Non-Cochrane · DTA · PM40181403
Accuracy of artificial intelligence in caries detection: a systematic review and meta-analysis.
Luke Alexander Maniangat, Rezallah Nader Nabil Fouad · Head & face medicine · 2025 · PubMed · doi:10.1186/s13005-025-00496-8
Full textEvery question was extracted from the full text, figures and supplements by two independent LLM extractors, disagreements adjudicated. Rows cite where in the review the value comes from.
Questions
6
current version
GRADE rated
0
No studies
0
Included studies
35
linked below
No studies 0 · 0%Ungraded 6 · 100%Very low 0 · 0%Low 0 · 0%Moderate 0 · 0%High 0 · 0%
Ungraded means an effect was reported without a GRADE rating; it is not a gap. No studies means the review found nothing to answer the question. The six categories are a display order, not a score.
Review question
| Element | Planned (methods) | Found (results) |
|---|---|---|
| Population | Patients undergoing dental X-ray imaging | Patients imaged for caries; 21 studies (2017-2023) across China, South Korea, Brazil, Turkey, France, USA, UK, Taiwan, Germany, India, Netherlands (n=21) |
| Intervention | AI systems for detection and segmentation of dental structures and lesions on X-ray images (CNNs, DNNs, SVMs, decision trees, random forests) | AI models, predominantly CNNs; also ANNs, DCNN, FCNN, EfficientNet, ResNet, AlexNet (n=21) |
| Comparison | Human dentists or traditional methods of dental X-ray image interpretation | Trained dental professionals/dental surgeons or reference AI models (n varies by study) |
| Outcomes | Accuracy, sensitivity, and specificity of AI in detecting dental caries | Diagnostic accuracy (14 studies, pooled OR ~2.72, I2=88.0%); Sensitivity (9 studies, 17,190 participants, pooled OR 1.258 [95% CI 0.493-2.540], I2=86.03%); Specificity (7 studies, weighted mean ~87.90%, I2=96.03%) |
| Study designs | Prospective and retrospective cohort studies and case-control studies | Mostly retrospective and cross-sectional diagnostic accuracy studies; 2 randomized controlled trials (Mertens 2021; Devlin 2021) (n=21) |
| Timing | Not specified | Included studies published 2017-2023 |
Search date: Up to January 2024 (data collected up to January 2024; electronic search completed 19 May 2023, final search included research up to January 2024). Databases: PubMed; Scopus; Web of Science; Embase; Cochrane Library; IEEE Xplore; Google Scholar
Questions and results
AI systems for caries detection/segmentation on dental X-ray images (predominantly convolutional neural networks; also ANNs, DNNs, DCNN, FCNN, SVMs, decision trees, random forests; e.g. EfficientNet, ResNet, AlexNet) vs Human dentists / trained dental professionals or traditional radiographic interpretation (reference standard / reference models)
Population: Patients undergoing dental X-ray (radiographic) imaging for caries detection
| Outcome | Effect | Evidence | Status / GRADE |
|---|---|---|---|
Diagnostic accuracy (prevalence of accuracy) Digital bitewing X-rays (Fig 4a) On digital bitewing X-rays, AI accuracy was around 80% (single study, narrow CI; ungraded subgroup, numeric estimate not precisely readable). Source: Results 'Accuracy of caries detection by different Xray technique'; Fig 4a · same question in other reviews | ~80 | 1 study | Ungraded |
Diagnostic accuracy (prevalence of accuracy) Periapical X-rays (Fig 4c) On periapical X-rays, AI accuracy was around 76% (wide CI; ungraded subgroup, numeric estimate not precisely readable). Source: Results 'Accuracy of caries detection by different Xray technique'; Fig 4c · same question in other reviews | ~76 | NR | Ungraded |
Specificity Overall (all studies pooled) Across 7 studies, AI showed high weighted-mean specificity (~87.9%) for caries detection, with very high heterogeneity (I2=96%); ungraded. Source: Results 'Overall sensitivity of AI accuracy in caries detection' (specificity paragraph); Fig 6; Table 6 · same question in other reviews | Proportion (weighted mean specificity, %) 87.90 | 7 studies | Ungraded |
Sensitivity Overall (all studies pooled) Pooling 9 studies (17,190 participants), AI sensitivity for caries detection showed no significant association (pooled OR 1.258, 95% CI 0.493-2.540) with high heterogeneity (I2=86%); ungraded. Source: Results 'Overall sensitivity of AI accuracy in caries detection'; Fig 5; Table 5 · same question in other reviews | OR 1.258 [0.493, 2.540] | 9 studies · 17190 participants | Ungraded |
Diagnostic accuracy (prevalence of accuracy) Digital panoramic X-rays (Fig 4b) On digital panoramic X-rays, AI accuracy was around 70% (single study, wide CI; ungraded subgroup, numeric estimate not precisely readable). Source: Results 'Accuracy of caries detection by different Xray technique'; Fig 4b · same question in other reviews | ~70 | 1 study | Ungraded |
Diagnostic accuracy (prevalence of accuracy) Overall (all X-ray types pooled) Across 14 studies, AI showed high diagnostic accuracy for caries detection on dental X-rays (pooled OR ~2.72), with substantial heterogeneity (I2=88%) and no GRADE rating (ungraded). Source: Results 'Overall prevalence of accuracy of the caries detection'; Fig 3; Table 4 · same question in other reviews | OR 2.72 [14.65, 14.65] | 14 studies | Ungraded |
Included primary studies 35
Open as a trial list| Study | Also included by | Outcomes it informs here | Trial PICO | Results |
|---|---|---|---|---|
| Bayraktar 2022 Diagnosis of interproximal caries lesions with deep convolutional neural network in digital bitewing radiographs. PMID:34173051 | 1 other review | not stated | not extracted yet | not extracted yet |
| De Araujo Faria 2021 Prediction of Radiation-Related Dental Caries Through PyRadiomics Features and Artificial Neural Network on Panoramic Radiography. PMID:34254199 | 1 other review | not stated | not extracted yet | not extracted yet |
| Devlin 2021 The ADEPT study: a comparative study of dentists' ability to detect enamel-only proximal caries in bitewing radiographs with and without… PMID:34686815 | 1 other review | not stated | not extracted yet | not extracted yet |
| Geetha 2020 Dental caries diagnosis in digital radiographs using back-propagation neural network. PMID:31949895 | 1 other review | not stated | not extracted yet | not extracted yet |
| Lee 2018 Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm. PMID:30056118 | 1 other review | not stated | not extracted yet | not extracted yet |
| Zheng 2021 Artificial intelligence in digital cariology: a new tool for the diagnosis of deep caries and pulpitis using convolutional neural networks. PMID:34268376 | 1 other review | not stated | not extracted yet | not extracted yet |
| Bayrakdar 2021 Reduction-induced surface modification of human hair. PMID:20211112 | this review only | Sensitivity | not extracted yet | not extracted yet |
| Bayrakdar 2021 Deep-learning approach for caries detection and segmentation on dental bitewing radiographs. PMID:34807344 | this review only | not stated | not extracted yet | not extracted yet |
| Bayraktar 2022 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Cantu 2020 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Cantu 2020 Detecting caries lesions of different radiographic extension on bitewings using deep learning. PMID:32634466 | this review only | not stated | not extracted yet | not extracted yet |
| Chen 2021 matched by author and year (probable) | this review only | not stated | not extracted yet | not extracted yet |
| Chen 2021 Dental disease detection on periapical radiographs based on deep convolutional neural networks. PMID:33651298 | this review only | not stated | not extracted yet | not extracted yet |
| Chen 2023 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Chen 2023 Deep Learning-Based Recognition of Periodontitis and Dental Caries in Dental X-ray Images. PMID:37627796 | this review only | not stated | not extracted yet | not extracted yet |
| Choi 2018 matched by author and year (probable) | this review only | not stated | not extracted yet | not extracted yet |
| De Araujo Faria 2021 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Devlin 2021 matched by author and year (probable) | this review only | Sensitivity | not extracted yet | not extracted yet |
| Geetha 2020 [Changes in skin enzyme activity in experimental burns]. PMID:202088 | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Huang 2021 Deep Learning for Caries Detection using Optical Coherence Tomography DOI:10.1101/2021.05.04.21256502 | this review only | not stated | not extracted yet | not extracted yet |
| Huang 2021 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Hur 2021 Machine learning to predict distal caries in mandibular second molars associated with impacted third molars. PMID:34326441 | this review only | not stated | not extracted yet | not extracted yet |
| Lee 2021 Deep learning for early dental caries detection in bitewing radiographs. PMID:34413414 | this review only | not stated | not extracted yet | not extracted yet |
| Lian 2021 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Lian 2021 Deep Learning for Caries Detection and Classification. PMID:34574013 | this review only | not stated | not extracted yet | not extracted yet |
| Mao 2021 Caries and Restoration Detection Using Bitewing Film Based on Transfer Learning with CNNs. PMID:34283167 | this review only | not stated | not extracted yet | not extracted yet |
| Mertens 2021 matched by author and year (probable) | this review only | Sensitivity | not extracted yet | not extracted yet |
| Mertens 2021 Artificial intelligence for caries detection: Randomized trial. PMID:34656656 | this review only | not stated | not extracted yet | not extracted yet |
| Moran 2021 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Moran 2021 Classification of Approximal Caries in Bitewing Radiographs Using Convolutional Neural Networks. PMID:34372429 | this review only | not stated | not extracted yet | not extracted yet |
| Srivastava 2017 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Vinayahalingam 2021 matched by author and year (probable) | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Vinayahalingam 2021 Classification of caries in third molars on panoramic radiographs using deep learning. PMID:34131266 | this review only | not stated | not extracted yet | not extracted yet |
| Zheng 2021 Tracking of serum 25-hydroxyvitamin D levels during 14 years in a population-based study and during 12 months in an intervention study. PMID:20219763 | this review only | Diagnostic accuracy (prevalence of accuracy) | not extracted yet | not extracted yet |
| Zhu 2022 CariesNet: a deep learning approach for segmentation of multi-stage caries lesion from oral panoramic X-ray image. PMID:35017793 | this review only | not stated | not extracted yet | not extracted yet |
Overlap with other reviews
Reviews that include the same primary studies are not independent evidence. Shared counts match studies by PMID, DOI or NCT, and otherwise by author and year (a probable match).