Hanglok’s Article at the IEEE Transactions on Medical Imaging (TMI)
2024-03-14 14:39

We are excited to share with you our article published in the IEEE Transactions on Medical Imaging:“3D/2D Vessel Registration based on Monte Carlo Tree Search and Manifold Regularization”. DOI: 10.1109/TMI.2023.3347896

(https://ieeexplore.ieee.org/document/10375552). The following is an introduction to the article:

The augmented intra-operative real-time imaging in vascular interventional surgery, which is generally performed by projecting preoperative computed tomography angiography images onto intraoperative digital subtraction angiography (DSA) images, can compensate for the deficiencies of DSA-based navigation, such as lack of depth information and excessive use of toxic contrast agents. 3D/2D vessel registration is the critical step in image augmentation. A 3D/2D registration method based on vessel graph matching is proposed in this study. For rigid registration, the matching of vessel graphs can be decomposed into continuous states, thus 3D/2D vascular registration is formulated as a search tree problem. The Monte Carlo tree search method is applied to find the optimal vessel matching associated with the highest rigid registration score. For nonrigid registration, we propose a novel vessel deformation model based on manifold regularization. This model incorporates the smoothness constraint of vessel topology into the objective function. Furthermore, we derive simplified gradient formulas that enable fast registration. The proposed technique undergoes evaluation against seven rigid and three nonrigid methods using a variety of data - simulated, algorithmically generated, and manually annotated - across three vascular anatomies: the hepatic artery, coronary artery, and aorta. Our findings show the proposed method’s resistance to pose variations, noise, and deformations, outperforming existing methods in terms of registration accuracy and computational efficiency. The proposed method demonstrates average registration errors of 2.14 mm and 0.34 mm for rigid and nonrigid registration, and an average computation time of 0.51 s.

Work flow 4 consecutive steps

In the method part, the paper firstly defines the vascular topological model, the vascular registration problem and the solution space of the vascular registration, and gives a detailed mathematical derivation of the solution of rigid registration and elastic registration. Detailed calculation process, pleasereferto the original paper (http://doi.org/10.1109/TMI.2023.3347896).

In the experimental part, the author conducted a comprehensive data test on the proposed algorithm, including the robustness test of the algorithm using simulated vascular data with artificial defects (noise, deformation), the robustness test of the algorithm using the vascular centerline calculated by the segmentation algorithm, and the registration test using manually labeled clinical hepatic artery, aorta and coronary artery data. Quantitative and qualitative experiments demonstrate the excellent performance of the proposed method.

(Algorithm in artificial defect simulation of vascular data registration experiment results)

(Algorithm in automatic segmentation of LCA and RCA centerline data registration experimental results demonstrated)

(The algorithm is demonstrated in the registration results of manually labeled hepatic arteries, coronary arteries and aorta)

This method is a key technology to realize spatial positioning and navigation of vascular interventional surgical instruments, which can map the guide wire catheter and other instruments of real-time DSA images into the preoperative 3D vascular space. Hanglokcooperated with Zhongda Hospital Affiliated to Southeast University(东南大学附属中大医院)and the School of Instrument Science and Engineering of Southeast University(东南大学仪器科学与工程学院)to develop a new generation of vascular interventional robot system platform, including the device loading and control of multiple interventional processes and the immersive intelligent teleoperation master platform. At the same time, the research on key technologies such as force feedback force sensing of vascular interventional robot, medical image artificial intelligence large model, robot intelligent control, and surgical robot underlying operating system was carried out. The project was funded by China Postdoctoral Science Foundation (2021M700772).

Experimental scene outside the operating room