Stefano Pieralli^1,2,*; Christoph Beyer¹; Christian Wesemann¹; Kirstin Vach³; Florian Kernen⁴; Katja Nelson⁴; Benedikt Christopher Spies¹

¹Department of Prosthetic Dentistry, Center for Dental Medicine, Medical Center - University of Freiburg, Faculty of Medicine - University of Freiburg, Freiburg, Germany; ²Department of Prosthodontics, Dental School, USP - Universidade de São Paulo, São Paulo, Brazil; ³Institute of Medical Biometry and Statistics, Faculty of Medicine and Medical Center - University of Freiburg; ⁴Department of Oral and Maxillofacial Surgery, Center for Dental Medicine, Medical Center - University of Freiburg, Faculty of Medicine - University of Freiburg, Freiburg, Germany

Background: Virtual implant planning by means of a backward planning is a key factor to achieve prosthetically pleasing results. For this purpose, intraoral surface and 3D radiographic datasets, mostly exceeding the area of surgical interest and associated with increased exposure to ionizing radiation, are merged using dedicated software. To date, little is known regarding the necessary extension of the irradiated field of view (FOV) required for accurate matching of the aforementioned data.

Aim/Hypothesis: To assess the accuracy of virtual implant positioning in case of a single missing tooth (ST) or multiple posterior missing teeth (MT) without and with radiographic artifacts (AR) using different FOV extensions. Null hypothesis assumed that FOV extension has no effect on matching accuracy.

Material and Methods: Intraoral radiographic and surface datasets of n = 15 patients corresponding to ST (n = 5), MT (n = 5) and AR (n = 5), were retrospectively collected. Thereafter, the original full-arch (F) FOV of each case was reduced to the quadrant (Q) of the surgical site or to the adjacent tooth/teeth (A) using a beta version of a planning software. A trained operator in dataset merging performed the alignments with the intraoral scans, while an expert delivered the reference. The combination F-expert was considered the reference. To assess discrepancies, all datasets were imported in a measuring software and aligned to the reference. Horizontal and vertical deviations were calculated at the implant apex and shoulder. Linear mixed models with random intercepts were fitted for each implant position and coordinate to evaluate differences between the different FOV. The method of Scheffé was applied to correct for multiple testing and probability level for statistical significance was set to p < 0.05.

Results: Ten superimpositions per FOV extension (n = 3) of the included cases (n = 15) were performed by the trained operator resulting in a total of 450 alignments. Overall, highest mean discrepancies to FA-expert resulted in the AR-A group: 0.10 ± 0.33 mm (apex-horizontal) and -0.09 ± 0.25 mm (apex/shoulder-vertical). For the ST group, largest deviations for each volume resulted -0.06 ± 0.2 mm (F), -0.06 ± 0.1 mm (Q) and -0.05 ± 0.15 mm (A), respectively, and there were horizontally significant differences between F and S at both the implant apex and shoulder (p≤0.05). In the MT group, the main discrepancy resulted for Q: -0.07 ± 0.24 mm (apex-horizontal). However, no significant differences between F, Q and A could be assessed. Finally, the AR group showed greatest variation within the A subvolume with significant vertical differences compared to F at the level of the implant shoulder (p≤0.05).

Conclusion and Clinical implications: Superimpositions with different FOV extensions significantly affected the expected implant position. Therefore, the null-hypothesis was rejected. However, for the selected cases, maximum mean deviations from the reference of ≤0.1 mm suggest no clinical relevance of the assessed inaccuracies. Regarding an increased amount of implants/units to be installed, discrepancies calculated for the AR-A group might indicate a reduced FOV as potential risk-factor for inaccuracies in presence of artifacts.

Disclosure of Interest: None Declared

Keywords: accuracy, CBCT, intraoral scanner

斯特凡诺·皮拉利^1,2,* ; 克里斯托夫·拜尔¹ ; 克里斯蒂安·韦斯曼¹ ; 克尔斯汀·瓦赫³ ; 弗洛里安·克南⁴ ; 卡佳纳尔逊⁴ ; 本尼迪克特克里斯托弗间谍¹

¹修复牙科系，牙科医学中心，医学中心 - 弗莱堡大学，医学院 - 弗莱堡大学，德国弗莱堡；²口腔修复学系，USP - Universidade de São Paulo，巴西圣保罗；³医学生物计量和统计研究所，医学和医学中心 - 弗莱堡大学；⁴口腔颌面外科，牙科医学中心，医学中心 - 弗莱堡大学，医学院 - 弗莱堡大学，弗莱堡，德国

背景：通过反向规划进行的虚拟种植体规划是获得令人愉悦的修复效果的关键因素。为此，使用专用软件合并口内表面和 3D 射线照相数据集，其中大部分超出了手术感兴趣的区域并与增加的电离辐射暴露相关。迄今为止，对于准确匹配上述数据所需的辐照视场 (FOV) 的必要扩展知之甚少。

目的/假设：在单个缺失牙 (ST) 或多个后牙缺失牙 (MT) 的情况下，使用不同的 FOV 扩展评估虚拟种植体定位的准确性。零假设假设 FOV 扩展对匹配精度没有影响。

材料和方法：的口内的射线照相和表面数据集Ñ对应于ST（= 15位患者Ñ = 5），MT（Ñ = 5）和AR（Ñ= 5)，追溯收集。此后，使用 Beta 版计划软件将每个病例的原始全牙弓 (F) FOV 缩小到手术部位的象限 (Q) 或相邻的牙齿/牙齿 (A)。在数据集合并方面训练有素的操作员执行与口内扫描的对齐，而专家则提供参考。组合 F 专家被认为是参考。为了评估差异，将所有数据集导入测量软件并与参考对齐。计算种植体顶端和肩部的水平和垂直偏差。为每个种植体位置和坐标拟合具有随机截距的线性混合模型，以评估不同 FOV 之间的差异。p < 0.05。

结果：受过训练的操作员对所包含案例 ( n = 15) 的每个 FOV 扩展 ( n = 3) 进行了10 次叠加，总共进行了 450 次对齐。总体而言，与 FA 专家的最高平均差异导致 AR-A 组：0.10 ± 0.33 毫米（顶点水平）和 -0.09 ± 0.25 毫米（顶点/肩垂直）。对于 ST 组，每个体积的最大偏差分别导致 -0.06 ± 0.2 mm (F)、-0.06 ± 0.1 mm (Q) 和 -0.05 ± 0.15 mm (A)，并且 F 和 S 之间存在水平显着差异在种植体顶端和肩部 ( p≤0.05)。在 MT 组中，Q 的主要差异为：-0.07 ± 0.24 毫米（顶点-水平）。但是，无法评估 F、Q 和 A 之间的显着差异。最后，与种植体肩部水平的 F 相比，AR 组在 A 子体积内表现出最大的变化，具有显着的垂直差异（p ≤ 0.05）。

结论和临床意义：具有不同 FOV 扩展的叠加显着影响了预期的种植体位置。因此，原假设被拒绝。然而，对于选定的病例，与参考值的最大平均偏差≤0.1 mm 表明评估的不准确度没有临床相关性。关于要安装的植入物/单元数量的增加，为 AR-A 组计算的差异可能表明 FOV 降低，这是存在伪影时不准确的潜在风险因素。

利益披露：无申报

关键词：准确性，CBCT，口内扫描仪