Learning-Based Approach for Personalized Craniomaxillofacial Surgical Planning

摘要： 我们这个项目的主要临床目标是为患有以下疾病的患者提供个性化的精准护理 通过显著改进手术计划方法，改善颅颌面(CMF)畸形。CMF 畸形包括先天和后天的颌骨和面部畸形。中国有大量的病人 美国和世界各地都患有CMF畸形。CMF手术的基本原则包括 截骨、自体、骨修复畸形CMF结构恢复正常解剖和功能 移植物，或带血管的游离皮瓣。CMF手术的成功不仅取决于手术的技术方面 手术，在很大程度上也是手术计划的精确制定。然而，CMF手术计划 由于CMF解剖和畸形的复杂性，极具挑战性。在例行CMF期间 外科手术计划，外科医生首先获取病人头骨的三维(3D)模型。然后他表演 三维头影测量分析以量化畸形。最后，通过虚拟切割3D来模拟手术 建模为多个骨骼分段。然后，外科医生会尽最大努力分别移动和旋转每个节段 至正常头影测量值范围内的理想位置(目前的护理标准)。这是 有问题的“正常”头影测量值是正常人群的平均值，其中每个值 有一个平均值和一个标准差。由于正常值的变化，外科医生必须经常 猜猜患者的头影测量应该校正到什么值。此外, 头影测量只是一组线性和角度的测量，它肯定不能代表复杂性 人类CMF结构的性质。因此，手术结果往往是主观的，并且在很大程度上取决于 外科医生的经验和艺术天赋。因为每个人的脸都是不同的，所以平均来说， 值“不能代表每一张脸的复杂形态。 为此，我们假设，如果外科医生能够预测患者的正常CMF形状， 手术计划将是客观和个性化的。因此，在这个项目中，我们建议开发和 验证一种新的手术计划方法，该方法使用患者特定的和解剖学上正确的参考模型。 我们的初步研究已经证明了我们方法的可行性。这个项目的成果 将通过开发个性化和精准的外科手术显著提高患者护理质量 客观制定CMF手术方案。这一结果将对颌骨畸形患者特别有利， 症状性和非症状性颅面畸形、创伤和CMF癌。在未来，我们的方法可以 还可用于设计和打印具有组织工程能力的3D患者专用可吸收骨植入物 用于骨骼再生。

Abstract: Our main clinical objective for this project is to provide personalized precision care to patients with craniomaxillofacial (CMF) deformities by significantly improving the surgical planning method. CMF deformities involve congenital and acquired deformities of the jaws and face. A large number of patients in the US and around the world suffer from CMF deformities. The basic principles of CMF surgery involve the restoration of deformed CMF structures back to normal anatomy and functions with osteotomy, autologous, bone grafts, or vascularized free flaps. The success of CMF surgery depends on not only the technical aspects of the operation, but also, to a large extent, the precise formulation of a surgical plan. However, CMF surgical planning is extremely challenging due to the complex nature of CMF anatomy and deformity. During a routine CMF surgical planning, a surgeon first acquires a three-dimensional (3D) model of the patient's skull. He then performs 3D cephalometric analysis to quantify the deformity. Finally, the surgery is simulated by virtually cutting the 3D model into multiple bony segments. The surgeon then tries his best to move and rotate each segment individually to a desired position within the normal range of cephalometric values (the current standard of care). This is problematic as “normal” cephalometric values are the averageness of normal population, in which each value has a mean and a standard deviation. Due to the variation within the normal values, the surgeon must often guess what the exact value the patient's cephalometric measurement should be corrected to. In addition, cephalometry is a group of only linear and angular measurements, which certainly cannot represent the complex nature of human CMF structures. Therefore, surgical outcomes are often subjective and heavily dependent on the surgeons' experience and artistic talent. Because each human face is different, the average “normal values” cannot represent the complex morphology of each individual face. To this end, we hypothesize that if a surgeon can foresee what the normal CMF shape of the patient should be, the surgical plan will be objective and personalized. Therefore, in this project, we propose developing and validating a new surgical planning method of using patient-specific and anatomically-correct reference models. The feasibility of our approach has already been proven by our preliminary studies. The results of this project will significantly improve the quality of patient care by developing personalized and precise surgical plans for CMF surgery objectively. The results will be especially beneficial to patients with jaw deformities, syndromic and non-syndromic craniofacial deformities, trauma, and CMF cancer. In the future, our approach can also be used to design and print 3D patient-specific resorbable bone implants with tissue engineering capability for bone regeneration.