In robotic radiosurgery, it is necessary to compensate for systematic latencies arising from target tracking and mechanical constraints. This compensation is usually achieved by means of an algorithm which computes the future target position. In most scientific works on respiratory motion prediction, only one or two algorithms are evaluated on a limited amount of very short motion traces. The purpose of this work is to gain more insight into the real world capabilities of respiratory motion prediction methods by evaluating many algorithms on an unprecedented amount of data. We have evaluated six algorithms, the normalized least mean squares (nLMS), recursive least squares (RLS), multi-step linear methods (MULIN), wavelet-based multiscale autoregression (wLMS), extended Kalman filtering, and ε-support vector regression (SVRpred) methods, on an extensive database of 304 respiratory motion traces. The traces were collected during treatment with the CyberKnife (Accuray, Inc., Sunnyvale, CA, USA) and feature an average length of 71 min. Evaluation was done using a graphical prediction toolkit, which is available to the general public, as is the data we used. The experiments show that the nLMS algorithm-which is one of the algorithms currently used in the CyberKnife-is outperformed by all other methods. This is especially true in the case of the wLMS, the SVRpred, and the MULIN algorithms, which perform much better. The nLMS algorithm produces a relative root mean square (RMS) error of 75% or less (i.e., a reduction in error of 25% or more when compared to not doing prediction) in only 38% of the test cases, whereas the MULIN and SVRpred methods reach this level in more than 77%, the wLMS algorithm in more than 84% of the test cases. Our work shows that the wLMS algorithm is the most accurate algorithm and does not require parameter tuning, making it an ideal candidate for clinical implementation. Additionally, we have seen that the structure of a patient's respiratory motion trace has strong influence on the outcome of prediction. Further work is needed to determine a priori the suitability of an individual's respiratory behaviour to motion prediction.
The new SVR-based correlation method outperforms traditional polynomial correlation methods for motion tracking. This method is suitable for clinical implementation and may improve the overall accuracy of targeted radiotherapy.
In robotic radiosurgery, a photon beam source, moved by a robot arm, is used to ablate tumors. The accuracy of the treatment can be improved by predicting respiratory motion to compensate for system delay. We consider a wavelet-based multiscale autoregressive prediction method. The algorithm is extended by introducing a new exponential averaging parameter and the use of the Moore-Penrose pseudo inverse to cope with long-term signal dependencies and system matrix irregularity, respectively. In test cases, this new algorithm outperforms normalized LMS predictors by as much as 50%. With real patient data, we achieve an improvement of around 5 to 10%.
A major challenge during endovascular interventions is visualising the position and orientation of the catheter being inserted. This is typically achieved by intermittent X-ray imaging. Since the radiation exposure to the surgeon is considerable, it is desirable to reduce X-ray exposure to the bare minimum needed. Additionally, transferring two-dimensional (2D) X-ray images to 3D locations is challenging. The authors present the development of a real-time navigation framework, which allows a 3D holographic view of the vascular system without any need of radiation. They extract the patient's surface and vascular tree from pre-operative computed tomography data and register it to the patient using a magnetic tracking system. The system was evaluated on an anthropomorphic full-body phantom by experienced clinicians using a four-point questionnaire. The average score of the system (maximum of 20) was found to be 17.5. The authors’ approach shows great potential to improve the workflow for endovascular procedures, by simultaneously reducing X-ray exposure. It will also improve the learning curve and help novices to more quickly master the required skills.
This new algorithm is a feasible tool for the prediction of human respiratory motion signals significantly outperforming previous algorithms. The only drawback is the high computational complexity and the resulting slow prediction speed. High performance computers will be needed to use the algorithm in live prediction of signals sampled at a high resolution.
IntroductionEndovascular aortic repair (EVAR) is a minimal-invasive technique that prevents life-threatening rupture in patients with aortic pathologies by implantation of an endoluminal stent graft. During the endovascular procedure, device navigation is currently performed by fluoroscopy in combination with digital subtraction angiography. This study presents the current iterative process of biomedical engineering within the disruptive interdisciplinary project Nav EVAR, which includes advanced navigation, image techniques and augmented reality with the aim of reducing side effects (namely radiation exposure and contrast agent administration) and optimising visualisation during EVAR procedures. This article describes the current prototype developed in this project and the experiments conducted to evaluate it.MethodsThe current approach of the Nav EVAR project is guiding EVAR interventions in real-time with an electromagnetic tracking system after attaching a sensor on the catheter tip and displaying this information on Microsoft HoloLens glasses. This augmented reality technology enables the visualisation of virtual objects superimposed on the real environment. These virtual objects include three-dimensional (3D) objects (namely 3D models of the skin and vascular structures) and two-dimensional (2D) objects [namely orthogonal views of computed tomography (CT) angiograms, 2D images of 3D vascular models, and 2D images of a new virtual angioscopy whose appearance of the vessel wall follows that shown in ex vivo and in vivo angioscopies]. Specific external markers were designed to be used as landmarks in the registration process to map the tracking data and radiological data into a common space. In addition, the use of real-time 3D ultrasound (US) is also under evaluation in the Nav EVAR project for guiding endovascular tools and updating navigation with intraoperative imaging. US volumes are streamed from the US system to HoloLens and visualised at a certain distance from the probe by tracking augmented reality markers. A human model torso that includes a 3D printed patient-specific aortic model was built to provide a realistic test environment for evaluation of technical components in the Nav EVAR project. The solutions presented in this study were tested by using an US training model and the aortic-aneurysm phantom.ResultsDuring the navigation of the catheter tip in the US training model, the 3D models of the phantom surface and vessels were visualised on HoloLens. In addition, a virtual angioscopy was also built from a CT scan of the aortic-aneurysm phantom. The external markers designed for this study were visible in the CT scan and the electromagnetically tracked pointer fitted in each marker hole. US volumes of the US training model were sent from the US system to HoloLens in order to display them, showing a latency of 259±86 ms (mean±standard deviation).ConclusionThe Nav EVAR project tackles the problem of radiation exposure and contrast agent administration during EVAR interventions by using a multidisciplinary approach to guide the endovascular tools. Its current state presents several limitations such as the rigid alignment between preoperative data and the simulated patient. Nevertheless, the techniques shown in this study in combination with fibre Bragg gratings and optical coherence tomography are a promising approach to overcome the problems of EVAR interventions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
334 Leonard St
Brooklyn, NY 11211
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.