Diagnostic accuracy of proximal enamel subsurface demineralization and its relationship with calcium loss and lesion depth

Artificial intelligence for caries detection: a novel diagnostic tool using deep learning algorithms

OBJECTIVES

The aim of this study was to develop an assessment tool for automatic detection of dental caries in periapical radiographs using convolutional neural network (CNN) architecture.

METHODS

A novel diagnostic model named ResNet + SAM was established using numerous periapical radiographs (4278 images) annotated by medical experts to automatically detect dental caries. The performance of the model was compared to the traditional CNNs (VGG19, ResNet-50), and the dentists. The Gradient-weighted Class Activation Mapping (Grad-CAM) technique shows the region of interest in the image for the CNNs.

RESULTS

ResNet + SAM demonstrated significantly improved performance compared to the modified ResNet-50 model, with an average F1 score of 0.886 (95% CI 0.855-0.918), accuracy of 0.885 (95% CI 0.862-0.901) and AUC of 0.954 (95% CI 0.924-0.980). The comparison between the performance of the model and the dentists revealed that the model achieved higher accuracy than that of the junior dentists. With the assist of the tool, the dentists achieved superior metrics with a mean F1 score of 0.827 and the interobserver agreement for dental caries is enhanced from 0.592/0.610 to 0.706/0.723.

CONCLUSIONS

According to the results obtained from the experiments, the automatic assessment tool using the ResNet + SAM model shows remarkable performance and has excellent possibilities in identifying dental caries. The use of the assessment tool in clinical practice can be of great benefit as a clinical decision-making support in dentistry and reduce the workload of dentists.

The power of light - From dental materials processing to diagnostics and therapeutics

Harnessing the power of light and its photonic energy is a powerful tool in biomedical applications. Its use ranges from biomaterials processing and fabrication of polymers to diagnostics and therapeutics. Dental light curable materials have evolved over several decades and now offer very fast (≤ 10 s) and reliable polymerization through depth (4-6 mm thick). This has been achieved by developments on two fronts: (1) chemistries with more efficient light absorption characteristics (camphorquinone [CQ], ~30 L mol-1 cm1 [ʎmax 470 nm]; monoacylphosphine oxides [MAPO], ~800 L mol-1 cm-1 [ʎmax 385 nm]; bisacylphosphine oxide [BAPO], ~1,000 L mol-1 cm-1 [ʎmax 385 nm]) as well mechanistically efficient and prolonged radical generation processes during and after light irradiation, and; (2) introducing light curing technologies (light emitting diodes [LEDs] and less common lasers) with higher powers (≤ 2 W), better spectral range using multiple diodes (short: 390-405 nm; intermediate: 410-450 nm; and long: 450-480 nm), and better spatial power distribution (i.e. homogenous irradiance). However, adequate cure of materials falls short for several reasons, including improper selection of materials and lights, limitations in the chemistry of the materials, and limitations in delivering light through depth. Photonic energy has further applications in dentistry which include transillumination for diagnostics, and therapeutic applications that include photodynamic therapy, photobiomodulation, and photodisinfection. Light interactions with materials and biological tissues are complex and it is important to understand the advantages and limitations of these interactions for successful treatment outcomes. This article highlights the advent of photonic technologies in dentistry, its applications, the advantages and limitations, and possible future developments.

Detecting Proximal Caries on Periapical Radiographs Using Convolutional Neural Networks with Different Training Strategies on Small Datasets

The present study aimed to evaluate the performance of convolutional neural networks (CNNs) that were trained with small datasets using different strategies in the detection of proximal caries at different levels of severity on periapical radiographs. Small datasets containing 800 periapical radiographs were randomly categorized into a training and validation dataset (n = 600) and a test dataset (n = 200). A pretrained Cifar-10Net CNN was used in the present study. Different training strategies were used to train the CNN model independently; these strategies were defined as image recognition (IR), edge extraction (EE), and image segmentation (IS). Different metrics, such as sensitivity and area under the receiver operating characteristic curve (AUC), for the trained CNN and human observers were analysed to evaluate the performance in detecting proximal caries. IR, EE, and IS recognition modes and human eyes achieved AUCs of 0.805, 0.860, 0.549, and 0.767, respectively, with the EE recognition mode having the highest values (p all < 0.05). The EE recognition mode was significantly more sensitive in detecting both enamel and dentin caries than human eyes (p all < 0.05). The CNN trained with the EE strategy, the best performer in the present study, showed potential utility in detecting proximal caries on periapical radiographs when using small datasets.

Artificial intelligence in digital cariology: a new tool for the diagnosis of deep caries and pulpitis using convolutional neural networks

Background

An accurate diagnosis of deep caries and pulpitis on periapical radiographs is a clinical challenge.

Methods

A total of 844 radiographs were included in this study. Of the 844, 717 (85%) were used for training and 127 (15%) were used for testing the three convolutional neural networks (CNNs) (VGG19, Inception V3, and ResNet18). The performance [accuracy, precision, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC)] of the CNNs were evaluated and compared. The CNN model with the best performance was further integrated with clinical parameters to see whether multi-modal CNN could provide an enhanced performance. The Gradient-weighted Class Activation Mapping (Grad-CAM) technique illustrates what image feature was the most important for the CNNs.

Results

The CNN of ResNet18 demonstrated the best performance [accuracy =0.82, 95% confidence interval (CI): 0.80–0.84; precision =0.81, 95% CI: 0.73–0.89; sensitivity =0.85, 95% CI: 0.79–0.91; specificity =0.82, 95% CI: 0.76–0.88; and AUC =0.89, 95% CI: 0.86–0.92], compared with VGG19 and Inception V3 as well as the comparator dentists. Therefore, ResNet18 was chosen to integrate with clinical parameters to produce the multi-modal CNN of ResNet18 + C, which showed a significantly enhanced performance (accuracy =0.86, 95% CI: 0.84–0.88; precision =0.85, 95% CI: 0.76–0.94; sensitivity =0.89, 95% CI: 0.83–0.95; specificity =0.86, 95% CI: 0.79–0.93; and AUC =0.94, 95% CI: 0.91–0.97).

Conclusions

The CNN of ResNet18 showed good performance (accuracy, precision, sensitivity, specificity, and AUC) for the diagnosis of deep caries and pulpitis. The multi-modal CNN of ResNet18 + C (ResNet18 integrated with clinical parameters) demonstrated a significantly enhanced performance, with promising potential for the diagnosis of deep caries and pulpitis.

Intraoral versus extraoral bitewing radiography in detection of enamel proximal caries: an ex vivo study

OBJECTIVE

To compare the diagnostic accuracy of digital intraoral and extraoral bitewing (EO BW) radiography in the detection of enamel proximal caries regardless of their ability to separate contacts.

METHODS

Artificial caries with different degrees of decalcification was induced in 80 human sound premolars and molars using formic acid. Intraoral radiographs were taken with photostimulable phosphor plate (PSP) and complementary metal oxide semiconductor (CMOS), using the paralleling bitewing technique. Extraoral bitewing radiographs were obtained using Sirona digital panoramic X-ray unit (Sirona Siemens, Bensheim, Germany). In total, 160 proximal surfaces were assessed by 2 observers twice. Area under the receiver operating characteristic curve (Az) values for each image type, observer and reading were compared using z-tests, with a significance level of p ≤ 0.05. Sensitivity, specificity, positive-predictive value and negative-predictive value for each observer and reading were calculated.

RESULTS

Spearman's test showed a strong positive correlation between the duration of demineralization and histological grading of carious teeth surfaces. For the three radiographic techniques, intraobserver reliability was strong to excellent. Moreover, interobserver agreement was strong. The differences between all detection methods were not statistically significant (p > 0.05). Intraoral bitewing using CMOS sensor had the highest sensitivity while EO BW had the highest specificity in the detection of incipient proximal caries.

CONCLUSIONS

Within the limits of the ex vivo design, the difference in diagnostic accuracy between the three radiographic techniques was not significant. EO BW could be used in the detection of enamel proximal caries with results comparable with intraoral bitewing with PSP plate and CMOS sensor.

Effects of different orthodontic primers on enamel demineralization around orthodontic brackets

AIM

The purpose of this work is to evaluate the effectiveness of one self-etching and two filled orthodontic primers on enamel demineralization around orthodontic brackets.

METHODS

Brackets were bonded to 84 bovine teeth and the vestibular enamel surfaces covered with acid-resistant nail varnish exposing 1 mm of space on each side of the bracket base. The teeth were allocated to four groups, using either Transbond XT conventional primer on etched enamel (group 1), Transbond Plus Self-Etching Primer on untreated enamel (group 2), Pro Seal filled resin primer on etched enamel (group 3), or Opal Seal filled resin primer on etched enamel (group 4). Each tooth was subjected to 15,000 strokes of brushing followed by exposure to an acid challenge. Calcium-ion release from each sample was calculated using atomic absorption spectrophotometry. Data were analyzed using one-way ANOVA and a post hoc Tukey test. Differences were considered statistically significant at p ≤ 0.05.

RESULTS

Statistically significant differences were observed between the four groups (p < 0.001). No significant difference was found between the controls (group 1) and the Opal Seal group. Higher calcium release was observed in the Pro Seal group and the self-etching primer group compared to the controls. The highest calcium release was recorded in the self-etching primer group.

CONCLUSION

Filled sealants may not have a protective effect against enamel demineralization. Transbond Plus Self-Etching Primer should be used cautiously, considering the risk of demineralization involved in its application.

Detection of artificial demineralization bordering different types of laminate veneers using visual inspection and storage phosphor radiography

OBJECTIVE

The objective of this study is to compare the diagnostic accuracy of visual inspection (VI) and storage phosphor plate (SPP) radiography for the detection of artificial demineralization bordering different laminate veneers.

MATERIALS AND METHODS

Twenty human maxillary canine teeth were prepared. All-ceramic (A) and hybrid ceramic (H) laminate veneers were fabricated and luted. Veneered teeth were covered except for a circular window on the proximal surface bordering restorations. Teeth were kept in acetic acid buffer to create demineralization and imaged with a SPP system. Ten observers evaluated all teeth first visually then with SPP images for the presence/absence of demineralization. Teeth were examined using scanning electron microscopy (SEM) as well. The accuracy was expressed as the area under the ROC curves (A(z)). Pair-wise comparisons were performed using two-way ANOVA and post hoc t test (p = 0.05). Fleiss kappa (κ) was used for agreement.

RESULTS

SPP radiography was better than the VI for both veneers (p = 0.004). The A(z)s of two veneers were different for both VI (p < 0.005) and SPP (p < 0.005). SEM evaluation revealed lesions confined to enamel. κ was fair for H, and fair to moderate for A. Agreement was higher for the radiographic evaluation for both veneers.

CONCLUSION

Enamel demineralizations bordering hybrid and ceramic laminate veneers can be detected better with SPP radiography than VI and detectability was better for all-ceramic veneers than the hybrid ceramic ones.

CLINICAL RELEVANCE

Early detection of enamel demineralizations bordering laminate veneers would result in time-saving and less-invasive treatment methods; therefore, SPP radiography may be recommended in clinically suspicious cases since it provides better diagnostic accuracy.