Hove L H, Mulic A, Tveit A B, Stenhagen K R, Skaare A B, Espelid I
Department of Cariology and Gerodontology, Faculty of Dentistry, Institute of Clinical Dentistry, University of Oslo, P.O. Box 1109, Blindern, 0317, Oslo, Norway,
Eur Arch Paediatr Dent. 2013 Feb;14(1):29-34. doi: 10.1007/s40368-012-0004-5. Epub 2013 Feb 7.
Clinical photographs and study models may provide permanent records of dental erosion and be useful supplements to clinical registration.
To assess the reliability and validity of registrations on clinical photographs and study models performed by a group of examiners.
Thirty tooth surfaces were selected and assessed clinically, using the visual erosion dental examination system. The chosen surfaces provided the whole range of dental erosions including sound surfaces. The tooth surfaces were photographed and impressions were obtained for preparation of study models. Thirty-three dentists examined and scored the selected surfaces both on photographs and study models.
The quality of diagnosis (AUC, area under curve) was slightly higher using photographs as compared to study models. The difference was statistically significant when the validation criterion was erosion, assuming dentine exposure. The inter-method agreement on photographs and study models versus the clinical evaluation were approximately in the same range with a mean κw of 0.48 and 0.43, respectively. When comparing study models with photographs the mean κw was 0.52. The intra-examiner agreement was strong/substantial for both (photographs mean κw = 0.63 and study models mean κw = 0.60).
Linear weighted Cohen's kappa (κw) was used to evaluate inter-method and intra-examiner agreement. Receiver operating characteristic and area under the curves were used to express diagnostic quality according to a clinical examination.
The results indicated that photographs were as good as study models for recording erosive lesions.
临床照片和研究模型可为牙齿侵蚀提供永久记录，并作为临床记录的有益补充。
评估一组检查人员对临床照片和研究模型进行记录的可靠性和有效性。
选择30个牙面，使用视觉牙齿侵蚀检查系统进行临床评估。所选牙面涵盖了包括健康牙面在内的整个牙齿侵蚀范围。对牙面进行拍照，并制取印模以制作研究模型。33名牙医对所选牙面在照片和研究模型上进行检查并评分。
与研究模型相比，使用照片时诊断质量（曲线下面积，AUC）略高。当验证标准为侵蚀（假设牙本质暴露）时，差异具有统计学意义。照片和研究模型与临床评估之间的方法间一致性大致处于相同范围，平均κw分别为0.48和0.43。比较研究模型和照片时，平均κw为0.52。两者的检查者内一致性均较强/高度一致（照片平均κw = 0.63，研究模型平均κw = 0.60）。
使用线性加权Cohen's kappa（κw）评估方法间和检查者内一致性。根据临床检查，使用受试者工作特征曲线和曲线下面积来表示诊断质量。
结果表明，照片在记录侵蚀性病变方面与研究模型效果相当。
