All-on-4® 治療之植體配置涉及植入位置、長度、直徑及傾斜角度等多項設計參數，不同配置與患者顎骨幾何差異皆可能影響植體周圍骨組織之力學反應。有限元素分析雖可量化不同植體配置之生物力學表現，並結合最佳化演算法搜尋較佳設計，但反覆建模與計算所需時間龐大，限制其即時應用。因此，本研究建立一套納入患者下顎骨幾何特徵之生物力學指標預測方法，以快速預測不同 All-on-4® 植體配置下植體周圍骨組織之應力表現。
本研究首先將六組真實患者下顎骨模型以固定體素間距轉換為 1283 之三維體素資料，並利用三維卷積自動編碼器將高維度顎骨體素資料壓縮為 128 維潛在特徵，作為患者顎骨幾何之低維數值化表徵，並透過自動化有限元素建模建立不同植體位置、長度、直徑及傾斜角度所對應之生物力學指標資料集。將顎骨特徵與植體配置參數共同輸入特徵線性調制模型 (Feature-wise Linear Modulation, FiLM)，建立生物學指標預測模型。並採用 Leave-One-Subject-Out 交叉驗證方式，以評估模型對未參與訓練顎骨之跨個體預測能力。
在最佳解鄰域之驗證中，六組顎骨各 20 個鄰近設計點之平均預測誤差介於 7.469% 至 14.558%，整體平均誤差為 10.257%。於最佳化搜尋前期、中期及後期代表點之驗證中，各階段平均預測誤差分別為 10.17%、10.13% 及 11.70%，共 18個設計點之整體平均誤差為 10.67%
綜合上述結果，FiLM 模型於未參與訓練之臨床患者顎骨上，仍能根據顎骨幾何特徵與植體配置參數預測其生物力學指標，本研究所建立之 FiLM 模型具備應用於不同臨床顎骨條件下，快速評估 All-on-4® 植體配置生物力學表現之可行性，並具有作為有限元素分析代理模型及後續植體配置最佳化工具之潛力。

All-on-4® implant placement involves multiple design parameters, while patient-specific mandibular geometry can influence peri-implant bone stress. Although finite element analysis (FEA) can evaluate biomechanical performance, repeated simulations are computationally expensive. This study developed a rapid biomechanical prediction model incorporating mandibular geometric features.
Six patient-specific mandibular models were converted into 1283 voxel data and compressed into 128-dimensional latent features using a three-dimensional convolutional autoencoder. Automated FEA was used to generate biomechanical data for different implant configurations. Mandibular features and implant parameters were then integrated into a Feature-wise Linear Modulation (FiLM) model, and Leave-One-Subject-Out cross-validation was used to evaluate prediction performance on unseen mandibles.
For 20 design points near the optimal solution of each mandible, the mean prediction error ranged from 7.469% to 14.558%, with an overall mean of 10.257%. Across 18 representative points from different optimization stages, the overall mean error was 10.67%.
The results indicate that the FiLM model can provide rapid biomechanical predictions for unseen mandibular geometries and has potential as an FEA surrogate for All-on-4® implant configuration optimization.

[1] Tribst, João Paulo Mendes, et al. "Influence of framework material and posterior implant angulation in full-arch all-on-4 implant-supported prosthesis stress concentration." Dentistry Journal 10.1 (2022): 12.
[2] Maló, Paulo, Bo Rangert, and Miguel Nobre. "“All‐on‐Four” immediate‐function concept with Brånemark System® implants for completely edentulous mandibles: a retrospective clinical study." Clinical implant dentistry and related research 5 (2003): 2-9.
[3] Mattsson, Torsten, et al. "Implant treatment without bone grafting in severely resorbed edentulous maxillae." Journal of oral and maxillofacial surgery 57.3 (1999): 281-287.
[4] Krekmanov, Leonard, et al. "Tilting of posterior mandibular and maxillary implants for improved prosthesis support." International Journal of Oral & Maxillofacial Implants 15.3 (2000).
[5] Maló, Paulo, B. O. Rangert, and Miguel Nobre. "All‐on‐4 immediate‐function concept with Brånemark System® implants for completely edentulous maxillae: a 1‐year retrospective clinical study." Clinical implant dentistry and related research 7 (2005): s88-s94.
[6] Maló, Paulo, et al. "“All‐on‐4” immediate‐function concept for completely edentulous maxillae: a clinical report on the medium (3 years) and long‐term (5 years) outcomes." Clinical implant dentistry and related research 14 (2012): e139-e150.
[7] Malo, Paulo, et al. "A longitudinal study of the survival of All-on-4 implants in the mandible with up to 10 years of follow-up." The Journal of the American Dental Association 142.3 (2011): 310-320.
[8] Maló, Paulo, et al. "The All‐on‐4 treatment concept for the rehabilitation of the completely edentulous mandible: A longitudinal study with 10 to 18 years of follow‐up." Clinical implant dentistry and related research 21.4 (2019): 565-577.
[9] Lopes, Armando, et al. "The N obel G uide® A ll‐on‐4® Treatment Concept for Rehabilitation of Edentulous Jaws: A Prospective Report on Medium‐and Long‐Term Outcomes." Clinical Implant Dentistry and Related Research 17 (2015): e406-e416.
[10] Himmlova, Lucie, et al. "Influence of implant length and diameter on stress distribution: a finite element analysis." The Journal of prosthetic dentistry 91.1 (2004): 20-25.
[11] Gümrükçü, Zeynep, and Yavuz Tolga Korkmaz. "Influence of implant number, length, and tilting degree on stress distribution in atrophic maxilla: a finite element study." Medical & biological engineering & computing 56.6 (2018): 979-989.
[12] Fazi, Giovanni, et al. "Three-dimensional finite element analysis of different implant configurations for a mandibular fixed prosthesis." International Journal of Oral & Maxillofacial Implants 26.4 (2011).
[13] Takahashi, Takeshi, Ichiro Shimamura, and Kaoru Sakurai. "Influence of number and inclination angle of implants on stress distribution in mandibular cortical bone with All-on-4 Concept." Journal of prosthodontic research 54.4 (2010): 179-184.
[14] Sezer, Taygun, Kerem Kilic, and Emir Esim. "Effect of anterior implant position on biomechanical performance in the maxillary all-on-four treatment: a 3-D finite element analysis." Journal of Oral Implantology 48.3 (2022): 177-186.
[15] Gümrükçü, Zeynep, Yavuz Tolga Korkmaz, and Fatih Mehmet Korkmaz. "Biomechanical evaluation of implant-supported prosthesis with various tilting implant angles and bone types in atrophic maxilla: A finite element study." Computers in biology and medicine 86 (2017): 47-54.
[16] Liang, Liang, et al. "A deep learning approach to estimate stress distribution: a fast and accurate surrogate of finite-element analysis." Journal of The Royal Society Interface 15.138 (2018): 20170844.
[17] Madani, Ali, et al. "Bridging finite element and machine learning modeling: stress prediction of arterial walls in atherosclerosis." Journal of biomechanical engineering 141.8 (2019): 084502.
[18] Tonutti, Michele, Gauthier Gras, and Guang-Zhong Yang. "A machine learning approach for real-time modelling of tissue deformation in image-guided neurosurgery." Artificial intelligence in medicine 80 (2017): 39-47.
[19] Mohammadizand, Mahmoud, Massoud Shariat-Panahi, and Morad Karimpour. "Surrogate-based positioning optimization of hip prostheses for minimal stress shielding." Medical Engineering & Physics 135.1 (2025): 104273.
[20] Perez, Ethan, et al. "Film: Visual reasoning with a general conditioning layer." Proceedings of the AAAI conference on artificial intelligence. Vol. 32. No. 1. 2018.
[21] 邱靖軒, "基於有限元素分析之下顎All-on-4®植體配置最佳化與臨床醫師決策之比較" 國立成功大學機械工程系學位論文,2025.
[22] Chen, Yung-Chung, et al. "Real-time optimization of prosthetic design for complete arch implant-supported treatments using finite element-based machine learning." The Journal of Prosthetic Dentistry (2024).