改良TARP技术与Goel技术治疗颅底凹陷症稳定性的有限元分析

发布时间：2018-05-12 17:18

本文选题：改良Goel技术 + C_2椎弓根螺钉　；参考：《南方医科大学》2015年硕士论文

【摘要】：背景颅底凹陷症(Basilar invagination, BI)是一种以颅颈交界区复杂骨结构畸形为基础的神经脊髓压迫综合征,其发病机制多与胚胎发育过程形成的扁平颅底、枕颈融合、Kleip-Feil畸形等有关,也可能与寰枢椎失稳后代偿有关,常继发于先天性畸形、类风湿性关节炎、甲状旁腺功能亢进、Paget病、成骨不全症和佝偻病等,多表现为寰枢关节脱位,齿状突向后、向上陷入枕骨大孔,压迫脑干,引起颈痛、四肢乏力、感觉麻木等神经症状。BI可分为斜坡型和齿状突型,斜坡型BI的病理解剖特征是,齿状突与寰椎始终保持正常解剖关系,但齿状突跟随寰椎及枕骨斜坡同步上移,导致颅底平坦,后颅窝容积减小,小脑被迫疝出枕骨大孔,从后方压迫脑干,引起相应的神经症状,因此,斜坡型BI一般采用后颅窝减压,扩大后颅窝容积的方法。与斜坡型BI不同,齿状突型BI存在寰椎脱位,并且齿状突向后、向上压迫延髓,因此,该型治疗的关键是复位寰枢椎,同时解除齿状突对延髓的压迫。有学者采用术前长时间卧床牵引或术中全麻下牵引来复位寰椎,然而,寰椎前弓和齿状突之间以及侧块关节之间有大量瘢痕组织形成,甚至侧块关节、寰齿关节间异常骨性融合,导致临床中大部分齿状突型BI的寰枢椎脱位是难复性的,即使在全麻下大重量颅骨牵引也无法复位。目前难复性寰枢椎脱位主要有两种治疗方法,分别是国内学者尹庆水等研制发明的经口咽寰枢复位钢板内固定系统(transoral atlantoaxial reduction plate,TARP)和印度学者Goel最初报道的后路寰枢植入垫片联合钉棒(或钉板)内固定系统(C1 lateral mass screw+C2 pedicle screw+Cage, C1LS+C2PS+Cage)。2004年印度学者Goel最初报道采用后路切断C2神经节及静脉丛,暴露并分离寰枢关节,去除软骨后植入自行设计的多孔金属垫片,迫使齿状突下移,最后辅以寰枢内固定,以上治疗BI的方法,我们称之为‘'Goel技术”(Goel technique)。2013年Chandra等指出,Goel技术仅仅复位齿状突垂直移位,并没有复位寰椎前脱位,他们在Goel技术的基础上利用植入的Cage为支点,通过器械对万向螺钉加压来复位寰椎,最后辅以寰枢内固定,以上以Cage为支点通过器械加压来复位寰椎的方法,我们称之为“改良Goel技术”(modified Goel technique)。为了减少术中椎动脉医源性损伤,他们采用C2双皮质椎板螺钉(bicortical C2 laminar screws,BC2LS)代替C2椎弓根螺钉(C2 pedicle screws, C2PS)。然而,后路C1侧块螺钉+Cage+C2双皮质椎板螺钉(C1LS+Cage+BC2LS)组成的钉棒系统,其生物力学稳定性未见相关报道。2004年尹庆水等首次报道TARP技术,随后他们应用TARP技术成功治疗大量齿状突型BI合并难复性寰枢椎脱位的患者,并获得良好临床效果。TARP技术不仅能有效复位寰枢椎脱位,而且也能为寰枢固定融合提供良好的生物力学稳定性。目前TARP内固定已更新至第三代,第三代TAPR内固定的主要改良之处是逆向椎弓根置钉方法以及钉板万向导钻和自锁机制的设计,从根本上确保了螺钉的坚强固定。TARP技术通过前路行寰枢融合,使松解、复位、减压、固定一步完成,避免后路手术对颈后肌的破坏,然而TARP技术植入颗粒骨(或髂骨块)存在取髂骨相关并发症,还有植骨塌陷、吸收、移位或脱落的可能,进而导致骨不连、内固定失效或感染发生,含有植骨的Cage取代颗粒骨理论上能增加稳定性,维持寰枢融合角度,并减少植骨塌陷、吸收、脱落等并发症,我们将Cage联合TARP内固定的方法称为“改良TARP技术’'(modified TARP technique,TARP+Cage)。目前国内外文献仅有少数关于改良TAPR技术和Goel技术的生物力学研究,未见有限元分析研究。有限元分析可定量表达颈椎运动,是体外生物力学实验的一个重要补充,它可以克服尸体实验和动物实验取材难、费用高、可重复性差等缺点。2000年,Puttlitz等首次报道上颈椎有限元模型,并将其应用于上颈椎类风湿关节炎的病理学研究,此后,随着有限元软件和计算机科学的发展,上颈椎有限元分析已经广泛应用于颈椎运动学研究,以及各种内固定器械的生物力学研究。实验一两种改良Goel技术治疗颅底凹陷症稳定性的有限元分析目的应用有限元分析评价C2双皮质椎板螺钉和C2椎弓根螺钉联合关节内Cage在寰枢固定中的生物力学差异。方法采集1名35岁正常男性上颈椎(C0-C2)CT数据,通过Mimics 10.01和Abaqus6.11软件建立C0-C2节段三维有限元完整模型(Intact)并进行有效性验证。正常有限元模型包括皮质骨、松质骨、软骨和上颈椎相关韧带,模型包含的韧带有横韧带、十字韧带上下韧带、翼状韧带、齿突尖韧带、前纵韧带、寰枕前膜、覆膜、寰枕后膜、寰枢后膜、C1-C2关节囊,由于横韧带呈低弹性组织且非常坚韧,采用膜单元来模拟,其余韧带参考相关文采用两节点T3D2单元来模拟,设置为只传导拉力。皮质骨的平均厚度设为1.5mm, C1-C2的关节软骨厚度设为3.0mm,关节软骨接触面之间采用滑动接触,摩擦系数设为0.1。皮质骨、松质骨、软骨、关节囊、韧带材料属性根据文献确定赋值。在已建立的Intact模型上,通过删除横韧带模拟横韧带断裂,建立上颈椎失稳模型(Unstable),并与体外颈椎生物力学实验数据比对。另外,BI常合并寰枕融合,合并率达92%,因此在已建立的Unstable模型上,删除Co-C1关节间软骨,添加单元格模拟寰枕融合状态。在失稳模型上分别建立后路C1侧块螺钉+Cage+C2双皮质椎板螺钉组成的钉棒系统模型(Cl lateral mass screw+Cage+bicortical C2 laminar screw, C1LS+Cage+BC2LS),后路C1侧块螺钉+Cage+C2椎弓根螺钉组成的钉棒系统模型(C1 lateral mass screw+Cage+C2 pedicle screw, C1LS+Cage+C2PS)。在枕骨髁上方施加40 N轴向压力模拟头颅重力,同时在枕骨髁上方施加1.5 Nm力矩使模型产生前屈、后伸、侧弯、旋转运动,记录C1LS+Cage+BC2LS组及ClLS+Cage+C2PS组的应力云图及应力峰值,计算C1-C2节段活动度(range of motion, ROM)。结果本研究成功建立正常人Co-C2非线性有限元模型,模型模拟了皮质骨、松质骨、关节软骨及关节囊、韧带的三维结构,共计单元26623个,节点26003个。Intact模型在前屈、后伸、侧弯、旋转载荷下Co-C1、C1-C2的ROM与Panjabi等体外颈椎标本实验结果和Zhang等上颈椎有限元模型结果吻合,验证了正常模型的有效性。Li等的颈椎生物力学研究和Zhang等的上颈椎有限元分析均采用切断横韧带的方法造成寰枢失稳,同样本研究也采用删除横韧带的方法模拟C1-C2失稳,本模型有限元结果表明,与Intact模型相比,Unstable模型在前屈、后伸、侧弯、旋转载荷下C1-C2的活动度分别增加35.2%、 16.4%、4.0%、5.6%,以上结果基本和Li等的颈椎生物力学实验结果吻合。在任何载荷下C1LS+Cage+BC2LS组和C1LS+Cage+C2PS组的C1-C2节段ROM差异均小于0.1°,且两组内固定所有螺钉的应力分布和应力峰值无明显差异。两组内固定系统在前屈、后伸载荷下C1螺钉比C2螺钉承受更大的应力,尤其是后伸载荷下C1螺钉的最大应力是C2螺钉的2倍左右。在后伸载荷下两组内固定Cage内植骨应力最小,存在明显应力遮挡,尤其是C1LS+Cage+C2PS组。结论本研究建立上颈椎有限元模型,并采用两组不同的内固定系统装配,进行生理载荷不同运动状态下的有限元分析结果表明,对于BI的治疗,当枢椎不适合置入C2椎弓根螺钉时,可采用C2双皮质椎板螺钉替代,两者提供的三维稳定性相当,均有利于融合。与C2PS技术相比,BC2LS技术简单、易行,同时能有效避免椎动脉和脊髓的损伤。目前仍需进一步研究颅底凹陷症患者的C2椎板影像学数据,为临床应用C2双皮质椎板螺钉技术治疗颅底凹陷症提供理论依据。实验二改良TARP技术与Goel技术治疗颅底凹陷症稳定性的有限元分析目的应用有限元分析比较前路改良’TARP技术与后路Goel技术治疗颅底凹陷症的生物力学稳定性差异。方法采集1名35岁正常男性上颈椎(C0-C2)CT数据,通过Mimics 10.01和Abaqus6.11软件建立Co-C2节段三维有限元完整模型并进行有效性验证。在失稳模型上参考TARP内固定植入方法建立前路改良TARP内固定模型(transoral atlantoaxial reduction plate+Cage, TARP+Cage),具体方法如下：C1逆向侧块螺钉采用单皮质螺钉,入针点位于Cl侧块内侧缘向外5mm处,向外侧倾斜5°-10°、向上倾斜10°-15。,确保沿着C1侧块轴进入；C2逆向椎弓根螺钉采用双皮质螺钉,入针点位于C2上关节面内侧顶点下方约5mm处,向外侧倾斜9.3°-28.3°、向下倾斜6.5°-2.15°,逆向沿C2椎弓根轴进入。同样,在失稳模型上参考后路C1侧块螺钉和C2椎弓根螺钉植入方法构建后路Goel内固定模型(C1 lateral mass screw+C2 pedicle screw+Cage, C1LS+C2PS+Cage),具体方法如下：C1侧块螺钉采用双皮质螺钉,入针点位于后弓下方C1侧块中部,沿C1-2关节面向内侧倾斜5。-10。、向上倾斜10。-15°,沿着C1侧块轴进入；C2椎弓根螺钉采用双皮质螺钉,遵循“高、内”原则,入针点位于C2椎弓根内侧,沿C1-2关节面向内侧倾斜16.5°-23.8°、向上倾斜25.3°-36.7°,沿着C2椎弓根轴进入。在枕骨髁上方施加40 N轴向压力模拟头颅重力,同时在枕骨髁上方施加1.5 Nm力矩使模型产生前屈、后伸、侧弯、旋转运动,记录TARP+Cage组及C1LS+C2PS+Cage组的应力云图及应力峰值,并计算C1-C2节段活动度(range of motion, ROM)。结果与完整模型相比,两组内固定均能减少C1-C2节段ROM。与C1LS+C2PS+Cage组相比,TARP+Cage组的C1-C2节段ROM在后伸、侧弯、旋转载荷下分别减少44.7%、30.0%、10.5%,但在前屈载荷下ROM增加30.0%。除了在后伸载荷下TARP+Cage组C2螺钉的应力峰值大于ClLS+C2PS+Cage组C2螺钉外,其余在前屈、后伸、侧弯和旋转载荷下TARP+Cage组C1螺钉、C2螺钉的应力峰值均相应的小于C1LS+C2PS+Cage组C1螺钉、C2螺钉。在前屈、后伸载荷下TARP板的应力峰值均小于后路棒。结论本研究建立上颈椎有限元模型,并采用两组不同的内固定系统装配,进行生理载荷不同运动状态下的有限元分析结果表明,与Goel技术相比,改良TARP技术可能在后伸、侧弯、旋转方向上具有更好的三维稳定性,但在前屈方向的稳定性可能不如Goel技术。与Goel技术相比,改良TARP技术不仅在载荷传递和应力分布上更加合理,而且能有效减压、复位和固定寰枢椎,同时获得寰枢生理融合角度,进而获得良好远期疗效,但临床仍需要关于改良TARP技术与Goel技术的前瞻性、随机、多中心临床研究来明确改良TARP技术在寰枢复位、固定融合中的优势。
[Abstract]:Background Basilar invagination (BI) is a kind of nerve spinal cord compression syndrome based on the complex bone structure malformation in the craniocede junction area. Its pathogenesis is related to the flat skull base, occipital neck fusion, Kleip-Feil malformation, and so on. It can also be related to the decompensation of atlantoaxial instability. Chang Jifa is innate Malformation, rheumatoid arthritis, hyperparathyroidism, Paget disease, osteogenesis imperfecta and rickets, such as atlantoaxial dislocation, odontoid backwards, upwards into the occipital foramen, compression of the brain stem, cervical pain, fatigue, and numbness of the limbs,.BI can be divided into the slope and odontoid, the pathological anatomy of the ramp type BI The odontoid process and Atlas always maintain normal anatomical relationship, but the odontoid process follows the atlas and occipital slope synchronously, which leads to the flat skull base, the decrease of the volume of the posterior fossa, the cerebellum forced to herniate the occipital foramen, the brain stem from the rear, and the corresponding neurological symptoms. Therefore, the posterior fossa decompression is generally used in the slope type BI to enlarge the volume of the posterior fossa. Method. Unlike the ramp type BI, odontoid BI has atlas dislocation, and the odontoid is backward and oppresses the medulla. Therefore, the key to this type of treatment is to reset the atlantoaxial vertebrae and relieve the oppression of the odontoid in the medulla. There are a large number of scar tissue, even the lateral block joints and abnormal osseous fusion between the atlantoodontoid joints, resulting in the atlantoaxial dislocation of most odontoid BI in the clinic, which is difficult to restore, even under general anesthesia with large weight cranium traction. There are two main treatments for the difficult atlantoaxial dislocation. The method is the internal fixation system (C1 lateral mass screw+C2 pedicle screw+Cage), which was originally developed by the domestic scholar Yin Qingshui, such as the transoral atlantoaxial reduction plate, TARP, and the India scholar Goel. In the first 04 years, the India scholar Goel reported that the C2 ganglion and the venous plexus were severed by the posterior approach, exposing and separating the atlantoaxial joints, removing the cartilage and implanting self designed porous metal gaskets, forcing the odontoid process to move down, and finally supplemented with the atlantoaxial internal fixation. The above treatment for BI was called ''Goel technology' (Goel technique).2013 Chandra and so on. It is pointed out that the Goel technique only replaces the vertical displacement of the odontoid process and does not reposition the anterior atlantoaxial dislocation. On the basis of the Goel technique, they use the implanted Cage as the fulcrum to reset the atlas through the instruments to the universal screw. Finally, the atlantoaxial internal fixation is added to the atlantoaxial internal fixation, and the method of reduction of the atlas with the pressure of Cage as a fulcrum is called. "Modified Goel technology" (modified Goel technique). In order to reduce the iatrogenic injury of the vertebral artery during the operation, they used C2 double cortical laminar screws (bicortical C2 laminar screws, BC2LS) instead of C2 pedicle screws. The system, its biomechanical stability did not report the first reports of TARP technology such as.2004 Yin Qingshui, and then they successfully treated a large number of patients with odontoid BI with refractory atlantoaxial dislocation with TARP technology, and obtained good clinical effect,.TARP technology not only effectively relocated the atlantoaxial dislocation, but also can be used for the atlantoaxial fixation. It provides good biomechanical stability. At present, the internal fixation of TARP has been updated to third generations. The main improvement of the third generation of TAPR internal fixation is the reverse pedicle screw method and the design of the universal guide and self locking mechanism of the nail plate. Position, decompression, fixed one step, to avoid the damage to the posterior cervical muscles by posterior operation. However, TARP technique is implanted in the bone (or iliac bone) associated with the complications of iliac bone, and the possibility of bone graft collapse, absorption, displacement or abscission, resulting in bone nonunion, internal fixation inefficiency or infection, and the theory of Cage replacing granular bone with bone graft theory can theoretically increase With stability, maintaining the atlantoaxial fusion angle, and reducing the complications such as bone graft collapse, absorption, and abscission, we call the Cage combined TARP internal fixation method as "modified TARP technique, TARP+Cage". There are only a few biomechanical studies on improved TAPR technology and Goel technology at present, and no finite element is found. Finite element analysis can express cervical motion quantitatively. It is an important supplement to the biomechanical experiment in vitro. It can overcome the shortcomings of body experiment and animal experiment, high cost, and poor repeatability,.2000 years, Puttlitz and so on. It is the first report of the upper cervical vertebra finite element model and applies it to the disease of the upper cervical spine. Since then, with the development of finite element software and computer science, the finite element analysis of the upper cervical spine has been widely applied to the research of cervical vertebrae kinematics and the biomechanical study of various internal fixations. The finite element analysis of the one or two modified Goel techniques for the treatment of the stability of the skull base depression has been applied to the evaluation of C2 The biomechanical difference between the C2 pedicle screw and the C2 pedicle screw combined with Cage in the atlantoaxial fixation. Methods 1 35 year old male normal male upper cervical (C0-C2) CT data were collected and the C0-C2 segment three-dimensional finite element complete model (Intact) was established by Mimics 10.01 and Abaqus6.11 software, and the validity was verified. The normal finite element model included Cortical bone, cancellous bone, cartilage, and upper cervical associated ligaments. The models contain ligaments with transverse ligaments, upper and lower ligaments of cruciate ligaments, pterygal ligaments, apex ligaments, anterior longitudinal ligaments, atlantooccipital anterior membrane, membrane, atlantooccipital posterior membrane, atlantoaxial posterior membrane, C1-C2 joint sac, because the transverse ligament is low elastic tissue and very tough, the membrane unit is used to simulate the ligaments and the rest ligaments. The reference related articles were simulated with two node T3D2 unit. The average thickness of cortical bone was set to 1.5mm, the thickness of articular cartilage of C1-C2 was set to 3.0mm, sliding contact was used between the contact surfaces of articular cartilage, the coefficient of friction was set to 0.1. cortical bone, cancellous bone, soft bone, joint capsule, and ligament material attributes were determined according to the literature. On the established Intact model, the upper cervical instability model (Unstable) was established by deleting the transverse ligament to simulate the transverse ligament rupture (Unstable), and compared with the biomechanical experimental data of the external cervical spine. In addition, BI often merged with the atlantooccipital fusion, with a combined rate of 92%. Therefore, on the established Unstable model, the Co-C1 articular cartilage was deleted and the cell was added to simulate the atlas. On the instability model, the system model of the nail rod system (Cl lateral mass screw+Cage+bicortical C2 laminar screw, C1LS+Cage+BC2LS), and the screw rod system model of the screw +Cage+C2 pedicle screw of the backward C1 side block screw were established on the instability model. Dicle screw, C1LS+Cage+C2PS). Apply 40 N axial pressure in the occipital condyle to simulate head gravity, and apply 1.5 Nm torque above the occipital condyle to make the model forward flexion, extension, lateral bending, rotation movement, record the stress cloud and stress peak of group C1LS+Cage+BC2LS and ClLS+Cage+C2PS group, calculate C1-C2 segment activity (range of motion, R). OM). Results this study successfully established the Co-C2 nonlinear finite element model of normal people. The model simulated cortical bone, cancellous bone, articular cartilage and joint capsule, the three-dimensional structure of ligaments, total unit 26623, 26003.Intact models of nodes in the flexion, extension, side bend, Co-C1, C1-C2 ROM and Panjabi, and the experimental results of external cervical specimens. The results of the finite element model of the upper cervical vertebra, such as Zhang, are in agreement. The biomechanical study of the cervical vertebra and the finite element analysis of the upper cervical spine of the normal model, such as the validity of the normal model, and the finite element analysis of the upper cervical vertebra, such as the Zhang and so on, are all used to make the atlantoaxial instability with the method of cutting the transverse ligaments. The same study also uses the method of removing the transverse ligaments to simulate the instability of the C1-C2. The finite element result table of this model is also used in this study. Compared with the Intact model, the Unstable model increased by 35.2%, 16.4%, 4%, 5.6%, respectively, under the flexion, extension, lateral bending and rotation load, and the results were basically consistent with the biomechanical results of the cervical vertebra, such as Li and so on. The C1-C2 segment ROM of C1LS+Cage+BC2LS and C1LS+Cage+C2PS groups were less than 0.1 degrees under any load and two, and two There was no significant difference in stress distribution and stress peak value of all screws fixed in the group. The two groups of internal fixation systems were under the flexion of the internal fixation system and the C1 screw under the extension load was more stressed than the C2 screw, especially the maximum stress of the C1 screw under the extension load was about 2 times of the C2 screw. Under the extension load, the stress of the two internal fixation in the Cage was the smallest. There was a bright future. Explicit stress occlusion, especially in group C1LS+Cage+C2PS. Conclusion a finite element model of the upper cervical spine was established in this study, and two different internal fixation systems were used to carry out the finite element analysis of the physiological load in different motion states. The results showed that the C2 double cortex laminectomy could be used when the axis was not fit for the C2 pedicle screw for the treatment of the BI. Compared with C2PS technology, BC2LS technology is simple, easy and effective to avoid the injury of vertebral artery and spinal cord compared with C2PS technology. It is still necessary to further study the imaging data of the cranial depression patients, and to treat the skull base depression by using C2 double cortical laminar screw technique in the bed. Provide a theoretical basis for the disease. Experiment two the finite element analysis of the modified TARP technique and Goel technique in the treatment of the stability of the skull base depression. The finite element analysis was used to compare the differences of biomechanical stability between the anterior improved TARP technique and the posterior Goel technique in the treatment of the skull base depression. Methods 1 35 year old male normal men's upper cervical (C0-C2) CT data were collected. Mimics 10.01 and Abaqus6.11 software were used to establish a complete three-dimensional finite element model of Co-C2 segment and verify the validity of the model. In the instability model, a modified TARP internal fixation model (transoral atlantoaxial reduction plate+Cage, TARP+Cage) was established by using the internal fixation method of TARP internal fixation (transoral atlantoaxial reduction plate+Cage, TARP+Cage). The concrete methods are as follows: C1 reverse side block screws use the single skin. The screw point is located at the outside 5mm of the medial edge of the Cl side block, and the outward side inclines 5 degree -10 degrees to 10 [degree] -15. to ensure the entry on the C1 side block axis; C2 reverse pedicle screw adopts the double cortical screw, and the point of entry is located about 5mm below the medial apex of the joint surface on the C2, and is tilted to the outside 9.3 degrees -28.3 degrees to the outside, and down to 6.5 [degree] -2.15 degrees downward, reverse along C2. The pedicle axis enters. Similarly, the posterior Goel internal fixation model (C1 lateral mass screw+C2 pedicle screw+Cage, C1LS+C2PS+Cage) is constructed by reference to the posterior C1 lateral mass screw and C2 pedicle screw implantation on the instability model. The concrete method is as follows: the C1 lateral block screw adopts the double cortical screw and the insertion point is located at the middle of the C1 side block below the posterior bow. The -2 joint was inclined to the medial tilt 5.-10., upward inclined 10.-15 degree, along the C1 side block axis; C2 pedicle screw adopted the double cortical screw, followed the "high, internal" principle, the needle point was located on the inside of the C2 pedicle, inclined 16.5 degree -23.8 to the inside of the C1-2 joint, inclined upward 25.3 degrees, along the C2 vertebral arch axis. 40 N axial pressure was used to simulate the head gravity, and the 1.5 Nm torque was applied in the occipital condyle to make the model forward flexion, extension, lateral bending and rotation movement, record the stress cloud map and the peak stress of group TARP+Cage and C1LS+C2PS+Cage, and calculate the C1-C2 segment activity (range of motion, ROM). Compared with the complete model, the results of the two groups were all fixed. To reduce the C1-C2 segment ROM. and the C1LS+C2PS+Cage group, the C1-C2 segment ROM of the TARP+Cage group decreased by 44.7%, 30%, and 10.5% respectively under the load of back, side bending and rotation, but the stress peak of the TARP+Cage group C2 screw under the flexion load was greater than that of the ClLS+C2PS+Cage group C2 screws under the extension load, and the rest was in the forward, back, and side bending. The peak values of stress of C1 screws and C2 screws in group TARP+Cage were less than those in group C1LS+C2PS+Cage under C1 and C2 screws.

【学位授予单位】：南方医科大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R687.3

【参考文献】