Evaluating YOLO for dental caries diagnosis: a systematic review and meta-analysis.

IF 2.3 Q3 Dentistry

Evidence-based dentistry Pub Date : 2025-07-23 DOI:10.1038/s41432-025-01180-1

Quang Tuan Lam, Minh Huu Nhat Le, I-Ta Lee, Nguyen Quoc Khanh Le

{"title":"Evaluating YOLO for dental caries diagnosis: a systematic review and meta-analysis.","authors":"Quang Tuan Lam, Minh Huu Nhat Le, I-Ta Lee, Nguyen Quoc Khanh Le","doi":"10.1038/s41432-025-01180-1","DOIUrl":null,"url":null,"abstract":"Objectives: Recent advancements in the You Only Look Once (YOLO) algorithm show promise for dental caries diagnosis. We aimed to evaluate the diagnostic performance of different YOLO versions using photographic and radiographic images for caries detection.Methods: We searched PubMed (MEDLINE), EMBASE, Web of Science, and Scopus for studies up to December 12, 2024. Studies using any YOLO version for caries detection were included. Binary diagnostic accuracy data were extracted to calculate pooled sensitivity, specificity, and area under the curve (AUC) using a bivariate random-effects model. Quality was assessed with QUADAS-2 and the Radiomics Quality Score (RQS). This review is registered in PROSPERO (CRD42024615440).Results: We included 15 studies in the systematic review and 14 in the meta-analysis. Overall, YOLO-based models achieved a pooled sensitivity of 79.3% and specificity of 84.9%, with an AUC of 0.832. YOLO using radiographic images demonstrated higher specificity (92.5% vs 72.0%) and AUC (0.847 vs 0.735) than using photographic images, while sensitivity was similar (78.6% vs 80.0%). Differences between YOLO versions (v5 and earlier vs v6 and later) and the use of external validation did not significantly affect diagnostic accuracy.Discussion: Radiograph-based YOLO models showed superior specificity to photograph-based models, reflecting the higher diagnostic detail of radiographs. However, photographic approaches are completely radiation-free and more accessible, which could benefit screening in low-resource settings. Newer YOLO versions did not significantly outperform older versions, likely due to the limited complexity of the task and dataset constraints in current studies.Conclusions: YOLO algorithms provide a reliable tool for dental caries detection. Radiograph imaging combined with YOLO offers enhanced diagnostic specificity, while even older YOLO versions remain effective for caries detection in practice.","PeriodicalId":12234,"journal":{"name":"Evidence-based dentistry","volume":" ","pages":""},"PeriodicalIF":2.3000,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Evidence-based dentistry","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1038/s41432-025-01180-1","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"Dentistry","Score":null,"Total":0}

引用次数: 0

Abstract

Objectives: Recent advancements in the You Only Look Once (YOLO) algorithm show promise for dental caries diagnosis. We aimed to evaluate the diagnostic performance of different YOLO versions using photographic and radiographic images for caries detection.

Methods: We searched PubMed (MEDLINE), EMBASE, Web of Science, and Scopus for studies up to December 12, 2024. Studies using any YOLO version for caries detection were included. Binary diagnostic accuracy data were extracted to calculate pooled sensitivity, specificity, and area under the curve (AUC) using a bivariate random-effects model. Quality was assessed with QUADAS-2 and the Radiomics Quality Score (RQS). This review is registered in PROSPERO (CRD42024615440).

Results: We included 15 studies in the systematic review and 14 in the meta-analysis. Overall, YOLO-based models achieved a pooled sensitivity of 79.3% and specificity of 84.9%, with an AUC of 0.832. YOLO using radiographic images demonstrated higher specificity (92.5% vs 72.0%) and AUC (0.847 vs 0.735) than using photographic images, while sensitivity was similar (78.6% vs 80.0%). Differences between YOLO versions (v5 and earlier vs v6 and later) and the use of external validation did not significantly affect diagnostic accuracy.

Discussion: Radiograph-based YOLO models showed superior specificity to photograph-based models, reflecting the higher diagnostic detail of radiographs. However, photographic approaches are completely radiation-free and more accessible, which could benefit screening in low-resource settings. Newer YOLO versions did not significantly outperform older versions, likely due to the limited complexity of the task and dataset constraints in current studies.

Conclusions: YOLO algorithms provide a reliable tool for dental caries detection. Radiograph imaging combined with YOLO offers enhanced diagnostic specificity, while even older YOLO versions remain effective for caries detection in practice.

查看原文本刊更多论文

评估YOLO对龋齿诊断的价值：一项系统回顾和荟萃分析。

目的：你只看一次（YOLO）算法的最新进展显示了龋齿诊断的希望。我们的目的是评估不同的YOLO版本的诊断性能，使用摄影和放射图像检测龋。方法：检索截至2024年12月12日的PubMed （MEDLINE）、EMBASE、Web of Science和Scopus。使用任何YOLO版本进行龋齿检测的研究都包括在内。提取二元诊断准确性数据，使用二元随机效应模型计算合并敏感性、特异性和曲线下面积（AUC）。采用QUADAS-2和放射组学质量评分（RQS）评估质量。本综述已在PROSPERO注册（CRD42024615440）。结果：我们在系统评价中纳入了15项研究，在meta分析中纳入了14项研究。总体而言，基于yolo的模型的总灵敏度为79.3%，特异性为84.9%，AUC为0.832。x线影像的YOLO特异性（92.5% vs 72.0%）和AUC （0.847 vs 0.735）高于摄影影像，而灵敏度相似（78.6% vs 80.0%）。YOLO版本之间的差异（v5和更早的版本与v6和更高的版本）和外部验证的使用并没有显著影响诊断的准确性。讨论：基于x线片的YOLO模型比基于照片的模型具有更好的特异性，反映了x线片更高的诊断细节。然而，照相方法是完全无辐射和更容易获得的，这可能有利于在资源匮乏的情况下进行筛查。较新的YOLO版本并没有明显优于旧版本，这可能是由于当前研究中任务和数据集约束的复杂性有限。结论：YOLO算法为龋病检测提供了可靠的工具。x线影像结合YOLO提供了增强的诊断特异性，而即使是较老的YOLO版本在龋齿检测中仍然有效。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Evidence-based dentistry Dentistry-Dentistry (all)

CiteScore

2.50

自引率

0.00%

发文量

期刊介绍： Evidence-Based Dentistry delivers the best available evidence on the latest developments in oral health. We evaluate the evidence and provide guidance concerning the value of the author''s conclusions. We keep dentistry up to date with new approaches, exploring a wide range of the latest developments through an accessible expert commentary. Original papers and relevant publications are condensed into digestible summaries, drawing attention to the current methods and findings. We are a central resource for the most cutting edge and relevant issues concerning the evidence-based approach in dentistry today. Evidence-Based Dentistry is published by Springer Nature on behalf of the British Dental Association.