Liu Che-Ming, Lu Tsung-Yueh, Wang Ching-Shuen, Feng Sheng-Wei, Lin Yu-Chieh, Lee Sheng-Yang, Lin Wei-Chun
Department of Dentistry, Wan-Fang Hospital, Taipei Medical University, Taipei, Taiwan.
School of Dentistry, College of Oral Medicine, Taipei Medical University, Taipei, Taiwan.
J Dent Sci. 2025 Jul;20(3):1665-1672. doi: 10.1016/j.jds.2025.03.016. Epub 2025 Mar 27.
BACKGROUND/PURPOSE: The uniqueness of human teeth necessitated that dental restorations be customized primarily through extensive manual labor. Therefore, this study explored the potential of AI designed dental restorations for clinical applications.
Digital artificial design and AI design crown restorations were replicated 10 times, for a total of 20 samples. The zirconia crown restoration was completed by strengthening and glazing according to standard clinical procedures. Samples were digitally archived using a dental scanner to assess reproducibility, precision, and occlusion. The human trial portion included natural tooth preparations by clinical standards. Three participants each designed two crowns, resulting in a total of six crowns. Dental x-rays were used for image evaluation.
The 3D accuracy showed that stereolithography (STL) and scan files of the AI design group were 3.4 and 6.6 times lower than the digital group, respectively ( < 0.05). The space of the occlusal surface of the AI-designed crown was 1.8-times higher than that of the digital design ( < 0.05). Intraoral optical images demonstrated that the AI designed crown closely resembled the human-designed counterpart in appearance. Comparison of color distribution showed more differences on the buccal and lingual sides between the two design patterns.
Clinical images indicate that the shape, precision, and space of AI designed crowns are comparable to those of digitally designed crowns. Despite the spatial differences in contact between AI designed and digitally designed crowns, the in vivo and in vitro test results demonstrated favorable realism and contact quality.
背景/目的：人类牙齿的独特性使得牙齿修复主要需要通过大量体力劳动进行定制。因此，本研究探讨了人工智能设计的牙齿修复在临床应用中的潜力。
数字人工设计和人工智能设计的牙冠修复体各复制10次，共20个样本。按照标准临床程序对氧化锆牙冠修复体进行强化和上釉处理。使用牙科扫描仪对样本进行数字存档，以评估可重复性、精度和咬合情况。人体试验部分包括按照临床标准制备天然牙齿。三名参与者每人设计两个牙冠，共设计出六个牙冠。使用牙科X光片进行图像评估。
三维精度显示，人工智能设计组的立体光刻（STL）文件和扫描文件分别比数字设计组低3.4倍和6.6倍（P<0.05）。人工智能设计的牙冠咬合面间隙比数字设计的高1.8倍（P<0.05）。口腔内光学图像显示，人工智能设计的牙冠在外观上与人工设计的牙冠非常相似。颜色分布比较显示，两种设计模式在颊侧和舌侧的差异更大。
临床图像表明，人工智能设计的牙冠在形状、精度和间隙方面与数字设计的牙冠相当。尽管人工智能设计的牙冠与数字设计的牙冠在接触空间上存在差异，但体内和体外测试结果显示出良好的逼真度和接触质量。
