Hierl Thomas, Kruber Daniel, Doerfler Hans-Martin, Huempfner-Hierl Heike, Krause Matthias
Head, Department of Oral & Maxillofacial Plastic Surgery, Helios Vogtland-Klinikum Plauen, Plauen, Germany.
Computer Scientist, Leipzig University, Leipzig, Germany.
J Oral Maxillofac Surg. 2019 Aug;77(8):1663-1672. doi: 10.1016/j.joms.2019.04.005. Epub 2019 Apr 12.
Preformed orbital meshes (POMs) are routinely used in orbital traumatology. As the shapes of commercially available POMs differ, we wanted to evaluate whether investigators would use the same POM in a given situation and perform similar virtual positioning. Furthermore, we investigated whether these tasks could be performed more quickly and more reproducibly using a new software workflow.
Four investigators performed virtual planning of orbital fracture treatment in the Brainlab system (Brainlab, Munich, Germany) using 3 different POMs. Time effort, POM selection, minimum-maximum deviations, and fitting within a distance corridor were measured and compared with the new software workflow.
Regarding time effort, a distinct learning curve was seen. An average of 15 minutes was needed, which compared inferiorly with 1 minute using the software algorithm. In POM selection, the individual choice of the investigators differed to a large extent. An identical POM was chosen in only 3 of 10 cases, and in 2 of 10 cases, all investigators and software choices were identical. Manual POM planning showed constant collision between bone and mesh with large interindividual maximum deviations. By defining collision prevention and achieving the highest concordance between bone and mesh geometry within a given distance slot, the new software was able to select and place the best-fitting POM and display regions of potential necessary presurgical modifications.
Planning of orbital fracture treatment in a navigation system using POMs seems highly individual. Integrating new software workflows should reduce time effort and lead to reproducible planning results, which could be used in intraoperative navigation. Furthermore, our new workflow could be used for training and expanded to other anatomic regions.
预成型眼眶网片（POMs）常用于眼眶创伤学。由于市售POMs的形状各异，我们想评估研究人员在给定情况下是否会使用相同的POM并进行相似的虚拟定位。此外，我们研究了使用新的软件工作流程是否能更快、更可重复地完成这些任务。
四名研究人员在Brainlab系统（德国慕尼黑Brainlab公司）中使用3种不同的POMs进行眼眶骨折治疗的虚拟规划。测量了时间花费、POM选择、最小-最大偏差以及在距离走廊内的拟合情况，并与新的软件工作流程进行比较。
在时间花费方面，呈现出明显的学习曲线。平均需要15分钟，与使用软件算法的1分钟相比要差。在POM选择上，研究人员的个人选择差异很大。在10个案例中只有3个选择了相同的POM，在10个案例中的2个，所有研究人员和软件的选择都是相同的。手动POM规划显示骨骼与网片之间持续碰撞，个体间最大偏差较大。通过定义防撞并在给定距离槽内实现骨骼与网片几何形状之间的最高一致性，新软件能够选择并放置最合适的POM，并显示潜在的术前必要修改区域。
在导航系统中使用POMs进行眼眶骨折治疗的规划似乎高度个体化。整合新的软件工作流程应能减少时间花费并产生可重复的规划结果，可用于术中导航。此外，我们的新工作流程可用于培训并扩展到其他解剖区域。
