We are developing a video see-through head-mounted display (HMD) augmented reality (AR) system for image-guided neurosurgical planning and navigation. The surgeon wears a HMD that presents him with the augmented stereo view. The HMD is custom fitted with two miniature color video cameras that capture a stereo view of the real-world scene. We are concentrating specifically at this point on cranial neurosurgery, so the images will be of the patient's head. A third video camera, operating in the near infrared, is also attached to the HMD and is used for head tracking. The pose (i.e., position and orientation) of the HMD is used to determine where to overlay anatomic structures segmented from preoperative tomographic images (e.g., CT, MR) on the intraoperative video images. Two SGI 540 Visual Workstation computers process the three video streams and render the augmented stereo views for display on the HMD. The AR system operates in real time at 30 frames/sec with a temporal latency of about three frames (100 ms) and zero relative lag between the virtual objects and the real-world scene. For an initial evaluation of the system, we created AR images using a head phantom with actual internal anatomic structures (segmented from CT and MR scans of a patient) realistically positioned inside the phantom. When using shaded renderings, many users had difficulty appreciating overlaid brain structures as being inside the head. When using wire frames, and texture-mapped dot patterns, most users correctly visualized brain anatomy as being internal and could generally appreciate spatial relationships among various objects. The 3D perception of these structures is based on both stereoscopic depth cues and kinetic depth cues, with the user looking at the head phantom from varying positions. The perception of the augmented visualization is natural and convincing. The brain structures appear rigidly anchored in the head, manifesting little or no apparent swimming or jitter. The initial evaluation of the system is encouraging, and we believe that AR visualization might become an important tool for image-guided neurosurgical planning and navigation.
Purpose: Extraneous irradiation is a major concern in pediatric radiation oncology. A modified version of the Siemens MV CBCT has been developed that allows for low dose and high contrast imaging. We will determine the required setup margin stratified by age, tumor site, sedation, and treatment position in an IRB approved pediatric localization protocol of where low‐dose megavoltage cone‐beam CT (LD‐MV CBCT) is a central component. Method and Materials: Research participants are assigned a non‐ionizing localization technique based on disease site, treatment position, age, and use of anesthesia. Along with the protocol localization technique, each research participant will receive a LD‐MV CBCT at the start of each fraction and the end of every other fraction. Per protocol, the LD‐MV CBCT delivers only 1.0cGy per scan at isocenter for each research participant. The proper CBCT settings are determined with aid of a treatment planning system. The pre‐treatment LD‐MV CBCT will be used to assess the accuracy of the localization technique and the post‐treatment LD‐MV CBCT will quantify the rigidity of the immobilization schema. The resultant data will provide quantitative information about the inter‐ and intra‐fractional motion of the target, which comprises the SM portion of the PTV. Results: In one week, the dose to the research participant due to the daily pre and post‐treatment LD‐MV CBCT will be approximately 7.5cGy. The CBCT will supplant routine verification ports films, which, if obtained twice a week, would deliver approximately 5.0cGy at isocenter. The research participant will receive only 2.5cGy a week in additional dose when daily LD‐MV CBCT is used compared to twice a week port films. Conclusion: We will have motion and imaging data on 40 research participants by August 2008, with a planned 200 research participants for the protocol. Conflict of Interest: Supported in part by Siemens Medical USA.
Purpose: The purpose of this work is to devise an efficient, effective and routine imaging modality to guide lung radiotherapy. Current methods involve acquiring a 4D-CBCT and comparing the digitally reconstructed radiographs (DRRs) for multiple breathing phases with online fluoroscopic images. A major shortcoming of DRRs and fluoroscopy is that unwanted structures such as bones occlude the target. Furthermore 4D-CBCT requires long acquisition times and large patient dose, making it impractical for routine use. Method and Materials: We propose a partial arc cone beam acquisition, which we call “cone beam digital tomosynthesis” (CBDT), to obtain cross-sectional images of a slab just thick enough to enclose the soft tissue target. Projections through this slab make “digitally reconstructed tomograms” (DRTs). Similar to DRRs, DRTs correct for beam divergence. However, different from DRRs, DRTs do not contain irrelevant structures outside the slab, making the target far more conspicuous. By gating this acquisition, dynamic cross sections and DRTs are obtained at multiple respiratory phases. These dynamic images are registered with those obtained from the planning 4D-CT dataset to guide treatment. Results: The DRRs constructed from multiple phases of a 4D-CT emphasize bony anatomy and other irrelevant structures overlaying the target, making the edges of the target difficult to find. We have demonstrated that this difficulty is overcome by the use of cross-sectional images from CBDT and DRTs, which are obtained with an image acquisition time that is significantly shorter than full-volume 4D-CBCT. Matching DRTs were obtained from the planning 4D-CT for each phase. 2D-2D registrations were performed to obtain the phase-varying offset. Conclusion: A new imaging technique has been introduced for image-guided lung treatment. With this approach, images are acquired faster and the appearance of the tumor is significantly enhanced by eliminating many extraneous structures. Conflict of Interest: This work is supported by Siemens.
The x‐ray fovea (U.S. patents pending) is a device for reducing x‐ray dose to patients and operators during x‐ray fluoroscopy. It consists of a semitransparent collimator with an open, circular, central hole. The fovea collimator is placed at the exit of the x‐ray tube, and the attenuation of the peripheral x‐ray beam reduces x‐ray exposure to patients and operators. The shadow caused by the x‐ray fovea can be compensated using real‐time image processing hardware. Accurate compensation is demonstrated for both linearly and logarithmically acquired images using a model that accounts for beam hardening in the fovea collimator. The central fovea region has improved image quality due to reduced scatter and veiling glare from the periphery. From beam‐stop measurements, a 40% reduction in scatter plus veiling glare is measured using the fovea. A contrast improvement ratio of 1.5 is measured throughout the central region. In the compensated periphery, noise is increased by a factor of 1.66 because fewer photons are detected, but a small amount of temporal filtering compensates this degradation. The Roentgen area product (RAP) exposure to patients is reduced by ≊70%, while scattered exposure to operators is reduced by ≊60%.
As-built reconstruction of existing industrial facilities such as power plants still involves much interactive work even though photogrammetry techniques are greatly used in this process. The majority of existing systems are based on 3D-point reconstruction. In general, these systems only use epipolar geometry to help the user. CyliCon is a software package for 3D reconstruction of industrial pipelines. This software enables its users to work with hundreds of images. It uses geometric features, such as occluding edges and vanishing points, and image processing methods, such as multi-resolution edge refinement, in order to provide semiautomatic user-friendly software. The software has been successfully tested on hundreds of indoor and outdoor industrial images.
In this paper we study the performance of the maximization of mutual information (MMI) based registration algorithm using a select number of medical image data sets. Where appropriate, we present a comparison of the MMI based algorithm with surface based and centerline (medial axes) based registration algorithms. In the paper we first show that the MMI based registration algorithm can accurately estimate the registration parameters using real patient image data sets. We then present a data set where the MMI based estimates are not as accurate. For this image data set, the MMI based algorithm requires sufficiently close initialization to the true parameters. Also, we show that for such image data sets feature based algorithms, such as, surface based and the centerline based methods, can accurately estimate the registration parameters. The paper also discusses the effects of the optimization method (Powell's versus Simplex) on the quality and computation time of the MMI based and correlation based registration algorithms.
Purpose: To quantitatively and qualitatively assess improvement in megavoltage cone beam CT (MVCBCT) image quality afforded by a 4.2 MV imaging beam line (IBL) with a carbon electron target and new pixelated ultrafast ceramic scintillator (UFC) detector, relative to the 6 MV treatment beam line (TBL). Detector blur is reduced with the UFC since light generated in a pixel is less likely to escape laterally. Methods: A prototype IBL+UFC system was installed on a Siemens ONCOR linear accelerator equipped with the MVision™ IGRT system. A UFC strip consisting of four tiles and measuring ∼40 cm by ∼10 cm was installed on the flat panel imager, with the long dimension in the cross-plane direction. Phantom images were acquired at doses from 2–60 cGy with the TBL, IBL with conventional scintillator, and IBL+UFC. Several head and neck, thoracic, and pelvic patients were imaged with the three systems at doses from 2–15 cGy. Results: Phantom images indicate that the IBL+UFC images have lower noise and higher contrast than the IBL and TBL images. The contrast-to-noise ratio (CNR) for the IBL was 1.5–2 times higher than for the TBL, and the IBL+UFC CNR was 1.5–2 times better than for the IBL at all doses. CNR saturated near 30 cGy for the IBL+UFC and at 60 cGy or greater for the IBL and TBL. IBL+UFC patient images showed improved soft tissue contrast at all doses and sites examined. Conclusions: The IBL+UFC combination increases the CNR by up to a factor of four relative to the TBL with the conventional scintillator, and a factor of two relative to the IBL with the conventional scintillator. Image noise and soft tissue contrast in head and neck, thoracic, and pelvic patients improved dramatically in the IBL+UFC images relative to the TBL images. Sponsored partially by Siemens Oncology Care Systems.
Purpose: Large patient anatomies and limited imaging field‐of‐view (FOV) lead to truncation of CT projections. Truncation introduces serious artifacts into reconstructed images, including central cupping and bright external rings. FOV may be increased using laterally offset detectors, but this requires advanced imaging hardware and full angular scanning. When linacs equipped with mini‐MLCs are used for megavoltage cone beam CT imaging, truncation is inevitable. We propose a novel method to complete truncated projections based on the observation that the thickness of the patient at all angles may be estimated along the projection rays by calculating water‐equivalent thicknesses (WET). These values are not at all affected by truncation and constitute valuable auxiliary information. Methods: We parameterize a pair of points along each ray that intersects the unknown object boundary. These points are separated by the measured WET value (obtained from projections that have been corrected for scatter and beam‐hardening). We assume, for all large body parts, that the patient outline may be roughly approximated as an ellipse. We use a deterministic optimization algorithm to simultaneously estimate the point positions and ellipse parameters by minimizing the distance between point sets and the ellipse boundary. The ellipse obtained as solution is used to complete the truncated projections before reconstruction. We apply the algorithm to a severely truncated CT dataset of a typical abdomen. Results: The RMS error between non‐truncated and truncated reconstructions is 176%. The correction algorithm reduces this error to 1.0%. Conclusion: Even thought the algorithm assumes an elliptical patient cross‐section, truly impressive increases in quantitative image quality are observed. The presence of pelvic bone in the image does not appreciably bias the ellipse position even though it produces incorrect thickness estimates for some rays. The algorithm incurs low computational burden and is suitable for on‐line clinical work flows. Supported by Siemens.
