<![CDATA[ABSTRAKPendahuluan: Osseointegrasi merupakan faktor kunci dalam keberhasilan implan gigi, namun evaluasi konvensional seperti radiografi 2D dan histomorfometri memiliki keterbatasan dalam subjektivitas dan efektivitas. Artificial Intelligence (AI) menawarkan solusi inovatif untuk meningkatkan presisi dan kecepatan analisis. Tujuan penelitian ini untuk menganalisis aplikasi dan pengembangan AI dalam analisis osseointegrasi implan gigi menggunakan radiografi melalui scoping review. Metode: Pencarian sistematis dilakukan di PubMed, Scopus, MEDLINE, Embase, dan Web of Science (2014–2024) dengan kerangka PCC (Population: pasien implan gigi, Concept: AI, Context: klinis) menggunakan framework Arksey dan O’Malley serta panduan dari Joanna Briggs Institute (JBI). Hasil: Dari 11 artikel terpilih (2019–2024), mayoritas menggunakan radiografi periapikal dan CBCT sebagai modalitas utama, dengan model deep learning berbasis CNN (Convolution Neural Network) (seperti YOLOv7 dan ResNet-50) menunjukkan kinerja optimal dalam memprediksi kehilangan tulang marginal (akurasi 70,2–96,13%) dan stabilitas implan. Radiografi periapikal unggul dalam akurasi (94,74%) dan presisi (100%), sementara CBCT (Cone Beam Computed Tomography) menawarkan analisis volumetrik lebih detail dengan kecepatan pemrosesan hingga 76 ms. Meski demikian, variasi parameter radiografi dan ketergantungan pada dataset kecil (44–2920 gambar) berpotensi menyebabkan overfitting. Kolaborasi multi-institusi dan standarisasi teknik radiografi diperlukan untuk meningkatkan kemampuan AI dalam praktik klinis. Simpulan: Model deep learning (CNN, YOLOv7) dan machine learning (SVM) terbukti efektif dalam analisis osseointegrasi, terutama untuk marginal bone loss menggunakan radiografi periapikal dan CBCT. AI berpotensi merevolusi evaluasi implan gigi, namun implementasi klinis memerlukan validasi eksternal dan standardisasi data.KATA KUNCI: Artificial intelligence, implan gigi, osseointegrasi, deep learning, analisis radiografiCurrent applications and development of artificial intelligence for osseointegration dental implant analysis: Scoping ReviewABSTRACTIntroduction: Osseointegration is a key factor in dental implant success, but conventional evaluations such as two dimensional (2D) radiography and histomorphometry are limited by subjectivity and restricted diagnostic capacity. Artificial Intelligence (AI) offers an innovative solution to improve both precision and speed of analysis. This study aims to explore current applications and advancements in AI-based osseointegration analysis of dental implants using radiographs through a scoping review. Methods: A systematic search was conducted in PubMed, Scopus, MEDLINE, Embase, and Web of Science (2014–2024), using the PCC framework (Population: dental implant patients, Concept: AI, Context: clinical). The review followed the Arksey and O’Malley methodological framework and the Joanna Briggs Institute (JBI) guidelines. Results: Of the 11 selected articles (2019-2024), the majority used periapical radiography and CBCT (Cone Beam Computed Tomography) as the primary imaging modalities, with CNN (Convolution Neural Network)-based deep learning models (such as YOLOv7 and ResNet-50) demonstrated strong predictive performance for marginal bone loss (accuracy 70.2-96.13%) and implant stability. Periapical radiographs achieved high accuracy (94.74%) and precision (100%), while CBCT enabled more detailed volumetric analysis with processing speeds of up to 76 ms. However, variability in radiographic parameters and reliance on small datasets (44-2920 images) could lead to model overfitting. Multi-institutional collaboration and standardization of imaging protocols are required to enhance AI performance and generalizability in clinical practice. Conclusion: Deep learning (CNN, YOLOv7) and machine learning (SVM) models have proven effective in osseointegration analysis, particularly in predicting marginal bone loss using periapical radiographs and CBCT. AI has the potential to revolutionize dental implant evaluation, but clinical implementation requires external validation and data standardization.KEY WORDS: Artificial intelligence, dental implant, osseointegration, deep learning, radiographic analysis]]>
