一种基于2D/3D配准的脊柱术中校正方法

摘要

背景:脊柱术前三维影像有助于诊断和治疗,术中患者体位变化将引起脊柱形态改变,致使术前影像不能反映术中实际情况,无法确保手术的顺利实施.目的:利用脊髓手术中影像校正术前脊柱模型形态.方法:实验提出了一种基于2D/3D配准的脊柱术中校正方法,借助数字影像重建技术完成术前X射线图像与CT体数据的2D/3D配准,进一步完成术中、术前X射线图像中独立椎段的特征匹配,利用上述配准结果实现术前脊柱CT模型的术中快速校正.结果与结论:采用附有标记的颈椎标本进行实验,校正后可基本消除术前脊柱模型与术中形态的偏差,其误差可控制在1mm以内,能够满足医学临床要求.%BACKGROUND: The preoperative three-dimensional (3D) diagnosis and treatment. As patient's posture changes, the intra-operative shape of spine is deferent form the preoperative. So the preoperative images can not reflect the intra-operative situation. And doctor can not ensure the smooth implementation of the operation.OBJECTIVE: To correct preoperative spine models using intra-operative image.METHODS: This paper presented an approach based on 2D/3D registration to correct preoperative images during operation. We use digitally reconstructed radiograph (DRR) to register CT to X-ray image. Then complete the feature registrations of all independent vertebra segments between intra-operative and preoperative X-ray images. The results of above registrations are exploited to fast correct the preoperative spine models.RESULTS AND CONCLUSION: Using marked cervical spine specimens for experiment, the deviation of the spine model can be basically eliminated after correction. The error between intra-operative and corrected preoperative spine model is less than 1mm.And clinical requirements can be met.