
203-8.
applications. Furthermore, the standardization and 13. Morais P, Queirós S, Moreira AHJ, Ferreira A, Ferreira E, Duque D, et al.
benchmarking of data sets might increase the accuracy of Computer-aided recognition of dental implants in X-ray images. Proc SPIE 9414,
Medical Imaging 2015: Computer-Aided Diagnosis, 94142E 20 March 2015.
AI models in identifying implant type or predicting 14. Hung K, Montalvao C, Tanaka R, Kawai T, Bornstein MM. The use and
implant success. The availability of open data sets will performance of artificial intelligence applications in dental and maxillofacial
radiology: A systematic review. Dentomaxillofac Radiol 2020;49:1.
promote the growth of AI models. 15. Chang HJ, Lee SJ, Yong TH, Shin NY, Jang BG, Kim JE, et al. Deep learning
hybrid method to automatically diagnose periodontal bone loss and stage
periodontitis. Sci Rep 2020;5:7531.
CONCLUSIONS 16. Li H, Zhou J, Zhou Y, Chen J, Gao F, Xu Y, et al. Automatic and interpretable
model for periodontitis diagnosis in panoramic radiographsMartel ALet al.,
Based on the findings of this systematic review, the editors. Medical image computing and computer assisted intervention e
subsequent conclusions were drawn: MICCAI 2020. MICCAI 2020. Lecture Notes in Computer Science. 12262.
Springer; 2020.
1. AI models have the potential to recognize implant type, 17. Feres M, Louzoun Y, Haber S, Faveri M, Figueiredo LC, Levin L. Support
vector machine-based differentiation between aggressive and chronic peri-
predict implant success by using patient risk factors and odontitis using microbial profiles. Int Dent J 2018;68:39-46.
ontology criteria, and optimize implant designs, but 18. Pitts NB. Detection and measurement of approximal radiolucencies by
computer-aided image analysis. Oral Surg Oral Med Oral Pathol 1984;58:
they are still being developed. With its application in 358-66.
implant dentistry rapidly expanding, the effectiveness 19. Heaven TJ, Firestone AR, Feagin FF. Computer-based image analysis of
natural approximal caries on radiographic films. J Dent Res 1992;71:
and reliability of AI models should be evaluated before 846-9.
recommending them for clinical practice. 20. Duncan RC, Heaven T, Weems RA, Firestone AR, Greer DF, Patel JR. Using
computers to diagnose and plan treatment of approximal caries. Detected in
2. Based on the reviewed studies, the AI models radiographs. J Am Dent Assoc 1995;126:873-82.
developed to recognize implant type by using 21. Firestone AR, Sema D, Heaven TJ, Weems RA. The effect of a knowledge-
based, image analysis and clinical decision support system on observer
radiographical images were the more developed performance in the diagnosis of approximal caries from radiographic images.
application of AI in implant dentistry, obtaining an Caries Res 1998;32:127-34.
22. Wenzel A, Hintze H, Kold LM, Kold S. Accuracy of computer-automated
overall accuracy ranging from 93.8% to 98%. caries detection in digital radiographs compared with human observers. Eur J
3. The AI models to predict osteointegration or Oral Sci 2002;110:199-203.
23. Devito KL, de Souza Barbosa F, Felippe Filho WN. An artificial multilayer
implant success varied from 62.4% to 80.5% among perceptron neural network for diagnosis of proximal dental caries. Oral Surg
the studies. Oral Med Oral Pathol Oral Radiol Endod 2008;106:879-84.
24. Hung K, Yeung AWK, Tanaka R, Bornstein MM. Current applications, op- prediction of success of dental implants. In: Sanfeliu A, Cortés ML, edi-
portunities, and limitations of AI for 3D imaging in dental research and tors. Progress in pattern recognition, image analysis and applications.
practice. Int J Environ Res Public Health 2020;17:4424. Lecture Notes in Computer Science3773. Berlin, Heidelberg: Springer;
25. Moher D, Liberati A, Tetzlaff J, Altman DG; The PRISMA Group. Preferred 2005. p. 1015-26. CIARP.
reporting items for systematic reviews and meta-analyses: The PRISMA 39. Oliveira ALI, Baldisserotto C, Baldisserotto J. A comparative study on ma-
statement. PLoS Med 2009;6:e1000097. chine learning techniques for prediction of success of dental implants. In:
26. The Joanna Briggs Institute (JBI). Critical appraisal checklist for quasi- Proceedings of the 4th Mexican international conference on advances in
experimental studies (non-randomized experimental studies). Available at: artificial intelligence. Berlin, Heidelberg, MICAI: Springer-Verlag; 2005. p.
https://joannabriggs.org/sites/default/files/2019-05/JBI_Quasi-Experimental_ 939-48.
Appraisal_Tool2017_0.pdf. Accessed April 30, 2020. 40. Li H, Shi M, Liu X, Shi Y. Uncertainty optimization of dental implant based
27. Hadj Saïd M, Le Roux MK, Catherine JH, Lan R. Development of an artificial on finite element method, global sensitivity analysis and support vector
intelligence model to identify a dental implant from a radiograph. Int J Oral regression. Proc Inst Mech Eng H 2019;233:232-43.
Maxillofac Implants 2020;36:1077-82. 41. Roy S, Dey S, Khutia N, Roy Chowdhury A, Datta S. Design of patient
28. Lee JH, Jeong SN. Efficacy of deep convolutional neural network algorithm specific dental implant using FE analysis and computational intelligence
for the identification and classification of dental implant systems, using techniques. Appl Soft Comput 2018;65:272-9.
panoramic and periapical radiographs: A pilot study. Medicine (Baltimore) 42. Zaw K, Liu GR, Deng B, Tan KB. Rapid identification of elastic modulus of
2020;99:e20787. the interface tissue on dental implants surfaces using reduced-basis method
29. Sukegawa S, Yoshii K, Hara T, Yamashita K, Nakano K, Yamamoto N, et al. and a neural network. J Biomech 2009;42:634-41.
Deep neural networks for dental implant system classification. Biomolecules 43. Jaskari J, Sahlsten J, Järnstedt J, Mehtonen H, Karhu K, Sundqvist O,
2020;10:984. et al. Deep learning method for mandibular canal segmentation in
30. Kim JE, Nam NE, Shim JS, Jung YH, Cho BH, Hwang JJ. Transfer learning via dental cone beam computed tomography volumes. Sci Rep 2020;10:
deep neural networks for implant fixture system classification using periapical 5842.
radiographs. J Clin Med 2020;9:1117. 44. Yang H, Liang K, Zhang L, Kang K, Xing Y. Improve 3D cone-beam CT
31. Lee JH, Kim YT, Lee JB, Jeong SN. A Performance comparison between reconstruction by slice-wise deep learning. In: 2018 IEEE Nuclear Science
automated deep learning and dental professionals in classification of dental Symposium and Medical Imaging Conference Proceedings. Sydney,
implant systems from dental imaging: A multi-center study. Diagnostics Australia: NSS/MIC; 2018. p. 1-3.
(Basel) 2020;10:910. 45. Lübbeke A, Silman AJ, Barea C, Prieto-Alhambra D, Carr AJ. Mapping
32. Takahashi T, Nozaki K, Gonda T, Mameno T, Wada M, Ikebe K. Identifi- existing hip and knee replacement registries in Europe. Health Policy
cation of dental implants using deep learning-pilot study. Int J Implant Dent 2008;122:548-57.
2020;6:53.
33. Zhang H, Shan J, Zhang P, Chen X, Jiang H. Trabeculae microstructure Corresponding author:
parameters serve as effective predictors for marginal bone loss of dental Dr Miguel Gómez-Polo
implant in the mandible. Sci Rep 2020;10:18437. Department of Conservative Dentistry and Prosthodontics
34. Ha SR, Park HS, Kim EH, Kim HK, Yang JY, Heo J, et al. A pilot study using School of Dentistry
machine learning methods about factors influencing prognosis of dental Complutense University of Madrid
implants. J Adv Prosthodont 2018;10:395-400. 28040 Madrid
35. Liu C, Lin CH, Hu YH, You ZH. Predicting the failure of dental implants SPAIN
using supervised learning techniques. Appl Sci 2018;8:698. Email: mgomezpo@ucm.es
36. Papantonopoulos G, Gogos C, Housos E, Bountis T, Loos BG. Prediction of
individual implant bone levels and the existence of implant “phenotypes”.
Clin Oral Implants Res 2017;28:823-32. Copyright © 2021 by the Editorial Council for The Journal of Prosthetic Dentistry.
37. Moayeri RS, Mehdi K, Mahsa N. A hybrid method to predict success of dental This is an open access article under the CC BY-NC-ND license (http://
implants. IJACSA 2016;7:2016. creativecommons.org/licenses/by-nc-nd/4.0/).
38. Oliveira ALI, Baldisserotto C, Baldisserotto J. A comparative study on https://doi.org/10.1016/j.prosdent.2021.05.008
support vector machine and constructive RBF neural network for