Towards real-time MRI-guided 3D localization of deforming targets for non-invasive cardiac radiosurgery.

Radiosurgery to the pulmonary vein antrum in the left atrium (LA) has recently been proposed for non-invasive treatment of atrial fibrillation (AF). Precise real-time target localization during treatment is necessary due to complex respiratory and cardiac motion and high radiation doses. To determine the 3D position of the LA for motion compensation during radiosurgery, a tracking method based on orthogonal real-time MRI planes was developed for AF treatments with an MRI-guided radiotherapy system. Four healthy volunteers underwent cardiac MRI of the LA. Contractile motion was quantified on 3D LA models derived from 4D scans with 10 phases acquired in end-exhalation. Three localization strategies were developed and tested retrospectively on 2D real-time scans (sagittal, temporal resolution 100 ms, free breathing). The best-performing method was then used to measure 3D target positions in 2D-2D orthogonal planes (sagittal-coronal, temporal resolution 200-252 ms, free breathing) in 20 configurations of a digital phantom and in the volunteer data. The 3D target localization accuracy was quantified in the phantom and qualitatively assessed in the real data. Mean cardiac contraction was ⩽ 3.9 mm between maximum dilation and contraction but anisotropic. A template matching approach with two distinct template phases and ECG-based selection yielded the highest 2D accuracy of 1.2 mm. 3D target localization showed a mean error of 3.2 mm in the customized digital phantoms. Our algorithms were successfully applied to the 2D-2D volunteer data in which we measured a mean 3D LA motion extent of 16.5 mm (SI), 5.8 mm (AP) and 3.1 mm (LR). Real-time target localization on orthogonal MRI planes was successfully implemented for highly deformable targets treated in cardiac radiosurgery. The developed method measures target shifts caused by respiration and cardiac contraction. If the detected motion can be compensated accordingly, an MRI-guided radiotherapy system could potentially enable completely non-invasive treatment of AF.

Transcranial magnetic stimulation reveals complex cognitive control representations in the rostral frontal cortex.

Convergent evidence suggests that the lateral frontal cortex is at the heart of a brain network subserving cognitive control. Recent theories assume a functional segregation along the rostro-caudal axis of the lateral frontal cortex based on differences in the degree of complexity of cognitive control. However, the functional contribution of specific rostral and caudal sub-regions remains elusive. Here we investigate the impact of disrupting rostral and caudal target regions on cognitive control processes, using Transcranial Magnetic Stimulation (TMS). Participants performed three different task-switching conditions that assessed differences in the degree of complexity of cognitive control processes, after temporally disrupting rostral, or caudal target regions, or a control region. Disrupting the rostral lateral frontal region specifically impaired behavioral performance of the most complex task-switching condition, in comparison to the caudal target region and the control region. These novel findings shed light on the neuroanatomical architecture supporting control over goal-directed behavior.

Radiotherapy beyond cancer: target localization in real-time MRI and treatment planning for cardiac radiosurgery.

PURPOSE: Atrial fibrillation (AFib) is the most common cardiac arrhythmia that affects millions of patients world-wide. AFib is usually treated with minimally invasive, time consuming catheter ablation techniques. While recently noninvasive radiosurgery to the pulmonary vein antrum (PVA) in the left atrium has been proposed for AFib treatment, precise target location during treatment is challenging due to complex respiratory and cardiac motion. A MRI linear accelerator (MRI-Linac) could solve the problems of motion tracking and compensation using real-time image guidance. In this study, the authors quantified target motion ranges on cardiac magnetic resonance imaging (MRI) and analyzed the dosimetric benefits of margin reduction assuming real-time motion compensation was applied. METHODS: For the imaging study, six human subjects underwent real-time cardiac MRI under free breathing. The target motion was analyzed retrospectively using a template matching algorithm. The planning study was conducted on a CT of an AFib patient with a centrally located esophagus undergoing catheter ablation, representing an ideal case for cardiac radiosurgery. The target definition was similar to the ablation lesions at the PVA created during catheter treatment. Safety margins of 0 mm (perfect tracking) to 8 mm (untracked respiratory motion) were added to the target, defining the planning target volume (PTV). For each margin, a 30 Gy single fraction IMRT plan was generated. Additionally, the influence of 1 and 3 T magnetic fields on the treatment beam delivery was simulated using Monte Carlo calculations to determine the dosimetric impact of MRI guidance for two different Linac positions. RESULTS: Real-time cardiac MRI showed mean respiratory target motion of 10.2 mm (superior-inferior), 2.4 mm (anterior-posterior), and 2 mm (left-right). The planning study showed that increasing safety margins to encompass untracked respiratory motion leads to overlapping structures even in the ideal scenario, compromising either normal tissue dose constraints or PTV coverage. The magnetic field caused a slight increase in the PTV dose with the in-line MRI-Linac configuration. CONCLUSIONS: The authors' results indicate that real-time tracking and motion compensation are mandatory for cardiac radiosurgery and MRI-guidance is feasible, opening the possibility of treating cardiac arrhythmia patients completely noninvasively.

Multivariate respiratory motion prediction.

In extracranial robotic radiotherapy, tumour motion is compensated by tracking external and internal surrogates. To compensate system specific time delays, time series prediction of the external optical surrogates is used. We investigate whether the prediction accuracy can be increased by expanding the current clinical setup by an accelerometer, a strain belt and a flow sensor. Four previously published prediction algorithms are adapted to multivariate inputs-normalized least mean squares (nLMS), wavelet-based least mean squares (wLMS), support vector regression (SVR) and relevance vector machines (RVM)-and evaluated for three different prediction horizons. The measurement involves 18 subjects and consists of two phases, focusing on long term trends (M1) and breathing artefacts (M2). To select the most relevant and least redundant sensors, a sequential forward selection (SFS) method is proposed. Using a multivariate setting, the results show that the clinically used nLMS algorithm is susceptible to large outliers. In the case of irregular breathing (M2), the mean root mean square error (RMSE) of a univariate nLMS algorithm is 0.66 mm and can be decreased to 0.46 mm by a multivariate RVM model (best algorithm on average). To investigate the full potential of this approach, the optimal sensor combination was also estimated on the complete test set. The results indicate that a further decrease in RMSE is possible for RVM (to 0.42 mm). This motivates further research about sensor selection methods. Besides the optical surrogates, the sensors most frequently selected by the algorithms are the accelerometer and the strain belt. These sensors could be easily integrated in the current clinical setup and would allow a more precise motion compensation.

Tracking post-error adaptation in the motor system by transcranial magnetic stimulation.

The commission of an error triggers cognitive control processes dedicated to error correction and prevention. Post-error adjustments leading to response slowing following an error ("post-error slowing"; PES) might be driven by changes in excitability of the motor regions and the corticospinal tract (CST). The time-course of such excitability modulations of the CST leading to PES is largely unknown. To track these presumed excitability changes after an error, single pulse transcranial magnetic stimulation (TMS) was applied to the motor cortex ipsilateral to the responding hand, while participants were performing an Eriksen flanker task. A robotic arm with a movement compensation system was used to maintain the TMS coil in the correct position during the experiment. Magnetic pulses were delivered over the primary motor cortex ipsilateral to the active hand at different intervals (150, 300, 450 ms) after correct and erroneous responses, and the motor-evoked potentials (MEP) of the first dorsal interosseous muscle (FDI) contralateral to the stimulated hemisphere were recorded. MEP amplitude was increased 450 ms after the error. Two additional experiments showed that this increase was neither associated to the correction of the erroneous responses nor to the characteristics of the motor command. To the extent to which the excitability of the motor cortex ipsi- and contralateral to the response hand are inversely related, these results suggest a decrease in the excitability of the active motor cortex after an erroneous response. This modulation of the activity of the CST serves to prevent further premature and erroneous responses. At a more general level, the study shows the power of the TMS technique for the exploration of the temporal evolution of post-error adjustments within the motor system.

Evaluating and comparing algorithms for respiratory motion prediction.

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.

Automatic scanning of large tissue areas in neurosurgery using optical coherence tomography.

BACKGROUND: With its high spatial and temporal resolution, optical coherence tomography (OCT) is an ideal modality for intra-operative imaging. One possible application is to detect tumour invaded tissue in neurosurgery, e.g. during complete resection of glioblastoma. Ideally, the whole resection cavity is scanned. However, OCT is limited to a small field of view (FOV) and scanning perpendicular to the tissue surface. METHODS: We present a new method to use OCT for scanning of the resection cavity during neurosurgical resection of brain tumours. The main challenges are creating a map of the cavity, scanning perpendicular to the surface and merging the three-dimensional (3D) data for intra-operative visualization and detection of residual tumour cells. RESULTS: Our results indicate that the proposed method enables creating high-resolution maps of the resection cavity. An overlay of these maps with the microscope images provides the surgeon with important information on the location of residual tumour tissue underneath the surface. CONCLUSION: We demonstrated that it is possible to automatically acquire an OCT image of the complete resection cavity. Overlaying microscopy images with depth information from OCT could lead to improved detection of residual tumour cells.

Motorization of a surgical microscope for intra-operative navigation and intuitive control.

BACKGROUND: During surgical procedures, various medical systems, e.g. microscope or C-arm, are used. Their precise and repeatable manual positioning can be very cumbersome and interrupts the surgeon's work flow. Robotized systems can assist the surgeon but they require suitable kinematics and control. However, positioning must be fast, flexible and intuitive. METHODS: We describe a fully motorized surgical microscope. Hardware components as well as implemented applications are specified. The kinematic equations are described and a novel control concept is proposed. RESULTS: Our microscope combines fast manual handling with accurate, automatic positioning. Intuitive control is provided by a small remote control mounted to one of the surgical instruments. Positioning accuracy and repeatability are < 1 mm and vibrations caused by automatic movements fade away in about 1 s. CONCLUSION: The robotic system assists the surgeon, so that he can position the microscope precisely and repeatedly without interrupting the clinical workflow. The combination of manual und automatic control guarantees fast and flexible positioning during surgical procedures.

Vascular tree reconstruction with discrete tomography: intensity based camera correction for 3D reconstruction.

PURPOSE: This paper is concerned with the reconstruction of vascular trees from few projections using discrete tomography. However, its computational cost is high and it lacks robustness when the data are inconsistent. We improve robustness by incorporating an intensity-based camera-correction method. The proposed approach is also capable of handling small motion artifacts by modeling them as repositionings of a virtual X-ray camera. We also present a parallel implementation which substantially reduces reconstruction time. METHODS: We propose a data-driven reduction of positional inconsistencies by minimizing the reconstruction residual to increase the robustness. Inspired by motion compen-sation algorithms in SPECT imaging, we combine an intensity-based 2D/3D-registration method with itera-tive reconstruction methods. Our objective is the robust vascular-tree reconstruction from positionally inconsistent data. The speed of the reconstruction is substantially increased by a volume-splitting scheme that allows parallel processing. RESULTS: Vascular trees in the liver can be accurately reconstructed from few positionally inconsistent projections using digitally reconstructed radiographs. We have tested the proposed method on synthetic projection data and on objects imaged with a new robotized C-arm. We measured a decrease in the average reconstruction residual of about 13% for real data compared to projection data without preprocessing. Over 4,600 reconstruction experiments were conducted to evaluate the speed-up obtained when employing the volume-splitting scheme. Reconstruction time decreased linearly with increased number of processor-cores, both for real and synthetic data. CONCLUSIONS: The proposed method reduces inconsistencies caused by positioning errors and small motion artifacts. No prior segmentation or detection of correspondences between projections is necessary, because all algorithms are intensity-based. As a result, the proposed method allows for robust, high-quality reconstructions, while reducing radiation dose substantially.

Achieving super-resolution X-ray imaging with mobile C-arm devices.

BACKGROUND: The term super-resolution refers to the process of combining a set of low-resolution images into a high-resolution image using image processing methods. This work is concerned with the reconstruction of high-resolution X-ray images. Specifically, we address the problem of acquiring X-ray images from multiple, very close view points. METHODS: We propose to use a novel experimental robotic C-arm device to create high-resolution X-ray images. For this purpose, we suggest different strategies for acquiring multiple low-resolution images, and we provide the steps to achieve acquisition-error compensation. Compared to visible light images, X-ray images have the particularity that parallax effects render super-resolution very difficult. Using the acquired multi-frame data, we evaluate recent well-known super-resolution reconstruction algorithms. The same algorithms are evaluated based on synthetic 3D phantom data and real X-ray images. RESULTS: In experiments with both synthetic and real projection data, we successfully reconstruct up to four times higher-resolution images. These images reveal structures and details which are not perceivable in the low-resolution images. CONCLUSIONS: The advantage of super-resolution techniques for X-ray is the potential reduction of radiation dose for patients and medical personnel. Potential medical applications include the diagnosis of early-stage osteoporosis and the detection of very small calcifications.

Towards multi-directional OCT for speckle noise reduction.

Multi-directional optical coherence tomography (MD-OCT) applies and extends the concept of angular compounding for speckle noise reduction to the area of OCT imaging. OCT images are acquired from a wide range of angles of view. Averaging of the rotated images therefore requires compensation of the parallax which is achieved by simple image registration for image reconstruction. Test measurements of a sample structure in a low and highly scattering environment show that the method improves the signal-to-noise ratio by a factor of 4 and hence reduces speckle noise significantly. Experimental results also show that the proposed averaging increases the performance of common edge-detection algorithms.

A robotic assistant for stereotactic neurosurgery on small animals.

BACKGROUND: This work presents the development and performance analysis of a robotic system for stereotactic neurosurgery on small animals. The system is dedicated to the precise placement of probes in the small animal brain, thus providing an improved framework for brain research. METHODS: Based on an analysis of small animal stereotaxy, the mechanical design of the robotic system is presented. Details of the structure and mechanical components and a kinematic description are outlined. The calibration process of the system for arbitrary probes is described. To analyse the mechanical positioning accuracy of the system, a testbed is presented. RESULTS: Positioning performance results show that the system features a mean mechanical positioning accuracy of 32 microm and a mean positioning repeatability of 11 microm. CONCLUSION: The system meets the requirements of targeting small functional areas within the brain of small animals and thus offers a new tool for small animal brain research.

Brain mapping with transcranial magnetic stimulation using a refined correlation ratio and Kendall's tau.

Transcranial magnetic stimulation provides a mean to stimulate the brain non-invasively and painlessly. The effect of the stimulation hereby depends on the stimulation coil used and on its placement. This paper presents a mapping algorithm based on the assumption of a monotonous functional relationship between the applied electric field strength at the representation point of a muscle and the evoked motor potential. We combine data from coil characteristics, coil placement, and stimulation outcome to calculate a likelihood map for the representation of stimulated muscles in the brain. Hereby, correlation ratio (CR) and Kendall's rank coefficient tau are used to find areas in the brain where there is most likely a functional or monotonous relationship between electric field strength applied to this area and the muscle response. First results show a good accordance of our method with mapping from functional magnetic resonance imaging. In our case, classical evaluation of CR with binning is impossible, because sample data sets are too small and data are continuous. We therefore introduce a refined CR formula based on a Parzen windowing of the X-data to solve the problem. In contrast to usual windowing approaches, which require numeric integration, it can be evaluated directly in O(n2) time. Hence, its advantage lies in fast evaluation while maintaining robust applicability to small sample sets. We suggest that the presented formula can generally be used in CR-related problems where sample size is small and data range is continuous.

Stepwise multi-criteria optimization for robotic radiosurgery.

Achieving good conformality and a steep dose gradient around the target volume remains a key aspect of radiosurgery. Clearly, this involves a trade-off between target coverage, conformality of the dose distribution, and sparing of critical structures. Yet, image guidance and robotic beam placement have extended highly conformal dose delivery to extracranial and moving targets. Therefore, the multi-criteria nature of the optimization problem becomes even more apparent, as multiple conflicting clinical goals need to be considered coordinate to obtain an optimal treatment plan. Typically, planning for robotic radiosurgery is based on constrained optimization, namely linear programming. An extension of that approach is presented, such that each of the clinical goals can be addressed separately and in any sequential order. For a set of common clinical goals the mapping to a mathematical objective and a corresponding constraint is defined. The trade-off among the clinical goals is explored by modifying the constraints and optimizing a simple objective, while retaining feasibility of the solution. Moreover, it becomes immediately obvious whether a desired goal can be achieved and where a trade-off is possible. No importance factors or predefined prioritizations of clinical goals are necessary. The presented framework forms the basis for interactive and automated planning procedures. It is demonstrated for a sample case that the linear programming formulation is suitable to search for a clinically optimal treatment, and that the optimization steps can be performed quickly to establish that a Pareto-efficient solution has been found. Furthermore, it is demonstrated how the stepwise approach is preferable compared to modifying importance factors.

Brain-mapping using robotized TMS.

We present first results of brain-mapping using robotic Transcranial Magnetic Stimulation. This non-invasive procedure enables the reliable detection of the representation of individual muscles or muscle groups in the motor-cortex. The accuracy is only exceeded by direct electrical stimulation of the brain during surgery. Brain-mapping using robotic TMS can also be used to detect displacements of brain regions caused by tumors. The advantage of TMS is that it is non-invasive. In this study, we compare results from statistical mapping with robotic TMS to results achieved from direct stimulation done during tumor surgery. To our knowledge this is the first study of this type. We mapped the representation of three muscle groups (forearm, pinky and thumb) in tumor patients with the robot-aided TMS protocol and with direct stimulation. The resulting maps agree within 5mm.

Prototype of an intraoperative navigation and documentation system for laparoscopic radiofrequency ablation: first experiences.

AIMS: Laparoscopic radiofrequency ablation (RFA) is an accepted approach to treat unresectable liver tumours, distinguishing itself from other techniques by combining minimal invasiveness and the advantages of a surgical approach. The major task of laparoscopic RFA is the accurate needle placement according to preoperative planning to achieve complete tumour ablation. This study investigates the value of an image-guided surgery system to accomplish this task. METHODS: An image-guided surgery system for laparoscopic liver treatments (LapAssistent) based on a 3D-navigation scene was developed. A laparoscopic ultrasound probe and a RFA needle could be navigated using an electromagnetic tracking system. The system was studied using a perfused tumour-mimic-model of a porcine liver. Navigating the RFA needle, the tumours were ablated. RESULTS: The system enables the surgeon to intraoperatively update the three-dimensional planning data in case of new findings. The RFA needle could be placed accurately in a targeted tumour, even out of the ultrasound plane. In case of multiple tumours lying in close spatial relationship, the documentation module helps to keep track of the already ablated tumours and those that still need to be treated. CONCLUSION: The system adds benefit to laparoscopic RFA enabling the surgeon to place the needle accurately inside the targeted tumours using the navigation scene. A manual alignment of the preoperative data to the physical space produces a feasible result for a restricted region. A precise measurement of the accuracy of this process has to be done. The possibility to update the three-dimensional model with new intraoperative findings enables the surgeon to adapt to a new intraoperative situation. Furthermore the possibility to mark ablated tumours helps to keep track of the operation plan.

Motion and positional error correction for cone beam 3D-reconstruction with mobile C-arms.

CT-images acquired by mobile C-arm devices can contain artefacts caused by positioning errors. We propose a data driven method based on iterative 3D-reconstruction and 2D/3D-registration to correct projection data inconsistencies. With a 2D/3D-registration algorithm, transformations are computed to align the acquired projection images to a previously reconstructed volume. In an iterative procedure, the reconstruction algorithm uses the results of the registration step. This algorithm also reduces small motion artefacts within 3D-reconstructions. Experiments with simulated projections from real patient data show the feasibility of the proposed method. In addition, experiments with real projection data acquired with an experimental robotised C-arm device have been performed with promising results.

Technical description, phantom accuracy, and clinical feasibility for single-session lung radiosurgery using robotic image-guided real-time respiratory tumor tracking.

To describe the technological background, the accuracy, and clinical feasibility for single session lung radiosurgery using a real-time robotic system with respiratory tracking. The latest version of image-guided real-time respiratory tracking software (Synchrony, Accuray Incorporated, Sunnyvale, CA) was applied and is described. Accuracy measurements were performed using a newly designed moving phantom model. We treated 15 patients with 19 lung tumors with robotic radiosurgery (CyberKnife, Accuray) using the same treatment parameters for all patients. Ten patients had primary tumors and five had metastatic tumors. All patients underwent computed tomography-guided percutaneous placement of one fiducial directly into the tumor, and were all treated with single session radiosurgery to a dose of 24 Gy. Follow up CT scanning was performed every two months. All patients could be treated with the automated robotic technique. The respiratory tracking error was less than 1 mm and the overall shape of the dose profile was not affected by target motion and/or phase shift between fiducial and optical marker motion. Two patients required a chest tube insertion after fiducial implantation because of pneumothorax. One patient experienced nausea after treatment. No other short-term adverse reactions were found. One patient showed imaging signs of pneumonitis without a clinical correlation. Single-session radiosurgery for lung tumor tracking using the described technology is a stable, safe, and feasible concept for respiratory tracking of tumors during robotic lung radiosurgery in selected patients. Longer follow-up is needed for definitive clinical results.

[Fluoroscopy-based 3D navigation of complex correction osteotomies at the proximal femur]. / Fluoroskopiebasierte 3D-Navigation komplexer Korrekturosteotomien des proximalen Femurs.

BACKGROUND: Despite great advances in hip alloarthroplasty there are still numerous indications for joint-saving procedures such as correction osteotomies. Often these procedures include complex 3D rearrangements of the proximal femur, which are for the surgeon technically very demanding. The project aim was to develop a precise intraoperative virtual 3D planning tool including a detailed biomechanical analysis and enable the surgeon to realize exactly this plan by using computer-assisted techniques. METHODS: Using only two different angled fluoro frames a simplified femoral model was inversely constructed. For navigation a passive optical system was used with a C-arm calibration kit and PC-based software. For in vitro evaluation complex osteotomies were performed on ten femora under simulated OR conditions. RESULTS: The mean difference between the planning and real surgical outcome for the wedge size was less then 2 degrees and for the femur head center position less then 4 mm. No implant penetrated the femur neck isthmus. CONCLUSION: Without changing the standard operative procedure the method can be of high clinical importance to improve planning accuracy and consecutive operative realization for precise fragment positioning and plate location without penetrating the isthmus of the femoral neck. And -- besides precision -- it can potentially help to reduce intraoperative complications such as implant penetration and minimize X-ray use.

Respiration tracking in radiosurgery without fiducials.

In robotic radiosurgery, a focused beam of radiation is moved by a robot arm. We investigated methods for soft-tissue navigation using robotic radiosurgery. In previous work we described a method for real-time tracking based on correlation between the motion of implanted fiducial markers and external skin markers. In this work we extend our method of correlation-based tracking to tracking without implanted fiducials. We propose to use deformation algorithms on CT data sets combined with registration of digitally reconstructed radiographs and intra-treatment X-ray images to obtain intermittent information on the target location. This information is then combined with our basic correlation method to achieve real-time tracking. Our study investigates the feasibility of this approach from the point of view of computing time and required level of user interaction. The term 7D registration is coined to describe the underlying method for performing this task.