Artificial intelligence applications in restorative dentistry: A systematic review

Intraoral scanners as tracking devices: A dental protocol for assessing volumetric changes between intraoral scans

Intraoral scanners (IOSs) are digital data acquisition technologies that ease the recording of virtual diagnostic casts. Some IOSs have a specific software tool to assess volumetric changes between 2 scans acquired on the patient at different times. The scans are superimposed and volumetric differences between both meshes are reported. However, these software tools may be limited to scans captured only by the IOS of the same manufacturer. The present manuscript describes a protocol for comparing volumetric changes between 2 scans recorded using any IOS. Additionally, 1 of the scans is divided into 3 sections to minimize the alignment distortion and maximize the evaluation of the volumetric changes.

Artificial intelligence-assisted full-mouth radiograph mounting in dental education

OBJECTIVES

With the increasing prevalence of artificial intelligence (AI) and the significant research gap in the application of AI within dentistry, this study aimed to (1) evaluate the efficiency and accuracy of dental students in full-mouth radiograph series (FMS) mounting with and without AI assistance, and (2) assess dental students' perceptions of AI in clinical education to address the impact of AI in dental education.

METHODS

An AI-based interface for mounting radiographs on FMS templates was designed and implemented in the study. Forty third-year dental students were randomly assigned to control and test groups. The control group manually mounted FMS radiographs, while the test group reviewed AI-pre-mounted radiographs for adjustments. Students' performance in efficiency and accuracy was evaluated. Pre- and post-study surveys were conducted to gauge students' confidence levels and opinions regarding the usefulness of the AI-assisted program.

RESULTS

The test group (using AI) demonstrated significantly faster radiograph mounting times than the control group (manual) (p < 0.05). Accuracy was lower in the test groups, when comparing AI-assisted and manual mounting of FMS (p < 0.01). Self-confidence and confidence in AI were consistent between the control and test groups, both before and after the study.

CONCLUSION

Students with AI presented with a decreased accuracy in FMS radiograph mounting. Therefore, AI automation could potentially have negative impacts in a learning environment with inexperienced clinicians.

Artificial intelligence systems in dental shade-matching: A systematic review

PURPOSE

Uses for artificial intelligence (AI) are being explored in contemporary dentistry, but artificial intelligence in dental shade-matching has not been systematically reviewed and evaluated. The purpose of this systematic review was to evaluate the accuracy of artificial intelligence in predicting dental shades in restorative dentistry.

METHODS

A systematic electronic search was performed with the databases MEDLINE (PubMed), Scopus, Cochrane Library, and Google Scholar. A manual search was also conducted. All titles and abstracts were subject to the inclusion criteria of observational, interventional studies, and studies published in the English language. Narrative reviews, systematic reviews, case reports, case series, letters to the editor, commentaries, studies that were not AI-based, studies that were not related to dentistry, and studies that were related to other disciplines in dentistry, other than restorative dentistry (prosthodontics and endodontics) were excluded. Two investigators independently evaluated the quality assessment of the studies by applying the Joanna Briggs Institute Critical Appraisal Checklist for Quasi-Experimental Studies (non-randomized experimental studies). A third investigator was consulted to resolve the lack of consensus.

RESULTS

Fifty-three articles were initially found from all the searches combined from articles published from 2008 till March 2023. A total of 15 articles met the inclusion criteria and were included in the systematic review. AI algorithms for shade-matching include fuzzy logic, a genetic algorithm with back-propagation neural network, back-propagation neural networks, convolutional neural networks, artificial neural networks, support vector machine algorithms, K-nearest neighbor with decision tree and random forest, deep learning for detection of dental prostheses based on object-detection applications, You Only Look Once-YOLO. Moment invariant was used for feature extraction.  XG (Xtreme Gradient) Boost was used in one study as a gradient-boosting machine learning algorithm. The highest accuracy in the prediction of dental shades was the decision tree regression model for leucite-based dental ceramics of 99.7% followed by the fuzzy decision of 99.62%, and support vector machine using cross-validation of 97%.

CONCLUSIONS

Lighting conditions, shade-matching devices and color space models, and the type of AI algorithm influence the accuracy of the prediction of dental shades. Knowledge-based systems and neural networks have shown better accuracy in predicting dental shades.

Changes of cardiac function: cardiac adaptation in patients with hypothyroidism assessed by cardiac magnetic resonance-a meta-analysis

Objective

The meta-analysis aimed to explore the cardiac adaptation in hypothyroidism patients by cardiac magnetic resonance.

Research methods and procedures

Databases including PubMed, Cochrane Library, Embase, CNKI, and Sinomed for clinical studies of hypothyroidism on cardiac function changes. Databases were searched from the earliest data to 15 June 2023. Two authors retrieved studies and evaluated their quality. Review Manager 5.4.1 and Stata18 were used to analyze the data. This study is registered with the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY), 202440114.

Results

Six studies were selected for further analysis. Five of them reported differences in cardiac function measures between patients with hypothyroidism and healthy controls, and three studies reported cardiac function parameters after treatment in patients with hypothyroidism. The fixed-effect model combined WMD values for left ventricular ejection fraction (LVEF) had a pooled effect size of -1.98 (95% CI -3.50 to -0.44], P=0.01), implying that LVEF was lower in patients with hypothyroidism than in healthy people. Analysis of heterogeneity found moderate heterogeneity (P = 0.08, I² = 50%). WMD values for stroke volume (SV), cardiac index (CI), left ventricular end-diastolic volume index(LVEDVI), left ventricular end-systolic volume (LESVI), and left ventricular mass index(LVMI) were also analyzed, and pooled effect sizes showed the CI and LVEDVI of patients with hypothyroidism ware significantly decrease (WMD=-0.47, 95% CI [-0.93 to -0.00], P=0.05, WMD=-7.99, 95%CI [-14.01 to -1.96], P=0.009, respectively). Patients with hypothyroidism tended to recover cardiac function after treatment [LVEF (WMD = 6.37, 95%CI [2.05, 10.69], P=0.004), SV (WMD = 7.67, 95%CI [1.61, 13.74], P=0.01), CI (WMD = 0.40, 95%CI [0.01, 0.79], P=0.05)], and there was no difference from the healthy controls.

Conclusion

Hypothyroidism could affect cardiac function, although this does not cause significant heart failure. It may be an adaptation of the heart to the hypothyroid state. There was a risk that this adaptation may turn into myocardial damage. Cardiac function could be restored after treatment in patients with hypothyroidism. Aggressive levothyroxine replacement therapy should be used to reverse cardiac function.

Systematic review registration

https://inplasy.com, identifier (INPLASY202440114).

Accuracy of the 3-dimensional virtual patient representation obtained by using 4 different techniques: An in vitro study

STATEMENT OF PROBLEM

Facial and intraoral scans can be aligned with or without the assistance of extraoral scan body systems to obtain a 3-dimensional (3D) virtual patient representation. However, the accuracy of the virtual patient remains uncertain.

PURPOSE

The purpose of this in vitro study was to measure the accuracy of the virtual patient representation obtained by superimposing facial and intraoral digital scans with 4 different techniques (with and without the usage of extraoral scan bodies) and to measure the operator influence on the accuracy of the virtual patient integration.

MATERIAL AND METHODS

Three markers were placed in the jaw simulation of a mannequin on the right (r), center (c), and left (l) surfaces. Five additional markers were attached to the mesiobuccal cusp of the right first molar (RM), cusp of the right canine (RC), buccal surface of the right central incisor (CI), cusp of the left canine (LC), and mesiobuccal cusp of the left first molar (LM). A reference scan (control scan) of the mannequin was obtained by using an industrial scanner (Gom ATOS Q 3D 12 M). Four different groups were created depending on the technique used: 3D scan body (3D scan body) (3D-SB group), AFT (AFT Dental System) (AFT group), Sat 3D (Sat 3D) (Sat3D group), and without using a scan body system (No-SB group). Additionally, a digital scan of the typodont was obtained with an intraoral scanner (TRIOS 4). The virtual patient integration was performed 10 times per group by 2 independent operators by using a software program (DentalCAD, Galway). Each operator obtained a total of 9 interlandmark measurements on the reference scan and on each virtual patient integration of each group with the measurement tool of the computer-aided design program. The data were analyzed by using 4-way ANOVA followed by the pairwise comparison Tukey tests (α=.05).

RESULTS

The group (P<.001), specimen (P<.001), and operator (P<.001) significantly influenced the trueness discrepancies obtained. Additionally, the 3D-SB group had the best trueness (244 μm), and the No-SB group had the worst trueness (346 μm). Operator 1 (279 μm) obtained significantly better trueness than operator 2 (295 μm). Group (P<.001), specimen (P<.001), and operator (P<.001) significantly influenced precision discrepancies, with the AFT (149 μm) and 3D-SB (154 μm) groups having the best precision and the No-SB group (269 μm) the worst precision. Operator 1 (176 μm) obtained significantly better precision than operator 2 (197 μm).

CONCLUSIONS

The techniques tested influenced the accuracy of the 3D virtual patient representation. The 3D-SB group had the best trueness, and the AFT and 3D-SB groups had the best precision, while the No-SB group showed the lowest trueness and precision values. Operator handling had a significant effect on the trueness and precision values of the virtual patient integrations tested.

Analysis of Pediatric Pulpotomy, Pulpectomy, and Extractions in Primary Teeth Revealed No Significant Association with Subsequent Root Canal Therapy and Extractions in Permanent Teeth: A Retrospective Study

Recent evidence suggests that an ever-growing number of pediatric patients require invasive treatments such as root canal therapy (RCT) in their permanent dentition, albeit with little information about risk factors such as prior invasive treatments of pulpotomy or pulpectomy in their primary dentition. Therefore, the primary objectives of this study were to determine the number of pediatric patients who have had any type of invasive treatment in their primary teeth, to assess their association with any subsequent invasive treatment (root canal therapy, extractions) in their permanent dentition, and to assess these trends over time. This retrospective study utilized summary data from a clinical pediatric patient pool (ages 0-17) over the period of 2013-2022. This analysis revealed that pediatric patients requiring pulpotomies and pulpectomies in primary dentition declined between 2013 (n = 417, n = 156) and 2022 (n = 250, n = 12), while root canal therapy (RCT) in permanent dentition increased six-fold from n = 54 to n = 330. In addition, few (7.8%) patients with RCT had a previous history of pulpotomy or pulpectomy, which suggests that invasive treatments performed in primary dentition have no direct association with the subsequent need for invasive treatments in permanent dentition, although more research is needed to determine the explanations for these observations.

An overview of the different digital facebow methods for transferring the maxillary cast into the virtual articulator

OBJECTIVES

The purposes of this study were to classify the described digital facebow techniques for transferring the maxillary cast into the semi-adjustable virtual articulator based on the digital data acquisition technology used and to review the reported accuracy values of the different digital facebow methods described.

OVERVIEW

Digital data acquisition technologies, including digital photographs, facial scanners, cone beam computed tomography (CBCT) imaging, and jaw tracking systems, can be used to transfer the maxillary cast into the virtual articulator. The reported techniques are reviewed, as well as the reported accuracy values of the different digital facebow methods.

CONCLUSIONS

Digital photographs can be used to transfer the maxillary cast into the virtual articulator using the true horizontal reference plane, but limited studies have assessed the accuracy of this method. Facial scanning and CBCT techniques can be used to transfer the maxillary cast into the virtual articulator, in which the most frequently selected references planes are the Frankfort horizontal, axis orbital, and true horizontal planes. Studies analyzing the accuracy of the maxillary cast transfer by using facial scanning and CBCT techniques are restricted. Lastly, optical jaw trackers can be selected for transferring the maxillary cast into the virtual articulator by using the axis orbital or true horizontal planes, yet the accuracy of these systems is unknown.

CLINICAL IMPLICATIONS

Digital data acquisition technologies, including digital photographs, facial scanning methods, CBCTs, and optical jaw tracking systems, can be used to transfer the maxillary cast into the virtual articulator. Studies are needed to assess the accuracy of these digital data acquisition technologies for transferring the maxillary cast into the virtual articulator.

An overview of artificial intelligence based applications for assisting digital data acquisition and implant planning procedures

OBJECTIVES

To provide an overview of the current artificial intelligence (AI) based applications for assisting digital data acquisition and implant planning procedures.

OVERVIEW

A review of the main AI-based applications integrated into digital data acquisitions technologies (facial scanners (FS), intraoral scanners (IOSs), cone beam computed tomography (CBCT) devices, and jaw trackers) and computer-aided static implant planning programs are provided.

CONCLUSIONS

The main AI-based application integrated in some FS's programs involves the automatic alignment of facial and intraoral scans for virtual patient integration. The AI-based applications integrated into IOSs programs include scan cleaning, assist scanning, and automatic alignment between the implant scan body with its corresponding CAD object while scanning. The more frequently AI-based applications integrated into the programs of CBCT units involve positioning assistant, noise and artifacts reduction, structures identification and segmentation, airway analysis, and alignment of facial, intraoral, and CBCT scans. Some computer-aided static implant planning programs include patient's digital files, identification, labeling, and segmentation of anatomical structures, mandibular nerve tracing, automatic implant placement, and surgical implant guide design.

Artificial Intelligence through Wireless Sensors Applied in Restorative Dentistry: A Systematic Review

The integration of wireless sensors with artificial intelligence could represent a transformative approach in restorative dentistry, offering a sophisticated means to enhance diagnostic precision, treatment planning, and patient outcomes. This systematic review was conducted to pinpoint and assess the efficacy of wireless sensors in restorative dentistry. The search methodology followed the guidelines outlined by PRISMA and involved the utilization of prominent scientific databases. Following the final phase of evaluating eligibility, the systematic review included six papers. Five experiments were conducted in vitro, while one was a randomized clinical trial. The investigations focused on wireless sensors for cavity diagnosis, toothbrush forces, facial mask applications, and physiological parameter detection from dental implants. All wireless sensors demonstrated efficacy in achieving the objectives established by each study and showed the validity, accuracy, and reproducibility of this device. The investigations examined in this systematic review illustrate the potential of wireless sensors in restorative dentistry, especially in the areas of caries detection, dental implant systems, face masks, and power brushes. These technologies hold promise for enhancing patient outcomes and alleviating the workload of dental practitioners.

Cephalometric analysis performance discrepancy between orthodontists and an artificial intelligence model using lateral cephalometric radiographs

PURPOSE

The purpose of the present clinical study was to compare the Ricketts and Steiner cephalometric analysis obtained by two experienced orthodontists and artificial intelligence (AI)-based software program and measure the orthodontist variability.

MATERIALS AND METHODS

A total of 50 lateral cephalometric radiographs from 50 patients were obtained. Two groups were created depending on the operator performing the cephalometric analysis: orthodontists (Orthod group) and an AI software program (AI group). In the Orthod group, two independent experienced orthodontists performed the measurements by performing a manual identification of the cephalometric landmarks and a software program (NemoCeph; Nemotec) to calculate the measurements. In the AI group, an AI software program (CephX; ORCA Dental AI) was selected for both the automatic landmark identification and cephalometric measurements. The Ricketts and Steiner cephalometric analyses were assessed in both groups including a total of 24 measurements. The Shapiro-Wilk test showed that the data was normally distributed. The t-test was used to analyze the data (α = 0.05).

RESULTS

The t-test analysis showed significant measurement discrepancies between the Orthod and AI group in seven of the 24 cephalometric parameters tested, namely the corpus length (p = 0.003), mandibular arc (p < 0.001), lower face height (p = 0.005), overjet (p = 0.019), and overbite (p = 0.022) in the Ricketts cephalometric analysis and occlusal to SN (p = 0.002) and GoGn-SN (p < 0.001) in the Steiner cephalometric analysis. The intraclass correlation coefficient (ICC) between both orthodontists of the Orthod group for each cephalometric measurement was calculated.

CONCLUSIONS

Significant discrepancies were found in seven of the 24 cephalometric measurements tested between the orthodontists and the AI-based program assessed. The intra-operator reliability analysis showed reproducible measurements between both orthodontists, except for the corpus length measurement.

CLINICAL SIGNIFICANCE

The artificial intelligence software program tested has the potential to automatically obtain cephalometric analysis using lateral cephalometric radiographs; however, additional studies are needed to further evaluate the accuracy of this AI-based system.

Detection of caries around restorations on bitewings using deep learning

OBJECTIVE

Secondary caries lesions adjacent to restorations, a leading cause of restoration failure, require accurate diagnostic methods to ensure an optimal treatment outcome. Traditional diagnostic strategies rely on visual inspection complemented by radiographs. Recent advancements in artificial intelligence (AI), particularly deep learning, provide potential improvements in caries detection. This study aimed to develop a convolutional neural network (CNN)-based algorithm for detecting primary caries and secondary caries around restorations using bitewings.

METHODS

Clinical data from 7 general dental practices in the Netherlands, comprising 425 bitewings of 383 patients, were utilized. The study used the Mask-RCNN architecture, for instance, segmentation, supported by the Swin Transformer backbone. After data augmentation, model training was performed through a ten-fold cross-validation. The diagnostic accuracy of the algorithm was evaluated by calculating the area under the Free-Response Receiver Operating Characteristics curve, sensitivity, precision, and F1 scores.

RESULTS

The model achieved areas under FROC curves of 0.806 and 0.804, and F1-scores of 0.689 and 0.719 for primary and secondary caries detection, respectively.

CONCLUSION

An accurate CNN-based automated system was developed to detect primary and secondary caries lesions on bitewings, highlighting a significant advancement in automated caries diagnostics.

CLINICAL SIGNIFICANCE

An accurate algorithm that integrates the detection of both primary and secondary caries will permit the development of automated systems to aid clinicians in their daily clinical practice.

Accuracy of an artificial intelligence-based program for locating the maxillomandibular relationship of scans acquired by using intraoral scanners

STATEMENT OF PROBLEM

An artificial-intelligence (AI) based program can be used to articulate scans in maximum intercuspal position (MIP) or correct occlusal collisions of articulated scans at MIP; however, the accuracy of the AI program determining the MIP relationship is unknown.

PURPOSE

The purpose of the present clinical study was to assess the influence of intraoral scanner (IOS) (TRIOS 5 or i700) and program (IOS or AI-based program) on the accuracy of the MIP relationship.

MATERIAL AND METHODS

Casts of a participant mounted on an articulator were digitized (T710). A maxillary and a mandibular scan of the participant were recorded by using 2 IOSs: TRIOS 5 and i700. The scans were duplicated 15 times. Then, each duplicated pair of scans was articulated in MIP using a bilateral occlusal record. Articulated scans were duplicated and allocated into 2 groups based on the automatic occlusal collisions' correction completed by using the corresponding IOS program: IOS-corrected and IOS-noncorrected group. Three subgroups were created based on the AI-based program (Bite Finder) method: AI-articulated, AI-IOS-corrected, and AI-IOS-noncorrected (n=15). In the AI-articulated subgroup, the nonarticulated scans were imported and articulated. In the AI-IOS-corrected subgroup, the articulated scans obtained in the IOS-corrected group were imported, and the occlusal collisions were corrected. In the AI-IOS-corrected subgroup, the articulated scans obtained in the IOS-noncorrected subgroup were imported, and the occlusal collisions were corrected. A total of 36 interlandmark measurements were calculated on each articulated scan (Geomagic Wrap). The distances computed on the reference scan were used as a reference to calculate the discrepancies with each experimental scan. Nonparametric 2-way ANOVA and pairwise multiple comparison Dwass-Steel-Critchlow-Fligner tests were used to analyze trueness. The general linear model procedure was used to analyze precision (α=.05).

RESULTS

Significant maxillomandibular trueness (P=.003) and precision (P<.001) differences were found among the subgroups. The IOS-corrected and IOS-noncorrected (P<.001) and AI-articulated and IOS-noncorrected subgroups (P=.011) were significantly different from each other. The IOS-corrected and AI-articulated subgroups obtained significantly better maxillomandibular trueness and precision than the IOS-noncorrected subgroups.

CONCLUSIONS

The IOSs tested obtained similar MIP accuracy; however, the program used to articulate or correct occlusal collusions impacted the accuracy of the MIP relationship.

Deep learning in the overall process of implant prosthodontics: A state-of-the-art review

Artificial intelligence represented by deep learning has attracted attention in the field of dental implant restoration. It is widely used in surgical image analysis, implant plan design, prosthesis shape design, and prognosis judgment. This article mainly describes the research progress of deep learning in the whole process of dental implant prosthodontics. It analyzes the limitations of current research, and looks forward to the future development direction.

Discrepancies in the occlusal devices designed by an experienced dental laboratory technician and by 2 artificial intelligence-based automatic programs

STATEMENT OF PROBLEM

Artificial intelligence (AI) models have been developed for different applications, including the automatic design of occlusal devices; however, the design discrepancies of an experienced dental laboratory technician and these AI automatic programs remain unknown.

PURPOSE

The purpose of this in vitro study was to compare the overall, intaglio, and occlusal surface discrepancies of the occlusal device designs completed by an experienced dental laboratory technician and two AI automatic design programs.

MATERIAL AND METHODS

Virtually articulated maxillary and mandibular diagnostic casts were obtained in a standard tessellation language (STL) file format. Three groups were created depending on the operator or program used to design the occlusal devices: an experienced dental laboratory technician (control group) and two AI programs, namely Medit Splints from Medit (Medit group) and Automate from 3Shape A/S (3Shape group) (n=10). To minimize the discrepancies in the parameter designs among the groups tested, the same printing material and design parameters were selected. In the control group, the dental laboratory technician imported the articulated scans into a dental design program (DentalCAD) and designed a maxillary occlusal device. The occlusal device designs were exported in STL format. In the Medit and 3Shape groups, the diagnostic casts were imported into the respective AI programs. The AI programs automatically designed the occlusal device without any further operator intervention. The occlusal device designs were exported in STL format. Among the 10 occlusal designs of the control group, a random design (shuffle deck of cards) was used as a reference file to calculate the overall, intaglio, and occlusal discrepancies in the specimens of the AI groups by using a program (Medit Design). The root mean square (RMS) error was calculated. Kruskal-Wallis, and post hoc Dwass-Steel-Critchlow-Fligner pairwise comparison tests were used to analyze the trueness of the data. The Levene test was used to assess the precision data (α=.05).

RESULTS

Significant overall (P<.001), intaglio (P<.001), and occlusal RMS median value (P<.001) discrepancies were found among the groups. Significant overall RMS median discrepancies were observed between the control and the Medit groups (P<.001) and the control and 3Shape groups (P<.001). Additionally, significant intaglio RMS median discrepancies were found between the control and the Medit groups (P<.001), the Medit and 3Shape groups (P<.001), and the control and 3Shape groups (P=.008). Lastly, significant occlusal RMS median discrepancies were found between the control and the 3Shape groups (P<.001) and the Medit and 3Shape groups (P<.001). The AI-based software programs tested were able to automatically design occlusal devices with less than a 100-µm trueness discrepancy compared with the dental laboratory technician. The Levene test revealed significant overall (P<.001), intaglio (P<.001), and occlusal (P<.001) precision among the groups tested.

CONCLUSIONS

The use of a dental laboratory technique influenced the overall, intaglio, and occlusal trueness of the occlusal device designs obtained. No differences were observed in the precision of occlusal device designs acquired among the groups tested.

Annual review of selected scientific literature: A report of the Committee on Scientific Investigation of the American Academy of Restorative Dentistry

The Scientific Investigation Committee of the American Academy of Restorative Dentistry offers this review of the 2022 dental literature to briefly touch on several topics of interest to modern restorative dentistry. Each committee member brings discipline-specific expertise in their subject areas that include (in order of the appearance in this report): prosthodontics; periodontics, alveolar bone, and peri-implant tissues; dental materials and therapeutics; occlusion and temporomandibular disorders; sleep-related breathing disorders; oral medicine and oral and maxillofacial surgery; and dental caries and cariology. The authors focused their efforts on reporting information likely to influence the daily dental treatment decisions of the reader with an emphasis on innovations, new materials and processes, and future trends in dentistry. With the tremendous volume of literature published daily in dentistry and related disciplines, this review cannot be comprehensive. Instead, its purpose is to update interested readers and provide valuable resource material for those willing to subsequently pursue greater detail on their own. Our intent remains to assist colleagues in navigating the tremendous volume of newly minted information produced annually. Finally, we hope that readers find this work helpful in managing patients.

Applications, functions, and accuracy of artificial intelligence in restorative dentistry: A literature review

OBJECTIVE

The applications of artificial intelligence (AI) are increasing in restorative dentistry; however, the AI performance is unclear for dental professionals. The purpose of this narrative review was to evaluate the applications, functions, and accuracy of AI in diverse aspects of restorative dentistry including caries detection, tooth preparation margin detection, tooth restoration design, metal structure casting, dental restoration/implant detection, removable partial denture design, and tooth shade determination.

OVERVIEW

An electronic search was performed on Medline/PubMed, Embase, Web of Science, Cochrane, Scopus, and Google Scholar databases. English-language articles, published from January 1, 2000, to March 1, 2022, relevant to the aforementioned aspects were selected using the key terms of artificial intelligence, machine learning, deep learning, artificial neural networks, convolutional neural networks, clustering, soft computing, automated planning, computational learning, computer vision, and automated reasoning as inclusion criteria. A manual search was also performed. Therefore, 157 articles were included, reviewed, and discussed.

CONCLUSIONS

Based on the current literature, the AI models have shown promising performance in the mentioned aspects when being compared with traditional approaches in terms of accuracy; however, as these models are still in development, more studies are required to validate their accuracy and apply them to routine clinical practice.

CLINICAL SIGNIFICANCE

AI with its specific functions has shown successful applications with acceptable accuracy in diverse aspects of restorative dentistry. The understanding of these functions may lead to novel applications with optimal accuracy for AI in restorative dentistry.

Artificial intelligence in the detection and classification of dental caries

STATEMENT OF PROBLEM

Automated detection of dental caries could enhance early detection, save clinician time, and enrich treatment decisions. However, a reliable system is lacking.

PURPOSE

The purpose of this study was to train a deep learning model and to assess its ability to detect and classify dental caries.

MATERIAL AND METHODS

Bitewings radiographs with a 1876×1402-pixel resolution were collected, segmented, and anonymized with a radiographic image analysis software program and were identified and classified according to the modified King Abdulaziz University (KAU) classification for dental caries. The method was based on supervised learning algorithms trained on semantic segmentation tasks.

RESULTS

The mean score for the intersection-over-union of the model was 0.55 for proximal carious lesions on a 5-category segmentation assignment and a mean F1 score of 0.535 using 554 training samples.

CONCLUSIONS

The study validated the high potential for developing an accurate caries detection model that will expedite caries identification, assess clinician decision-making, and improve the quality of patient care.

ChatGPT for shaping the future of dentistry: the potential of multi-modal large language model

The ChatGPT, a lite and conversational variant of Generative Pretrained Transformer 4 (GPT-4) developed by OpenAI, is one of the milestone Large Language Models (LLMs) with billions of parameters. LLMs have stirred up much interest among researchers and practitioners in their impressive skills in natural language processing tasks, which profoundly impact various fields. This paper mainly discusses the future applications of LLMs in dentistry. We introduce two primary LLM deployment methods in dentistry, including automated dental diagnosis and cross-modal dental diagnosis, and examine their potential applications. Especially, equipped with a cross-modal encoder, a single LLM can manage multi-source data and conduct advanced natural language reasoning to perform complex clinical operations. We also present cases to demonstrate the potential of a fully automatic Multi-Modal LLM AI system for dentistry clinical application. While LLMs offer significant potential benefits, the challenges, such as data privacy, data quality, and model bias, need further study. Overall, LLMs have the potential to revolutionize dental diagnosis and treatment, which indicates a promising avenue for clinical application and research in dentistry.

ORIENTATE: automated machine learning classifiers for oral health prediction and research

BACKGROUND

The application of data-driven methods is expected to play an increasingly important role in healthcare. However, a lack of personnel with the necessary skills to develop these models and interpret its output is preventing a wider adoption of these methods. To address this gap, we introduce and describe ORIENTATE, a software for automated application of machine learning classification algorithms by clinical practitioners lacking specific technical skills. ORIENTATE allows the selection of features and the target variable, then automatically generates a number of classification models and cross-validates them, finding the best model and evaluating it. It also implements a custom feature selection algorithm for systematic searches of the best combination of predictors for a given target variable. Finally, it outputs a comprehensive report with graphs that facilitates the explanation of the classification model results, using global interpretation methods, and an interface for the prediction of new input samples. Feature relevance and interaction plots provided by ORIENTATE allow to use it for statistical inference, which can replace and/or complement classical statistical studies.

RESULTS

Its application to a dataset with healthy and special health care needs (SHCN) children, treated under deep sedation, was discussed as case study. On the example dataset, despite its small size, the feature selection algorithm found a set of features able to predict the need for a second sedation with a f1 score of 0.83 and a ROC (AUC) of 0.92. Eight predictive factors for both populations were found and ordered by the relevance assigned to them by the model. A discussion of how to derive inferences from the relevance and interaction plots and a comparison with a classical study is also provided.

CONCLUSIONS

ORIENTATE automatically finds suitable features and generates accurate classifiers which can be used in preventive tasks. In addition, researchers without specific skills on data methods can use it for the application of machine learning classification and as a complement to classical studies for inferential analysis of features. In the case study, a high prediction accuracy for a second sedation in SHCN children was achieved. The analysis of the relevance of the features showed that the number of teeth with pulpar treatments at the first sedation is a predictive factor for a second sedation.

Artificial intelligence in the diagnosis of dental diseases on panoramic radiographs: a preliminary study

BACKGROUND

Artificial intelligence (AI) has been introduced to interpret the panoramic radiographs (PRs). The aim of this study was to develop an AI framework to diagnose multiple dental diseases on PRs, and to initially evaluate its performance.

METHODS

The AI framework was developed based on 2 deep convolutional neural networks (CNNs), BDU-Net and nnU-Net. 1996 PRs were used for training. Diagnostic evaluation was performed on a separate evaluation dataset including 282 PRs. Sensitivity, specificity, Youden's index, the area under the curve (AUC), and diagnostic time were calculated. Dentists with 3 different levels of seniority (H: high, M: medium, L: low) diagnosed the same evaluation dataset independently. Mann-Whitney U test and Delong test were conducted for statistical analysis (ɑ=0.05).

RESULTS

Sensitivity, specificity, and Youden's index of the framework for diagnosing 5 diseases were 0.964, 0.996, 0.960 (impacted teeth), 0.953, 0.998, 0.951 (full crowns), 0.871, 0.999, 0.870 (residual roots), 0.885, 0.994, 0.879 (missing teeth), and 0.554, 0.990, 0.544 (caries), respectively. AUC of the framework for the diseases were 0.980 (95%CI: 0.976-0.983, impacted teeth), 0.975 (95%CI: 0.972-0.978, full crowns), and 0.935 (95%CI: 0.929-0.940, residual roots), 0.939 (95%CI: 0.934-0.944, missing teeth), and 0.772 (95%CI: 0.764-0.781, caries), respectively. AUC of the AI framework was comparable to that of all dentists in diagnosing residual roots (p > 0.05), and its AUC values were similar to (p > 0.05) or better than (p < 0.05) that of M-level dentists for diagnosing 5 diseases. But AUC of the framework was statistically lower than some of H-level dentists for diagnosing impacted teeth, missing teeth, and caries (p < 0.05). The mean diagnostic time of the framework was significantly shorter than that of all dentists (p < 0.001).

CONCLUSIONS

The AI framework based on BDU-Net and nnU-Net demonstrated high specificity on diagnosing impacted teeth, full crowns, missing teeth, residual roots, and caries with high efficiency. The clinical feasibility of AI framework was preliminary verified since its performance was similar to or even better than the dentists with 3-10 years of experience. However, the AI framework for caries diagnosis should be improved.

Deep learning: A primer for dentists and dental researchers

OBJECTIVES

Despite deep learning's wide adoption in dental artificial intelligence (AI) research, researchers from other dental fields and, more so, dental professionals may find it challenging to understand and interpret deep learning studies, their employed methods, and outcomes. The objective of this primer is to explain the basic concept of deep learning. It will lay out the commonly used terms, and describe different deep learning approaches, their methods, and outcomes.

METHODS

Our research is based on the latest review studies, medical primers, as well as the state-of-the-art research on AI and deep learning, which have been gathered in the current study.

RESULTS

In this study, a basic understanding of deep learning models and various approaches to deep learning is presented. An overview of data management strategies for deep learning projects is presented, including data collection, data curation, data annotation, and data preprocessing. Additionally, we provided a step-by-step guide for completing a real-world project.

CONCLUSION

Researchers and clinicians can benefit from this study by gaining insight into deep learning. It can be used to critically appraise existing work or plan new deep learning projects.

CLINICAL SIGNIFICANCE

This study may be useful to dental researchers and professionals who are assessing and appraising deep learning studies within the field of dentistry.

Design of patient specific basal dental implant using Finite Element method and Artificial Neural Network technique

The bone conditions of mandibular bone vary from patient to patient, and as a result, a patient-specific dental implant needs to be designed. The basal dental implant is implanted in the cortical region of the bone since the top surface of the bone narrows down because of aging. Taguchi designs of experiments technique are used in which 25 optimum solid models of basal dental implants are modeled with variable geometrical parameters, viz. thread length, diameter, and pitch. In the solid models the implants are placed in the cortical part of the 3D models of cadaveric mandibles, that are prepared from CT data using image processing software. Patient-specific bone conditions are varied according to the strong, weak, and normal basal bone. A compressive force of 200 N is applied on the top surface of these implants and using finite element analysis software, the microstrain on the peri-implant bone ranges from 1000 to 4000 depending on the various bone conditions. According to the finite element data, it can be concluded that weak bone microstrain is comparatively high compared with normal and strong bone conditions. A surrogate artificial neural network model is prepared from the finite element analysis data. Surrogate model assisted genetic algorithm is used to find the optimum patient-specific basal dental implant for a better osseointegration-friendly mechanical environment.

Application and Performance of Artificial Intelligence Technology in Detection, Diagnosis and Prediction of Dental Caries (DC)—A Systematic Review

Evolution in the fields of science and technology has led to the development of newer applications based on Artificial Intelligence (AI) technology that have been widely used in medical sciences. AI-technology has been employed in a wide range of applications related to the diagnosis of oral diseases that have demonstrated phenomenal precision and accuracy in their performance. The aim of this systematic review is to report on the diagnostic accuracy and performance of AI-based models designed for detection, diagnosis, and prediction of dental caries (DC). Eminent electronic databases (PubMed, Google scholar, Scopus, Web of science, Embase, Cochrane, Saudi Digital Library) were searched for relevant articles that were published from January 2000 until February 2022. A total of 34 articles that met the selection criteria were critically analyzed based on QUADAS-2 guidelines. The certainty of the evidence of the included studies was assessed using the GRADE approach. AI has been widely applied for prediction of DC, for detection and diagnosis of DC and for classification of DC. These models have demonstrated excellent performance and can be used in clinical practice for enhancing the diagnostic performance, treatment quality and patient outcome and can also be applied to identify patients with a higher risk of developing DC.

Lingual bone thickness in the apical region of the horizontal mandibular third molar: A cross-sectional study in young Japanese

Background

Perforation of the lingual plate in the apical region of mandibular third molars will increase the risk of aberration and migration of the root tip and the risk of lingual nerve injury. The aim of this study was to analyze anatomical information, including relationships between the apical region of horizontally impacted mandibular third molars and lingual plates, in young Japanese.

Methods

Japanese patients, with horizontally impacted third molars, who underwent CT examination as a preoperative assessment for mandibular third molar extraction were included, and anatomical characteristics in the apical region of the right mandibular third molar were analyzed, in this study.

Results

A total of 121 patients were included based on the inclusion and exclusion criteria of this study. The mean and standard deviation of the bone thickness on the lingual side of the mandibular third molar in the apical region was 1.5 ± 1.6 mm, and the absence of lingual cortical bone in the apical region, namely, “perforation”, was observed in 44 patients. The statistical analysis revealed the predictors of cases with perforation as follows: gender, age, and the available space evaluated by Pell and Gregory classification.

Conclusions

This study clarified that “perforation” was sometimes observed in young Japanese, and that the predictors of those cases were as follows: gender, age, and the available space evaluated by Pell and Gregory classification.

Augmented, Virtual and Mixed Reality in Dentistry: A Narrative Review on the Existing Platforms and Future Challenges

The recent advancements in digital technologies have led to exponential progress in dentistry. This narrative review aims to summarize the applications of Augmented Reality, Virtual Reality and Mixed Reality in dentistry and describes future challenges in digitalization, such as Artificial Intelligence and Robotics. Augmented Reality, Virtual Reality and Mixed Reality represent effective tools in the educational technology, as they can enhance students’ learning and clinical training. Augmented Reality and Virtual Reality and can also be useful aids during clinical practice. Augmented Reality can be used to add digital data to real life clinical data. Clinicians can apply Virtual Reality for a digital wax-up that provides a pre-visualization of the final post treatment result. In addition, both these technologies may also be employed to eradicate dental phobia in patients and further enhance patient’s education. Similarly, they can be used to enhance communication between the dentist, patient, and technician. Artificial Intelligence and Robotics can also improve clinical practice. Artificial Intelligence is currently developed to improve dental diagnosis and provide more precise prognoses of dental diseases, whereas Robotics may be used to assist in daily practice.