基于MRI与超声图像融合的前列腺癌粒子植入过程实时路径控制

Image-guided robotic prostate cancer seed implantation is the inevitable trend in the future clinical treatment. But it is difficult to be located lesions precisely and dynamically. And the traditional intervention rigid needle can not bypass the sensitive position, such as phalanx and nerve, and can not realize interventional treatment for the arbitrary region of prostate. The above-mentioned problems seriously restrict the development of image-guided robotic prostate cancer seed implantation application. This project intends to use and fuse both of the advantages of the high precision of MRI image and fast convenience of Ultrasound image, thus to realize real-time dynamic positioning of prostate lesions. Feature information of prostate MRI and ultrasound image is extracted by contour feature extraction algorithm based on mathematical morphology. Automatic segmentation of lesions and flexible needle is carried out by multi-Gaussian mixture model. Image fusion method of preoperative MRI image and intra-operative ultrasound image of lesions is studied by unrestrained mark registration method with adding fuzzy rule. 3-D tree type seed implantation strategy is put forward based on the target-insertion principle of flexible needle. Multiple-target rapid-exploring random tree algorithm is implemented to realize real-time 3-D path planning and control of seed implantation by flexible needle. The robot experimental system of flexible needle prostate seed implantation based on MRI-Ultrasound image fusion is established, and the experimental study is executed. This project will theoretically and practically promote the efficiency and quality of prostate brachytherapy.

图像引导下的机器人前列腺癌粒子植入治疗是未来临床的必然趋势。但是前列腺及其病灶点在术中难以动态实时精确定位，传统的刚性介入针无法绕过趾骨、神经等敏感部位以实现对前列腺任意区域的介入治疗，这些问题严重地制约了它的发展。.提出利用MRI图像的高精度和超声图像的快速便捷性，融合二者的优点来进行前列腺癌病灶点的实时动态定位。基于数学形态学的轮廓特征算法提取前列腺图像的特征信息，建立多高斯混合模型实现病灶点和柔性针的自动分割。研究扰动情况下病灶点术前MRI图像和术中超声图像的融合方法。提出基于柔性针的三维树状粒子植入策略，建立多目标快速扩展随机树算法模型，实现柔性针粒子植入实时三维路径规划和控制。研究适应柔性针驱动和控制的前列腺介入机器人的结构和组成，构建MRI-超声图像融合的柔性针前列腺癌粒子植入机器人实验平台，并开展实验研究。本项目对提高前列腺介入治疗的效率和质量具有重要理论意义和价值。

图像引导下的机器人前列腺癌粒子植入治疗是未来临床的必然趋势。本项目提出提出利用MRI图像的高精度和超声图像的快速便捷性，融合了二者的优点来进行前列腺癌病灶点的实时动态定位。该项目基于数学形态学的轮廓特征算法提取前列腺图像的特征信息，建立了多高斯混合模型实现病灶点和柔性针的自动分割。该项目研究了扰动情况下病灶点术前MRI图像和术中超声图像的融合方法。该项目提出了基于柔性针的三维树状粒子植入策略，建立了多目标快速扩展随机树算法模型，实现了柔性针粒子植入实时三维路径规划和控制。该项目完成了研究适应柔性针驱动和控制的前列腺介入机器人的结构和组成，构建MRI-超声图像融合的柔性针前列腺癌粒子植入机器人实验平台，并开展实验研究。本项目对提高前列腺介入治疗的效率和质量具有重要理论意义和价值。