Outcome-Driven Approach to Minimize the Risks of Facial Distortion Following CMF Surgery

摘要： 我们这个项目的主要临床动机是为患者提供个性化的精确护理， 颌骨（包括上颌骨和下颌骨）畸形，通过显着改善手术计划方法。数量 患颌骨畸形的患者人数每年都在上升。 颌外科是通过骨段复位来治疗颌骨畸形的主要手术方式 的下巴。正颌外科手术的最终结果是由最终的面部外观来判断的。虽然 不直接对面部软组织进行手术，面部随骨骼变化而“自动”变化。 颌外科手术需要广泛的手术计划。虽然我们可以精确地计划骨骼的运动 并在手术过程中使用计算机辅助手术模拟（卡斯）和3D打印将其转移到患者身上， 外科医生仍然无法在手术计划期间实际预测面部变化， 希望术后能“自动”恢复正常面容。然而，这条“精神线索” 这种方法不可靠，因为面部变化并不完全跟随骨骼变化。问题是， 复合缺陷患者的血压更高。例如，如果患者有骨骼畸形和轻度面部缺陷， 外科医生在手术前必须知道如何过度矫正骨骼以掩盖软组织缺陷。但 这些信息只能通过准确的方法来预测面部变化。此外，患者的 从角度来看，最后的面部外观是他们非常关心的。因此，这一点非常重要，因为 医生和病人，以准确地预测面部变化。 在上一个项目期间，我们在预测面部变化方面取得了重大成就 使用有限元（FE）方法跟踪骨骼运动。然而，这种方法仍然需要相当大的 准备FE模型的时间。此外，不是决定最终的手术结果（ 首先，目前的方法仍然是被动地预测面部变化， 骨骼手术这些障碍极大地阻碍了外科医生在临床环境中实际使用它。 我们的假设是，个性化的精确治疗结果只能在外科医生 能够在计划骨外科手术之前确定最终的治疗结果，期望的术后面部。 为了验证我们的假设，我们建议将结果驱动和基于机器学习的技术整合在一起 首先估计期望的术后面部，然后计划骨手术。 拟议的项目将对提高病人护理质量产生重大的临床影响。它将使 临床医生根据面部和骨骼信息，使用单个 软件在日常临床实践中。它还将彻底改变手术计划技术， 驱动方法，即，首先估计期望的术后面部，然后计划骨手术。

Abstract: Our main clinical motivation of this project is to provide personalized precision care to patients with jaw (both maxilla and mandible) deformities by significantly improving surgical planning method. The number of patients suffering from jaw deformities is escalating each year. Orthognathic surgery is a main surgical procedure to treat jaw deformities by repositioning bony segments of the jaws. The ultimate outcomes of orthognathic surgery are judged by the final facial appearance. Although the facial soft tissues are not directly operated on, the face changes “automatically” following the bony changes. Orthognathic surgery requires extensive surgical planning. While we can accurately plan the bony movements and transfer it to the patient during the surgery using computer-aided surgical simulation (CASS) and 3D printing, surgeons are still unable to practically predict the facial changes during the surgical planning, and just hope that a postoperative normal face will be “automatically” restored. However, this “mental-clue” approach is not reliable because the facial change does not exactly follow bony change. The problem is even bigger in patients with composite defects. For example, if a patient has a skeletal deformity and mild facial defect, a surgeon must know, before surgery, how to overcorrect the skeleton to camouflage the soft-tissue defect. But this information can only be attained by accurate method to predict facial changes. In addition, from patient’s perspective, the final facial appearance is great concern to them. Therefore, it is extremely important, for both doctors and patients, to accurately predict facial changes. In the previous project period, we have made significant achievements in predicting facial changes following bony movements using finite element (FE) method. However, this approach still requires a considerable amount of time to prepare FE models. In addition, rather than determining the ultimate surgical outcome (the postoperative facial appearance) first, the current method is still to predict the facial change passively following the bony surgery. These hurdles greatly prevent surgeons from practically using it in the clinical setting. Our hypothesis is that a personalized precision treatment outcome can only be achieved if surgeons are able to determine the final treatment outcome, a desired postoperative face, before planning the bony surgery. To test our hypothesis, we propose to integrate outcome-driven and machine learning-based techniques together to first estimate a desired postoperative face, and then plan the bony surgery. The proposed project will have a significant clinical impact on improving patient care quality. It will enable clinicians to develop an optimal surgical plan based on both facial and bony information, on-the-fly, using a single software in their routine clinical practice. It will also revolutionize the surgical planning technique using outcome- driven approach, i.e., to first estimate a desired postoperative face and then plan the bony surgery.