Page 33 - 南京医科大学自然版
P. 33

第45卷第12期           赵   鹃,刘雯丽,祝非凡,等. 人工智能在口腔种植修复中的研究进展、应用与挑战[J].
                 2025年12月                    南京医科大学学报(自然科学版),2025,45(12):1709-1718                      ·1715 ·


                箱”问题,模型泛化能力不足,缺乏大规模临床验                                 intelligence in dental implant[J]. Chinese Journalof
                证,医师对 AI 认知水平差异,AI 软件及技术的低普                            Prosthodontics,2022,23(2):81-85
                及率,信息安全,伦理问题等。AI 也尚未具备人脑                         [6] 郑懿诺,孙沐毅,张虹云,等. 深度学习在口腔种植影像
                的综合分析和信息关联的能力,可以作为临床决策                                 学中的应用:研究进展与挑战[J]. 数据与计算发展前
                                                                       沿(中英文),2024,6(3):41-49
                的辅助,但仍需要人工主导综合判断。在人与人互
                                                                       ZHENG Y N,SUN M Y,ZHANG H Y,et al. Application
                动,特别是医患交流方面,计算机语言仍难以完全
                                                                       of deep learning in dental implant imaging:research
                替代人类语言。
                                                                       progress and challenges[J]. Frontiers of Data & Computing,
                    因此,为了适应口腔种植修复领域的发展,
                                                                       2024,6(3):41-49
                AI 应从数据端探索数据融合和增强策略,从算法端                         [7] YING S N,HUANG F,LIU W,et al. Deep learning in the
                探索半监督/弱监督/无监督学习框架,研发少标注/                               overall process of implant prosthodontics :a state⁃of⁃the⁃
                无标注大数据学习模型,利用数据脱敏、不可逆信                                 art review[J]. Clin Implant Dent Relat Res,2024,26(5):
                息转化等技术,实现数据安全及患者隐私保护,开                                 835-846
                展前瞻性多中心研究进行模型训练,提高准确度。                           [8] LIU C H,LIN C J,HU Y H,et al. Predicting the failure of
                未来,随着电子健康记录和多模态大数据的发展,                                 dental implants using supervised learning techniques[J].
                多模态数据拟合和多学科融合度将成为提升 AI 应                               Appl Sci,2018,8(5):698
                                                                 [9] LYAKHOV P A,DOLGALEV A A,LYAKHOVA U A,et
                用效果的重要方向。通过 AI 实现跨模态对齐多模
                                                                       al. Neural network system for analyzing statistical factors
                态数据,域适用及域泛化跨越学科间的壁垒,整合
                                                                       of patients for predicting the survival of dental implants
                多模态数据(影像学、口扫、面扫、口颌系统动静态
                                                                      [J]. Front Neuroinform,2022,16:1067040
                数据、基因组学、生物标志物等),生成更全面的患
                                                                 [10]PAPANTONOPOULOS G,GOGOS C,HOUSOS E,et al.
                者画像,构建全方位的虚拟患者,基于口腔全局观                                 Prediction of individual implant bone levels and the
                及多学科合作,优化诊断和临床决策,模拟和预测                                 existence of implant“phenotypes”[J]. Clin Oral Implants
                治疗效果,从而使治疗方案更科学合理,实现个性                                 Res,2017,28(7):823-832
                化医疗和精准治疗,提高诊疗效率,推动整个口腔                           [11]HUANG N N,LIU P,YAN Y L,et al. Predicting the risk
                医学领域向更高水平发展。                                           of dental implant loss using deep learning[J]. J Clin
                                                                       Periodontol,2022,49(9):872-883
               [参考文献]
                                                                 [12] MAHRI M,SHEN N,BERRIZBEITIA F,et al. Osseointe⁃
               [1] STOUMPOS A I,KITSIOS F,TALIAS M A. Digital trans⁃   gration pharmacology:a systematic mapping using artifi⁃
                    formation in healthcare:technology acceptance and its  cial intelligence[J]. Acta Biomater,2021,119:284-302
                    applications[J]. Int J Environ Res Public Health,2023,  [13] CHOI H R,SIADARI T S,KIM J E,et al. Automatic
                    20(4):3407                                         detection of teeth and dental treatment patterns on dental
               [2] ALHARBI M T,ALMUTIQ M M. Prediction of dental       panoramic radiographs using deep neural networks[J].
                     implants using machine learning algorithms[J]. J Healthc  Forensic Sci Res,2022,7(3):456-466
                     Eng,2022,2022:7307675                       [14]XU M M,WU Y J,XU Z N,et al. Robust automated teeth
               [3] VASHISHT R,SHARMA A,KIRAN T,et al. Artificial       identification from dental radiographs using deep learning
                    intelligence in dentistry ⁃ a scoping review[J]. J Oral  [J]. J Dent,2023,136:104607
                    Maxillofac Surg Med Pathol,2024,36(4):579-592  [15] LI W,WANG Y J,LIU Y. DMAF ⁃ Net:deformable
               [4] 朱可石,廖安琪,余优成. 机器学习在口腔种植学中的                           multi ⁃ scale adaptive fusion network for dental structure
                    应用研究进展[J]. 国际口腔医学杂志,2023,50(4):                    detection with panoramic radiographs[J]. Dentomaxillofac
                    491-498                                            Radiol,2024,53(5):296-307
                    ZHU K S,LIAO A Q,YU Y C. Research progress on  [16]JANG T J,KIM K C,CHO H C,et al. A fully automated
                    the application of machine learning in dental implantolo⁃  method for 3D individual tooth identification and segmen⁃
                    gy[J]. International Journal of Stomatology,2023,50(4):  tation in dental CBCT[J]. IEEE Trans Pattern Anal Mach
                    491-498                                            Intell,2022,44(10):6562-6568
               [5] 王俊成,时       权,刘洪臣. AI 在口腔种植中的应用[J].           [17] CUI Z M,FANG Y,MEI L,et al. A fully automatic AI
                    口腔颌面修复学杂志,2022,23(2):81-85                         system for tooth and alveolar bone segmentation from
                    WANG J C,SHI Q,LIU H C. Application of artificial  cone⁃beam CT images[J]. Nat Commun,2022,13(1):
   28   29   30   31   32   33   34   35   36   37   38