Registration of angiographic image on real-time fluoroscopic image for image-guided percutaneous coronary intervention

Purpose

In percutaneous coronary intervention (PCI), cardiologists must study two different X-ray image sources: a fluoroscopic image and an angiogram. Manipulating a guidewire while alternately monitoring the two separate images on separate screens requires a deep understanding of the anatomy of coronary vessels and substantial training. We propose 2D/2D spatiotemporal image registration of the two images in a single image in order to provide cardiologists with enhanced visual guidance in PCI.

Methods

The proposed 2D/2D spatiotemporal registration method uses a cross-correlation of two ECG series in each image to temporally synchronize two separate images and register an angiographic image onto the fluoroscopic image. A guidewire centerline is then extracted from the fluoroscopic image in real time, and the alignment of the centerline with vessel outlines of the chosen angiographic image is optimized using the iterative closest point algorithm for spatial registration.

Results

A proof-of-concept evaluation with a phantom coronary vessel model with engineering students showed an error reduction rate greater than 74% on wrong insertion to nontarget branches compared to the non-registration method and more than 47% reduction in the task completion time in performing guidewire manipulation for very difficult tasks. Evaluation with a small number of experienced doctors shows a potentially significant reduction in both task completion time and error rate for difficult tasks. The total registration time with real procedure X-ray (angiographic and fluoroscopic) images takes \(\sim \) 60 ms, which is within the fluoroscopic image acquisition rate of 15 Hz.

Conclusions

By providing cardiologists with better visual guidance in PCI, the proposed spatiotemporal image registration method is shown to be useful in advancing the guidewire to the coronary vessel branches, especially those difficult to insert into.

     

Fast and robust extraction of surrogate respiratory signal from intra-operative liver ultrasound images

Purpose

   In model-based respiratory motion estimation for the liver or other abdominal organs, the surrogate respiratory signal is usually obtained by using special tracking devices from skin or diaphragm, and subsequently applied to parameterize a 4D motion model for prediction or compensation. However, due to the intrinsic limits and economical costs of these tracking devices, the identification of the respiratory signal directly from intra-operative ultrasound images is a more attractive alternative.

Methods

   We propose a fast and robust method to extract the respiratory motion of the liver from an intra-operative 2D ultrasound image sequence. Our method employs a preprocess to remove speckle-like noises in the ultrasound images and utilizes the normalized cross-correlation to measure the image similarity fast. More importantly, we present a novel adaptive search strategy, which makes full use of the inter-frame dependency of the image sequence. This search strategy narrows the search range of the optimal matching, thus greatly reduces the search time, and makes the matching process more robust and accurate.

Results

   The experimental results on four volunteers demonstrate that our method is able to extract the respiratory signal from an image sequence of 256 image frames in 5 s. The quantitative evaluation using the correlation coefficient reveals that the respiratory motion, extracted near the liver boundaries and vessels, is highly consistent with the reference motion tracked by an EM device.

Conclusions

   Our method can use 2D ultrasound to track natural landmarks from the liver as surrogate respiratory signal and hence provide a feasible solution to replace special tracking devices.     

Shape–intensity prior level set combining probabilistic atlas and probability map constrains for automatic liver segmentation from abdominal CT images

Purpose

Propose a fully automatic 3D segmentation framework to segment liver on challenging cases that contain the low contrast of adjacent organs and the presence of pathologies from abdominal CT images.

Methods

First, all of the atlases are weighted in the selected training datasets by calculating the similarities between the atlases and the test image to dynamically generate a subject-specific probabilistic atlas for the test image. The most likely liver region of the test image is further determined based on the generated atlas. A rough segmentation is obtained by a maximum a posteriori classification of probability map, and the final liver segmentation is produced by a shape–intensity prior level set in the most likely liver region. Our method is evaluated and demonstrated on 25 test CT datasets from our partner site, and its results are compared with two state-of-the-art liver segmentation methods. Moreover, our performance results on 10 MICCAI test datasets are submitted to the organizers for comparison with the other automatic algorithms.

Results

Using the 25 test CT datasets, average symmetric surface distance is \(1.09 \pm 0.34\) mm (range 0.62–2.12 mm), root mean square symmetric surface distance error is \(1.72 \pm 0.46\) mm (range 0.97–3.01 mm), and maximum symmetric surface distance error is \(18.04 \pm 3.51\) mm (range 12.73–26.67 mm) by our method. Our method on 10 MICCAI test data sets ranks 10th in all the 47 automatic algorithms on the site as of July 2015. Quantitative results, as well as qualitative comparisons of segmentations, indicate that our method is a promising tool to improve the efficiency of both techniques.

Conclusion

The applicability of the proposed method to some challenging clinical problems and the segmentation of the liver are demonstrated with good results on both quantitative and qualitative experimentations. This study suggests that the proposed framework can be good enough to replace the time-consuming and tedious slice-by-slice manual segmentation approach.

     

2D-3D shape reconstruction of the distal femur from stereo X-ray imaging using statistical shape models

Three-dimensional patient specific bone models are required in a range of medical applications, such as pre-operative surgery planning and improved guidance during surgery, modeling and simulation, and in vivo bone motion tracking. Shape reconstruction from a small number of X-ray images is desired as it lowers both the acquisition costs and the radiation dose compared to CT. We propose a method for pose estimation and shape reconstruction of 3D bone surfaces from two (or more) calibrated X-ray images using a statistical shape model (SSM). User interaction is limited to manual initialization of the mean shape. The proposed method combines a 3D distance based objective function with automatic edge selection on a Canny edge map. Landmark-edge correspondences are weighted based on the orientation difference of the projected silhouette and the corresponding image edge. The method was evaluated by rigid pose estimation of ground truth shapes as well as 3D shape estimation using a SSM of the whole femur, from stereo cadaver X-rays, in vivo biplane fluoroscopy image-pairs, and an in vivo biplane fluoroscopic sequence. Ground truth shapes for all experiments were available in the form of CT segmentations. Rigid registration of the ground truth shape to the biplane fluoroscopy achieved sub-millimeter accuracy (0.68 mm) measured as root mean squared (RMS) point-to-surface (P2S) distance. The non-rigid reconstruction from the biplane fluoroscopy using the SSM also showed promising results (1.68 mm RMS P2S). A feasibility study on one fluoroscopic time series illustrates the potential of the method for motion and shape estimation from fluoroscopic sequences with minimal user interaction.     

Feasibility of respiratory motion-compensated stereoscopic X-ray tracking for bronchoscopy

Purpose

   Precise localization in bronchoscopy is challenging, particularly for peripheral lesions that cannot be reached by conventional bronchoscopes with a large working channel. Existing navigation methods are hampered by respiratory motion, e.g., in the lower lobes. We present an image-guided approach that considers respiratory motion and can localize instruments.

Methods

   We developed a rigid chest marker containing steel balls visible in X-ray images and a pattern for passive tracking with an optical camera system. An experimental setup to evaluate stereoscopic localization and to mimic chest motion was established in our interventional suite. The marker motion was recorded, and X-ray images were acquired from different angles using a standard C-arm. All coordinates were expressed with respect to the stationary tracking camera. The feasibility of motion-compensated stereoscopic localization was assessed.

Results

   The orientation of the C-arm could be established with a mean error of less than $1^{\circ }$ . Triangulation based on two different X-ray images from different angles resulted in a mean error of 1.8 ( $\pm $ 0.7) mm. A similar result was obtained when the marker was moved between X-ray acquisitions, and the mean error was 1.6 ( $\pm $ 1.4) mm. The latencies were approximately 80 and 380 ms for tracking camera and X-ray imaging, respectively. Stereoscopic localization of a moving target was feasible.

Conclusions

   The system presents a flexible alternative for precise stereoscopic localization of a bronchoscope or instruments using a standard C-arm. We demonstrated the ability to track multiple moving markers and to compensate for respiratory motion.     

Blood flow-induced physically based guidewire simulation for vascular intervention training

Purpose

A realistic guidewire behavior simulation is a vital component of a virtual vascular intervention system. Such systems are a safe, low-cost means of establishing a training environment to help inexperienced surgeons develop their intervention skills. Previous attempts to simulate the complex movement of a guidewire inside blood vessels have rarely considered the influence of blood flow. In this paper, we address this problem by integrating blood flow analysis and propose a novel guidewire simulation model.

Methods

The blood flow distribution inside the arterial vasculature was computed by separating the vascular model into discrete cylindrical vessels and modeling the flow in each vessel according to Poiseuille Law. The blood flow computation was then integrated into a robust Kirchhoff elastic model. With hardware acceleration, the guidewire simulation can be run in real time. To evaluate the simulation, an experimental environment with a 3D-printed vascular phantom and an electromagnetic tracking system was set up, with clinically used guidewire sensors applied to trace its motion as the standard for comparison.

Results

The virtual guidewire motion trace was assessed by comparing it to the comparison standard. The root-mean-square (RMS) value of the newly proposed guidewire model was 2.14 mm ± 1.24 mm, lower than the value of 4.81 mm ± 3.80 mm for the previous Kirchhoff model, while maintaining a computation speed of at least 30 fps.

Conclusion

The experimental results revealed that the blood flow-induced model exhibits better performance and physical credibility with a lower and more stable RMS error than the previous Kirchhoff model.

     

Respiratory motion compensation by model-based catheter tracking during EP procedures

In many cases, radio-frequency catheter ablation of the pulmonary veins attached to the left atrium still involves fluoroscopic image guidance. Two-dimensional X-ray navigation may also take advantage of overlay images derived from static pre-operative 3D volumetric data to add anatomical details otherwise not visible under X-ray. Unfortunately, respiratory motion may impair the utility of static overlay images for catheter navigation. We developed a novel approach for image-based 3D motion estimation and compensation as a solution to this problem. It is based on 3D catheter tracking which, in turn, relies on 2D/3D registration. To this end, a bi-plane C-arm system is used to take X-ray images of a special circumferential mapping catheter from two directions. In the first step of the method, a 3D model of the device is reconstructed. Three-dimensional respiratory motion at the site of ablation is then estimated by tracking the reconstructed catheter model in 3D based on bi-plane fluoroscopy. Phantom data and clinical data were used to assess model-based catheter tracking. Our phantom experiments yielded an average 2D tracking error of 1.4 mm and an average 3D tracking error of 1.1 mm. Our evaluation of clinical data sets comprised 469 bi-plane fluoroscopy frames (938 monoplane fluoroscopy frames). We observed an average 2D tracking error of 1.0 ± 0.4 mm and an average 3D tracking error of 0.8 ± 0.5 mm. These results demonstrate that model-based motion-compensation based on 2D/3D registration is both feasible and accurate.     

Respiratory-Induced Errors in Tumor Quantification and Delineation in CT Attenuation-Corrected PET Images: Effects of Tumor Size, Tumor Location, and Respiratory Trace: A Simulation Study Using the 4D XCAT Phantom

Purpose

We investigated the magnitude of respiratory-induced errors in tumor maximum standardized uptake value (SUV_max), localization, and volume for different respiratory motion traces and various lesion sizes in different locations of the thorax and abdomen in positron emission tomography (PET) images.

Procedures

Respiratory motion traces were simulated based on the common patient breathing cycle and three diaphragm motions used to drive the 4D XCAT phantom. Lesions with different diameters were simulated in different locations of lungs and liver. The generated PET sinograms were subsequently corrected using computed tomography attenuation correction involving the end exhalation, end inhalation, and average of the respiratory cycle. By considering respiration-averaged computed tomography as a true value, the lesion volume, displacement, and SUV_max were measured and analyzed for different respiratory motions.

Results

Respiration with 35-mm diaphragm motion results in a mean lesion SUV_max error of 24 %, a mean superior inferior displacement of 7.6 mm and a mean lesion volume overestimation of 129 % for a 9-mm lesion in the liver. Respiratory motion results in lesion volume overestimation of 50 % for a 9-mm lower lung lesion near the liver with just 15-mm diaphragm motion. Although there are larger errors in lesion SUV_max and volume for 35-mm motion amplitudes, respiration-averaged computed tomography results in smaller errors than the other two phases, except for the lower lung region.

Conclusions

The respiratory motion-induced errors in tumor quantification and delineation are highly dependent upon the motion amplitude, tumor location, tumor size, and choice of the attenuation map for PET image attenuation correction.     

Rest and stress transluminal attenuation gradient and contrast opacification difference for detection of hemodynamically significant stenoses in patients with suspected coronary artery disease

This study evaluated the feasibility of stress 320 detector CT coronary angiography (CTA) derived transluminal attenuation gradient (TAG320) and contrast opacification (CO) difference to detect hemodynamically significant stenoses as determined by invasive fractional flow reserve (FFR ≤ 0.80). Twenty-seven patients, including 51 vessels on rest CTA were studied. 16 (31 %) vessels were not interpretable on stress CTA largely secondary to motion artefacts. Receiver operating characteristic curve analysis showed a comparable area under the curve (AUC) for rest and stress TAG320 (0.78 and 0.75) which was higher than CTA alone (0.68), and rest and stress CO difference (0.76 and 0.67). Compared with rest CTA, stress CTA demonstrated inferior image quality (Median Likert score 4 vs. 3, P < 0.0001) and required a higher mean radiation exposure (3.2 vs. 5.1 mSv, P < 0.0001). Stress TAG320 and CO difference is less feasible and was not superior in diagnostic performance when compared with rest TAG320 and CO difference.     

Positioning error evaluation of GPU-based 3D ultrasound surgical navigation system for moving targets by using optical tracking system

Purpose

A near real-time three-dimensional (3D) ultrasound navigation system has been developed for guiding surgery involving internal organs that move and change shape (e.g., abdominal surgery, fetal surgery). In practical applications, significant errors arise between the actual navigation-image positions depending on the time delay of the system. Therefore, the positioning error of the system relative to the target velocity was evaluated.

Methods

We developed a method for evaluating the positioning error of a graphics processing unit-based 3D ultrasound surgical navigation system (with an optical tracking system) for moving targets. The effectiveness of this system was quantitatively evaluated in terms of its image processing runtime, target registration error (TRE), and positioning error for a moving target. The positioning error was evaluated for a phantom (with an optical tracking marker) moving at speeds of 5–25 mm/s, and the navigation target was the center point of the phantom. The imaging range of the volume data was set to the maximum angle and range of the ultrasound diagnostic system (update rate: 4 Hz).

Results

The image processing runtime was 27.43 ± 4.80 ms, and the TRE was 1.50 ± 0.28 mm. The positioning error was 4.24 ± 0.12 mm for a target moving at a speed of 10 mm/s and 5.36 ± 0.10 mm for one moving at 15 mm/s.

Conclusion

The effectiveness of an ultrasound navigation system was quantitatively evaluated by using the positioning error for a moving target. This navigation system demonstrated high calculation speed and positioning accuracy for a moving target. Therefore, it is suitable to guide the surgery of abdominal internal organs (e.g., in fetal and abdominal surgeries) that move or change shape during breathing and surgical approaches.     

Toward Fluoro-Free Interventions: Using Radial Intracardiac Ultrasound for Vascular Navigation

Transcatheter cardiovascular interventions have the advantage of patient safety, reduced surgery time and minimal trauma to the patient's body. Transcathether interventions, which are performed percutaneously, are limited by the lack of direct line of sight with the procedural tools and the patient anatomy. Therefore, such interventional procedures rely heavily on image guidance for navigating toward and delivering therapy at the target site. Vascular navigation via the inferior vena cava, from the groin to the heart, is an imperative part of most transcatheter cardiovascular interventions including heart valve repair surgeries and ablation therapy. Traditionally, the inferior vena cava is navigated using fluoroscopic techniques such as venography and computed tomography venography. These X-ray–based techniques can have detrimental effects on the patient as well as the surgical team, causing increased radiation exposure, leading to risk of cancer, fetal defects and eye cataracts. The use of a heavy lead apron has also been reported to cause back pain and spine issues, thus leading to interventionalist's disc disease. We propose the use of a catheter-based ultrasound augmented with electromagnetic tracking technology to generate a vascular roadmap in real time and perform navigation without harmful radiation. In this pilot study, we used spatially tracked intracardiac echocardiography to reconstruct a vessel from a phantom in a 3-D virtual environment. We illustrate how the proposed ultrasound-based navigation will appear in a virtual environment, by navigating a tracked guidewire within the vessels in the phantom without any radiation-based imaging. The geometric accuracy is assessed using a computed tomography scan of the phantom, with a Dice coefficient of 0.79. The average distance between the surfaces of the two models comes out to be 1.7 ± 1.12 mm.     

X-ray and optical stereo-based 3D sensor fusion system for image-guided neurosurgery

Purpose

In neurosurgery, an image-guided operation is performed to confirm that the surgical instruments reach the exact lesion position. Among the multiple imaging modalities, an X-ray fluoroscope mounted on C- or O-arm is widely used for monitoring the position of surgical instruments and the target position of the patient. However, frequently used fluoroscopy can result in relatively high radiation doses, particularly for complex interventional procedures. The proposed system can reduce radiation exposure and provide the accurate three-dimensional (3D) position information of surgical instruments and the target position.

Methods

X-ray and optical stereo vision systems have been proposed for the C- or O-arm. Two subsystems have same optical axis and are calibrated simultaneously. This provides easy augmentation of the camera image and the X-ray image. Further, the 3D measurement of both systems can be defined in a common coordinate space.

Results

The proposed dual stereoscopic imaging system is designed and implemented for mounting on an O-arm. The calibration error of the 3D coordinates of the optical stereo and X-ray stereo is within 0.1 mm in terms of the mean and the standard deviation. Further, image augmentation with the camera image and the X-ray image using an artificial skull phantom is achieved.

Conclusion

As the developed dual stereoscopic imaging system provides 3D coordinates of the point of interest in both optical images and fluoroscopic images, it can be used by surgeons to confirm the position of surgical instruments in a 3D space with minimum radiation exposure and to verify whether the instruments reach the surgical target observed in fluoroscopic images.

     

Automatic 3D reconstruction of electrophysiology catheters from two-view monoplane C-arm image sequences

Purpose

Catheter guidance is a vital task for the success of electrophysiology interventions. It is usually provided through fluoroscopic images that are taken intra-operatively. The cardiologists, who are typically equipped with C-arm systems, scan the patient from multiple views rotating the fluoroscope around one of its axes. The resulting sequences allow the cardiologists to build a mental model of the 3D position of the catheters and interest points from the multiple views.

Method

We describe and compare different 3D catheter reconstruction strategies and ultimately propose a novel and robust method for the automatic reconstruction of 3D catheters in non-synchronized fluoroscopic sequences. This approach does not purely rely on triangulation but incorporates prior knowledge about the catheters. In conjunction with an automatic detection method, we demonstrate the performance of our method compared to ground truth annotations.

Results

In our experiments that include 20 biplane datasets, we achieve an average reprojection error of 0.43 mm and an average reconstruction error of 0.67 mm compared to gold standard annotation.

Conclusions

In clinical practice, catheters suffer from complex motion due to the combined effect of heartbeat and respiratory motion. As a result, any 3D reconstruction algorithm via triangulation is imprecise. We have proposed a new method that is fully automatic and highly accurate to reconstruct catheters in three dimensions.

     

Robot-assisted automatic ultrasound calibration

Purpose

Ultrasound (US) calibration is the process of determining the unknown transformation from a coordinate frame such as the robot’s tooltip to the US image frame and is a necessary task for any robotic or tracked US system. US calibration requires submillimeter-range accuracy for most applications, but it is a time-consuming and repetitive task. We provide a new framework for automatic US calibration with robot assistance and without the need for temporal calibration.

Method

US calibration based on active echo (AE) phantom was previously proposed, and its superiority over conventional cross-wire phantom-based calibration was shown. In this work, we use AE to guide the robotic arm motion through the process of data collection; we combine the capability of the AE point to localize itself in the frame of the US image with the automatic motion of the robotic arm to provide a framework for calibrating the arm to the US image automatically.

Results

We demonstrated the efficacy of the automated method compared to the manual method through experiments. To highlight the necessity of frequent ultrasound calibration, it is demonstrated that the calibration precision changed from 1.67 to 3.20 mm if the data collection is not repeated after a dismounting/mounting of the probe holder. In a large data set experiment, similar reconstruction precision of automatic and manual data collection was observed, while the time was reduced by 58 %. In addition, we compared ten automatic calibrations with ten manual ones, each performed in 15 min, and showed that all the automatic ones could converge in the case of setting the initial matrix as identity, while this was not achieved by manual data sets. Given the same initial matrix, the repeatability of the automatic was [0.46, 0.34, 0.80, 0.47] versus [0.42, 0.51, 0.98, 1.15] mm in the manual case for the US image four corners.

Conclusions

The submillimeter accuracy requirement of US calibration makes frequent data collections unavoidable. We proposed an automated calibration setup and showed feasibility by implementing it for a robot tooltip to US image calibration. The automated method showed a similar reconstruction precision as well as repeatability compared to the manual method, while the time consumed for data collection was reduced. The automatic method also reduces the burden of data collection for the user. Thus, the automated method can be a viable solution for applications that require frequent calibrations.

     

Comparison of finite helical axes of normal and anatomically designed prosthetic knees

BackgroundDescribing three-dimensional joint motion using the finite helical axis has an advantage in understanding unknown coupling motion in prosthetic knee joints. We aimed to examine the differences in the orientations of finite helical axis of normal and anatomically designed cruciate-retaining and posterior-stabilized prosthetic knees after total knee arthroplasty.MethodsTen normal, 40 cruciate-retaining prosthetic knees of 33 patients and 19 posterior-stabilized prosthetic knees of 14 patients enabling to flex > 120° were analyzed during a squatting motion with deep knee bending. The motion was recorded by a fluoroscopic imaging system, and the pose of the bone and prostheses were determined by an image registration technique. The finite helical axes were calculated using 30° window.FindingsThe finite helical axis in the early flexion phase of the normal knees had a greater inferior inclination (mean − 19.0° (SD 7.2°)) than those of the cruciate-retaining (mean − 1.7 (SD 5.0°)) and posterior-stabilized (mean − 2.9° (SD 5.5°)) prosthetic knees (p < 0.001), and became almost horizontal and constant in the mid to deep flexion phases. In contrast, the cruciate-retaining and posterior-stabilized prosthetic knees demonstrated slightly inclined and almost constant vertical angles throughout the all phases.InterpretationThese results demonstrate that, in the normal knee, a clear coupling motion occurs during the early flexion phase. For the cruciate-retaining and posterior-stabilized prosthetic knees, an unclear coupling motion exists during all phases. These results suggest that the physiological motion is not possible to reproduce using shape-guided motion only even in an anatomically designed prosthetic knee.     

Catheter tip tracking for MR-guided interventions using discrete Kalman filter and mean shift localization

Purpose

This paper presents and evaluates stochastic computer algorithms used to automatically detect and track marked catheter tip during MR-guided catheterization. The algorithms developed employ extraction and matching of regional features of the catheter tip to perform the localization.

Method

To perform the tracking, a probability map that indicates the possible locations of the catheter tip in the MR images is first generated. This map is generated from the similarity to a given marker template. The method to assess the similarity between the marker template image and the different positions on each MR frame is based on speeded-up robust features extracted from the gradient image. The probability map is then used in two different stochastic localization frameworks mean shift (MS) localization and Kalman filter (KF) to track the position of the catheter using pairs of orthogonal projection of 2D MR images. The algorithm developed was tested on catheter tip marked with LC resonant circuit (of size $2\,\hbox {mm}\,\times \,2\,\hbox {cm}$ ) tuned to the Larmor frequency of the MRI scanner and for different image resolutions (1, 3, 5 and 7 mm squared pixel).

Results

The tracking performance was very robust for the two algorithms MS and KF with image resolution as low as 3 mm where the localization error was about 1 mm for KF and 0.9 mm for MS. For the 5-mm resolution images, the error was 2.2 mm for both KF and MS, and for the 7-mm resolution images, the error was 3.6 and 3.7 mm for KF and MS, respectively.

Conclusion

Both KF and MS gave comparable results when it comes to accuracy for the different image resolutions. The results showed that the two tracking algorithms tracked the catheter tip with high robustness for image resolution of 3 mm and with acceptable reliability for image resolution as poor as 5 mm with the resonant marker configuration used.     

Four-dimensional radiotherapeutic dose calculation using biomechanical respiratory motion description

Purpose

   Organ motion due to patient breathing introduces a technical challenge for dosimetry and lung tumor treatment by hadron therapy. Accurate dose distribution estimation requires patient-specific information on tumor position, size, and shape as well as information regarding the material density and stopping power of the media along the beam path. A new 4D dosimetry method was developed, which can be coupled to any motion estimation method. As an illustration, the new method was implemented and tested with a biomechanical model and clinical data.

Methods

   First, an anatomical model of the lung and tumor was synthesized with deformable tetrahedral grids using computed tomography (CT) images. The CT attenuation values were estimated at the grid vertices. Respiratory motion was simulated biomechanically based on nonlinear finite element analysis. Contrary to classical image-based methods where motion is described using deformable image registration algorithms, the dose distribution was accumulated over tetrahedral meshes that are deformed using biomechanical modeling based on finite element analysis.

Results

   The new method preserves the mass of the objects during simulation with an error between 1.6 and 3.6 %. The new method was compared to an existing dose calculation method demonstrating significant differences between the two approaches and overall superior performance using the new method.

Conclusion

   A unified model of 4D radiotherapy respiratory effects was developed where biomechanical simulations are coupled with dose calculations. Promising results demonstrate that this approach has significant potential for the treatment for moving tumors.     

Registration of 3D trans-esophageal echocardiography to X-ray fluoroscopy using image-based probe tracking

Two-dimensional (2D) X-ray imaging is the dominant imaging modality for cardiac interventions. However, the use of X-ray fluoroscopy alone is inadequate for the guidance of procedures that require soft-tissue information, for example, the treatment of structural heart disease. The recent availability of three-dimensional (3D) trans-esophageal echocardiography (TEE) provides cardiologists with real-time 3D imaging of cardiac anatomy. Increasingly X-ray imaging is now supported by using intra-procedure 3D TEE imaging. We hypothesize that the real-time co-registration and visualization of 3D TEE and X-ray fluoroscopy data will provide a powerful guidance tool for cardiologists. In this paper, we propose a novel, robust and efficient method for performing this registration. The major advantage of our method is that it does not rely on any additional tracking hardware and therefore can be deployed straightforwardly into any interventional laboratory. Our method consists of an image-based TEE probe localization algorithm and a calibration procedure. While the calibration needs to be done only once, the GPU-accelerated registration takes approximately from 2 to 15 s to complete depending on the number of X-ray images used in the registration and the image resolution. The accuracy of our method was assessed using a realistic heart phantom. The target registration error (TRE) for the heart phantom was less than 2 mm. In addition, we assess the accuracy and the clinical feasibility of our method using five patient datasets, two of which were acquired from cardiac electrophysiology procedures and three from trans-catheter aortic valve implantation procedures. The registration results showed our technique had mean registration errors of 1.5-4.2 mm and 95% capture range of 8.7-11.4 mm in terms of TRE.     

Registration of a CT-like atlas to fluoroscopic X-ray images using intensity correspondences

Objective  We present a novel method for intraoperative image-based bone surface reconstruction and its validation. Materials and methods  In the preoperative stage, we construct a CT-like intensity atlas of the anatomy of interest. In the intraoperative stage, we deformably register this atlas to fluoroscopic X-ray images of the patient anatomy. We iteratively refine the atlas-to-patient registration by establishing explicit correspondences between bone surfaces in the atlas and their projections in the X-ray images. The advantage of our method is its use of CT-quality intensity data for correspondence establishment, which eliminates the edge-detection problem and diminishes the miscorrespondence problem. We validate our method on two datasets: (1) an in vitro dry femur; (2) Digitally Reconstructed Radiographs, which were generated from 17 clinical CTs, and simulate realistic in vivo femurs. Results  The mean surface approximation error of our femur atlas was 0.85 ± 0.16 mm. On Digitally Reconstructed Radiographs, the mean surface reconstruction error was 1.40 ± 0.55 mm. On a dry femur, the mean surface reconstruction error was 1.44 mm. Conclusion  The results show that our reconstruction method is on par with the state of the art in reconstruction of ex vivo femurs. In addition, the results demonstrate that our method is effective in realistic simulations of the in vivo scenario.     

Automatic needle detection and real-time Bi-planar needle visualization during 3D ultrasound scanning of the liver

2D ultrasound (US) image guidance is used in minimally invasive procedures in the liver to visualize the target and the needle. Needle insertion using 2D ultrasound keeping the transducer position to view needle and reach target is challenging. Dedicated needle holders attached to the US transducer help to target in plane and at a specific angle. A drawback of this is that, the probe is fixed to the needle and cannot be rotated to assess the position of the needle in a perpendicular plane. In this study, we propose an automatic needle detection and tracking method using 3D US imaging to improve image guidance and visualization of the target in the liver with respect to the needle during these interventional procedures. The method utilizes a convolutional neural network for detection of the needle in 3D US images. In a subsequent step, the output of the convolutional neural network is used to detect needle candidates, which are fed into a final tracking step to determine the real needle position. The needle position is used to present two perpendicular cross-sectional planes of the 3D US image containing the needle in both directions. Performance of the method was evaluated in phantoms and in-vivo data by calculating the needle position distance and needle orientation angle between segmented needles and reference ground truth needles, which were manually annotated by an observer. The method successfully detects the needle position and orientation with mean errors of 1 mm and 2°, respectively. The proposed method yields a robust automatic needle detection and visualization at a frame rate of 3 Hz in 3D ultrasound imaging of the liver.