基于CT图像的三维个体化脊柱侧凸矫正手术辅助系统关键技术研究

Scoliosis is a high incidence of orthopedic diseases, the low success rate in traditional corrective operation for scoliosis always dues to the shortage of accurate pre-surgical guidance. Based on regular CT image data, this project studies on resolving key issues of three-dimensional individual correction system for scoliosis based on CT image data and provides precise guidance for scoliosis surgery. The following issues are offered in this project:(1) Research on polymorphic spinal 3D deep feature information extraction methods, and their applications to the automatic location and segmentation of spinal sections; (2) Research on accurate spinal morphology-oriented descriptions, and their applications to the establishment of new 3D scoliosis measurement parameters’ system; (3) Research on spatial curve modeling and clinical data mining technique, and their applications to the accurate modeling of 3D spinal resilience curve. This project attempts to reach break-throughs in both theory and clinical practice of 3D individual surgical navigation system, and finally builds an intelligent, efficient and precise 3D individual surgical assistant system for scoliosis correction.

脊柱侧凸是骨科高发性疾病，传统脊柱侧凸矫正手术由于缺乏术前精准指导造成手术成功率不高。本项目基于常规CT图像数据，研究解决三维个体化脊柱侧凸矫正系统构建中的关键技术问题,以求能为脊柱侧凸手术提供更为精准的指导。本项目提出: (1) 多形态脊柱深度三维特征信息提取方法研究，及其在自动脊柱节段定位与分割中的应用; (2) 面向三维脊柱形态特性的精准表达方法研究，及其在脊柱侧凸参数体系建立中的应用; (3) 空间曲线建模和临床数据挖掘方法研究，及其在精准三维脊柱恢复力线建模中的应用。本课题尝试在三维个体化手术导航系统的理论方法和临床实践两方面取得突破，最终建立一套智能、高效、精准的三维个体化脊柱侧凸矫正手术辅助系统。

本项目完成了多形态脊柱深度三维特征信息提取方法、面向三维脊柱形态特性的精准表达方法、空间曲线建模和临床数据挖掘方法三个方面的研究，并应用于自动脊柱节段定位与分割、脊柱侧凸参数体系建立以及精准三维脊柱恢复力线建模。解决了常规CT图像数据条件下三维个体化脊柱侧凸矫正系统构建中的关键技术问题，项目解决了测量面的建模并建立空间参数计算模型；在满足临床实际要求的条件下设定新参数，建立了相关测量方法；完成了三维测量参数医学性态和精度的稳定性评价；将医生的经验转化为脊柱几何形态数据表达指导恢复力线生成。项目最终建立了一套智能、高效、精准的三维个体化脊柱侧凸矫正手术辅助系统，在一定程度上提供了较为精确的术前指导，进而提高脊柱侧凸矫正手术的成功率。