研究背景

下颌骨是构成面下1/3 的骨性支架，对维持正常的面部外形起重要作用。借助附着的肌肉、韧带完成复杂的运动，进而实现吞咽、咀嚼、呼吸和语言等重要功能。累及下颌骨的病变，如颌面部良恶性肿瘤、感染及创伤往往会造成下颌骨节段性缺损 [1]，即下颌骨的连续性中断。这不仅会造成患者颌面部畸形，还严重影响患者咀嚼、吞咽和语言等功能并对患者造成沉重的心理负担和精神压力。
对下颌骨进行功能重建，最大限度恢复患者外貌和功能，可以极大的改善患者生活质量及身心健康，并为其回归社会提供巨大帮助。同时这也对颌面重建外科医生提出了很高的要求。
目前临床中最常用的修复方法是血管化腓骨瓣修复重建下颌骨缺损术，为提高重建的精准度，临床医生在数字化技术（主要包括数字化导板和术中导航）的辅助下可以将直线型的腓骨塑形成符合生理结构的、弧形的下颌骨。我们对传统数字化导板进行一定的改进，使其具备指导截骨和就位两种功能，希望能进一步提高手术精准度。
本实验拟通过回顾分析的方法来评价截骨就位复合导板在临床中的使用效果，并且尝试对混合现实技术在血管化腓骨瓣修复重建颌骨缺损术中应用的可行性进行探索性研究。
本实验共分为两部分：
第一部分：截骨就位复合导板在血管化腓骨瓣颌骨缺损术中的精准性研究
目的：验证截骨就位复合导板在血管化腓骨瓣重建颌骨缺损术中的精确性。
方法：选取自2014年10月至2018年1月于我院口腔颌面外科行血管化腓骨瓣下颌骨重建患者共10例。男7例，女3例，其中下颌骨成釉细胞瘤2例、粘液瘤2例、牙龈癌2例、口底癌1例、下颌骨透明细胞癌1例、骨化纤维瘤1例和下颌骨缺损二期重建1例。所有患者术前均行薄层头颅CT或CBCT，以及双下肢CTA检查，将数据导入Proplan CMF 3.0软件中设计重建方案，根据肿瘤性质、下颌骨受累范围行原发灶切除和腓骨瓣重建设计，之后借助快速成型技术制作颌骨截骨就位复合导板，并且预弯钛板。术中在导板的引导下行下颌骨病灶的切除、腓骨瓣的制备塑形、钛板固定和腓骨瓣的就位。术后一周患者拍摄头颅CT或CBCT，将术后影像数据通过Proplan CMF 3.0软件进行三维重建，并导入Geomagic软件将术后数据与术前设计方案进行拟合，验证复合导板的精确性。
结果：10例患者均按照术前设计顺利完成手术，术后腓骨肌皮瓣全部成活。经Geomagic拟合测量后整体重建误差为1.350mm±0.228mm（最大值：6.799mm），重建腓骨段误差为0.957mm±0.296mm（最大值为4.640mm），余留下颌骨误差为1.137mm±0.371mm（最大值为6.799mm）。术后数据分析显示，腓骨段重建位置与术前设计方案有较好的重合。术后随访显示患者咬合关系良好，对外形均较满意。
结论：根据术后拟合验证结果，它的精确性与导航技术以及传统导板相近，并且可以在术中指导钛板就位，提高下颌骨缺损修复重建的精度，达到理想的临床治疗效果，我们认为复合导板是一种简单、实用的方法。
关键词：颌骨截骨就位复合导板；血管化腓骨瓣；重建
第二部分 混合现实技术在血管化腓骨瓣重建颌骨缺损中应用的可行性研究
目的：对混合现实技术在血管化腓骨瓣颌骨缺损重建术中的应用可行性进行探索。
方法：共选取7根腓骨标本和1个下颌骨标本，所有标本均拍摄CT，将CT数据导入专业设计软件Proplan CMF 3.0，使用下颌骨标本建立标准的缺损模型，缺损累及同侧下颌角和尖牙区，为James Brown最新分类中II类缺损。运用这7根腓骨进行重建方案设计，在混合现实技术的辅助下进行腓骨截骨塑形，将腓骨段用钛板固定就位，完成重建，再次用混合现实设备HoloLens进行验证，之后拍摄CT，将术后重建模型与术前设计方案进行拟合，测算误差。
结果：运用Geomagic软件对术后重建模型进行拟合验证，有3组与术前设计方案基本相符，误差值为0.658±0.075mm，另外4组模型在拟合时软件提示“24%~38%不等的点距离模型太远，无法用于计算”，经Proplan重新对长度和重建角度测量计算后，4组模型长度的平均误差值为2.162±0.62mm，在重建角度上，第2组和第7组下颌角与术前设计方案相差4度，导致升支段腓骨无法与术前设计方案重合。而第1组和第5组模型则是因为下颌骨体段腓骨发生了旋转，导致拟合时与术前设计方案偏差较大。
结论：混合现实技术在医学中的应用尚处于起步阶段，目前HoloLens还无法实现空间精准定位，难以保证腓骨瓣重建颌骨缺损的精度。但混合现实技术为外科医生提供了一种可能，该技术在血管化腓骨瓣精确重建颌骨缺损中的应用仍需进一步探索与改进。
关键词：混合现实；HoloLens；血管化腓骨瓣；重建。

The mandible constructs the lower 1/3 of the facial structures and plays an influential role in maintaining a natural facial appearance. With the aid of attached muscles and ligaments, complex movements are carried out to achieve substantial functions including swallowing, mastication, respiration, and speech. Mandibular lesions such as benign and malignant tumors, infections, and trauma in the maxillofacial region can often cause mandibular segmental defect (MSD), that is, continuous interruption of the mandible. This will not only cause facial deformation, but also substantially affects the patient’s mastication, swallowing, and speech functions, causing psychological burden and mental stress.

Mandibular reconstruction maximumly recovers the patient’s facial appearance and functions, and can extensively improve the patient’s life quality, physical and mental health, assisting them to return to the society. However, this sets a high standard for facial reconstruction surgeons.

In current clinical practices, vascularized fibular flap technique is most commonly used to reconstruct mandibular defects. To increase the accuracy of the reconstruction, surgeons often use digital technology including intraoperative navigation and CAD/CAM guide template to modify the straight-lined fibula into the biological curved shape of the mandible. We modified the conventional CAD/CAM guide template to function in both guiding osteotomy and positioning Ti-plate. 

This experiment consists of two parts: one is a retrospective study on the clinical effects of application of cutting and placing composite template, another is to evaluate the feasibility of the mixed reality technique in mandibular reconstruction with vascularized fibular flap. 

1. The research on accuracy of mandibular cutting and placing composite template 

in mandible reconstruction with vascularized fibula flap.

Objective: To evaluate the accuracy of mandibular cutting and placing composite template in mandible reconstruction with vascularized fibula flap.

Methods: There are 10 patients (7 male to 3 female) with mandible segment defects caused by various reasons, who have received mandibular reconstruction with free vascularized fibular flap from October 2014 to January 2018 at our department were studied retrospectively. Disease classification included 2 mandibular ameloblastomas, 2 mandibular myxomas, 2 mandibular gingival carcinomas, 1 carcinoma of floor of mouth, 1 mandibular clear cell carcinoma, 1 mandibular ossifying fibroma and 1 secondary-stage reconstructions after tumor resection. Before surgery, all the patients received CTA of bilateral lower extremities and CBCT or thin-layer CT scanning of upper and lower jaw. Using software such as Proplan CMF 3.0 we simulated the excision and the design of reconstruction according to the nature of tumor and the involvement of mandible. The surgical model was made by rapid proto-typing technology. Whilst, we designed and manufacture the mandibular cutting and placing templates to guide the tumor resection, fibular osteotomy and fibular emplacement for patients intraoperatively. All of the patients received Cone Beam CT (CBCT) 1 week after surgery. And the data were imported into Geomagic software to analyze the error. 

Results: All the operative processes went smoothly and whole flaps survived. During the follow-up of 6 months, one patient died of local recurrence and metastasis, the others showed good prognoses. Postoperative CT and 3D error analysis revealed that osteotomy lines and reconstruction contour matched well with preoperative planning. Compared with preoperative positions, the average shift on the total mandible was 1.350mm±0.228mm（largest, 6.799mm）. Average shift on remaining mandible was 1.137±0.502 mm (largest, 1.886 mm). Average shift in the reconstructed mandible was 0.957±0.300 mm, largest being 2.441 mm. 

Conclusion: According to the fitting result, the accuracy of the mandibular cutting and placing template equals to the navigation technique and conventional template. It could guide positioning Ti-plate. the can improve the accuracy of surgery and obtain good surgical outcomes. We think that it’s a simple and useful method.

Key Words: Mandibular cutting and placing composite template; vascularized fibular flap; reconstruction.

2. The feasibility of the application of mixed reality in mandible reconstruction with vascularized fibula flap

Purpose: To study the feasibility of the application of mixed reality in mandible reconstruction with vascularized fibula flap.

Method: There are 7 fibulas and 1 mandible bone, which were all received CT examination. Using software such as Proplan CMF 3.0 we made a model of mandibular segmental defects which involves ipsilateral mandible angle and canine. The defect conforms to the classification raised by James Brown. Then we simulated the design of reconstruction by these 7 fibulas. The surgical plans were transferred to HoloLens. We used HoloLens to finish fibular osteotomy and fibular emplacement. After we fixed fibular segments by Ti-template, all of them received CT examination. And the data were imported into Geomagic software to analyze the error.

Result: We validate the error of the reconstructive model by Geomagic. We found that 3 groups obtained satisfactory results, the deviation is 0.658±0.075mm, the other 4 groups got large deviations about the contact angle of two fibula segments. Then we import these 4 groups data into Proplan CMF 3.0 to measure these lengths and angles. The average error of length is 2.16±0.62mm. For the reconstructive angle of mandible, the deviation of 2nd and 7th group are 3.6°and 3.9°. The mandibular body of 1st and 5th rotated, so they don’t match well.

Conclusion: Nowadays, mixed reality provides a possible hands-free approach as get rid of conventional guide template and screen. But HoloLens meet the demand of spatial accurate localization. It can’t guarantee the exact mandibular reconstruction with vascularized fibular flap. It needs to be improved. 

Key Words: Mixed Reality; HoloLens; vascularized fibular flap; reconstruction.

The mandible constructs the lower 1/3 of the facial structures and plays an influential role in maintaining a natural facial appearance. With the aid of attached muscles and ligaments, complex movements are carried out to achieve substantial functions including swallowing, mastication, respiration, and speech. Mandibular lesions such as benign and malignant tumors, infections, and trauma in the maxillofacial region can often cause mandibular segmental defect (MSD), that is, continuous interruption of the mandible. This will not only cause facial deformation, but also substantially affects the patient’s mastication, swallowing, and speech functions, causing psychological burden and mental stress.

Mandibular reconstruction maximumly recovers the patient’s facial appearance and functions, and can extensively improve the patient’s life quality, physical and mental health, assisting them to return to the society. However, this sets a high standard for facial reconstruction surgeons.

In current clinical practices, vascularized fibular flap technique is most commonly used to reconstruct mandibular defects. To increase the accuracy of the reconstruction, surgeons often use digital technology including intraoperative navigation and CAD/CAM guide template to modify the straight-lined fibula into the biological curved shape of the mandible. We modified the conventional CAD/CAM guide template to function in both guiding osteotomy and positioning Ti-plate. 

This experiment consists of two parts: one is a retrospective study on the clinical effects of application of cutting and placing composite template, another is to evaluate the feasibility of the mixed reality technique in mandibular reconstruction with vascularized fibular flap. 

1. The research on accuracy of mandibular cutting and placing composite template 

in mandible reconstruction with vascularized fibula flap.

Objective: To evaluate the accuracy of mandibular cutting and placing composite template in mandible reconstruction with vascularized fibula flap.

Methods: There are 10 patients (7 male to 3 female) with mandible segment defects caused by various reasons, who have received mandibular reconstruction with free vascularized fibular flap from October 2014 to January 2018 at our department were studied retrospectively. Disease classification included 2 mandibular ameloblastomas, 2 mandibular myxomas, 2 mandibular gingival carcinomas, 1 carcinoma of floor of mouth, 1 mandibular clear cell carcinoma, 1 mandibular ossifying fibroma and 1 secondary-stage reconstructions after tumor resection. Before surgery, all the patients received CTA of bilateral lower extremities and CBCT or thin-layer CT scanning of upper and lower jaw. Using software such as Proplan CMF 3.0 we simulated the excision and the design of reconstruction according to the nature of tumor and the involvement of mandible. The surgical model was made by rapid proto-typing technology. Whilst, we designed and manufacture the mandibular cutting and placing templates to guide the tumor resection, fibular osteotomy and fibular emplacement for patients intraoperatively. All of the patients received Cone Beam CT (CBCT) 1 week after surgery. And the data were imported into Geomagic software to analyze the error. 

Results: All the operative processes went smoothly and whole flaps survived. During the follow-up of 6 months, one patient died of local recurrence and metastasis, the others showed good prognoses. Postoperative CT and 3D error analysis revealed that osteotomy lines and reconstruction contour matched well with preoperative planning. Compared with preoperative positions, the average shift on the total mandible was 1.350mm±0.228mm（largest, 6.799mm）. Average shift on remaining mandible was 1.137±0.502 mm (largest, 1.886 mm). Average shift in the reconstructed mandible was 0.957±0.300 mm, largest being 2.441 mm. 

Conclusion: According to the fitting result, the accuracy of the mandibular cutting and placing template equals to the navigation technique and conventional template. It could guide positioning Ti-plate. the can improve the accuracy of surgery and obtain good surgical outcomes. We think that it’s a simple and useful method.

Key Words: Mandibular cutting and placing composite template; vascularized fibular flap; reconstruction.

2. The feasibility of the application of mixed reality in mandible reconstruction with vascularized fibula flap

Purpose: To study the feasibility of the application of mixed reality in mandible reconstruction with vascularized fibula flap.

Method: There are 7 fibulas and 1 mandible bone, which were all received CT examination. Using software such as Proplan CMF 3.0 we made a model of mandibular segmental defects which involves ipsilateral mandible angle and canine. The defect conforms to the classification raised by James Brown. Then we simulated the design of reconstruction by these 7 fibulas. The surgical plans were transferred to HoloLens. We used HoloLens to finish fibular osteotomy and fibular emplacement. After we fixed fibular segments by Ti-template, all of them received CT examination. And the data were imported into Geomagic software to analyze the error.

Result: We validate the error of the reconstructive model by Geomagic. We found that 3 groups obtained satisfactory results, the deviation is 0.658±0.075mm, the other 4 groups got large deviations about the contact angle of two fibula segments. Then we import these 4 groups data into Proplan CMF 3.0 to measure these lengths and angles. The average error of length is 2.16±0.62mm. For the reconstructive angle of mandible, the deviation of 2nd and 7th group are 3.6°and 3.9°. The mandibular body of 1st and 5th rotated, so they don’t match well.

Conclusion: Nowadays, mixed reality provides a possible hands-free approach as get rid of conventional guide template and screen. But HoloLens meet the demand of spatial accurate localization. It can’t guarantee the exact mandibular reconstruction with vascularized fibular flap. It needs to be improved. 

Key Words: Mixed Reality; HoloLens; vascularized fibular flap; reconstruction.