Sub-millimeter Accuracy Research Articles

Acoustic Signature: A focused ultrasound guidance technique with sub-millimeter accuracy

Accurate targeting is paramount for focused ultrasound (FUS) procedures. Here we describe a method for targeting FUS dubbed acoustic signature, which uses acoustic feedback from a target’s unique reflection patterns to guide the transducer to a previously defined orientation. We demonstrate convergence to the desired target in a 3 degree of freedom (DOF) water tank scenario and a 5 DOF robotic arm scenario. We also tested the acoustic signature technique for targeting with a phantom skull on a 3 DOF motor stage and a 5 DOF manual stereotactic frame designed for transcranial FUS procedures in an MRI environment. In both the water tank and the robotic arm convergence tests, the method converges to the target. The convergence was sufficiently smooth for automated gradient descent. In the 3-DOF test, the net targeting error was 0.30 ± 0.27 mm (n = 10). In the stereotactic frame 5-DOF test, the targeting error was 2.11 ± 0.99 mm and 3.2 ± 2.2° (n = 10) summed rotational error. The accuracy of this technique is limited by the positioning apparatus. This method may enable precise and repeatable FUS targeting for therapies which require multiple FUS sessions at the same target.

Comparison of image registration methods for combining laparoscopic video and spectral image data

Laparoscopic procedures can be assisted by intraoperative modalities, such as quantitative perfusion imaging based on fluorescence or hyperspectral data. If these modalities are not available at video frame rate, fast image registration is needed for the visualization in augmented reality. Three feature-based algorithms and one pre-trained deep homography neural network (DH-NN) were tested for single and multi-homography estimation. Fine-tuning was used to bridge the domain gap of the DH-NN for non-rigid registration of laparoscopic images. The methods were validated on two datasets: an open-source record of 750 manually annotated laparoscopic images, presented in this work, and in-vivo data from a novel laparoscopic hyperspectral imaging system. All feature-based single homography methods outperformed the fine-tuned DH-NN in terms of reprojection error, Structural Similarity Index Measure, and processing time. The feature detector and descriptor ORB1000 enabled video-rate registration of laparoscopic images on standard hardware with submillimeter accuracy.

Lunar ephemeris at sub microarcsecond accuracy (LESMA) leads to sub-millimeter positional accuracy of the moon

The most accurate LLR (lunar laser ranging) initiative, named APOLLO (apache point observatory lunar laser-ranging operation) demonstrated millimeter-range positional accuracy in 2009, thus improving LLR by one order-of-magnitude. Since, LLR is a foundational technique in studying gravity, Murphy (principal investigator of APOLLO) stated in 2009, that with this millimeter-range accuracy, the simulation model has been found to be the limiting-factor in extracting the theoretical science results, and hence, we should: (1) develop the science case and expand our ability to model LLR for a new regime of high precision, (2) develop the theoretical tools for honing the science case for submillimeter LLR, and (3) explore which model/code is worth putting our efforts into. (4) Since millimeter-quality data are a recent development, the model effort lags. (5) Finally, we will code-in new physics so that we may simulate sensitivities. In connection with simulation model/code, Murphy stated in 2013, that among the four available LLR simulation models: JPL (jet propulsion laboratory), CfA (the Harvard-Smithsonian center for astrophysics), LU (leibniz University, Hannover, Germany), and IMCCE (Institut de Mecanique celeste et de calcul des Ephemerides, France), the JPL model currently produces weighted RMS (root-mean-square) residuals at ∼18 mm, which is about half of the other models; so, clearly a gap exists from millimeter ranging-precision of APOLLO. Hence, the CfA, LU, and IMCCE are engaged, since 2013, in a stepwise comparative streamlining effort to identify the model-differences, errors, and shortcomings. All the four available LLR simulation models can be classified as GR (general relativity)-astronomers model; they are basically similar. Professor Douglas Currie of the University of Maryland, College Park, NASA Lunar Science Institute, stated in a Conference presentation, in 2012, that Ground stations, that is, the lunar observatories, have improved by a factor of 200, but the agreement between observations and fitted theory has plateaued at ∼2 cm over the past two decades. However, no substantial progress on improving the fit has been reported in the published literature, till date. Based on about a quarter-century of experience in doing high-precision numerical simulation of celestial orbits, the authors have developed LESMA (lunar Ephemeris at sub Microarcsecond accuracy) utilizing the methodology of evolved general relativity (EGR) that has incorporated the following two concepts: (1) Relativistic time for integration and (2) methodology of conservation of magnitude of the angular momentum, MΦ , for Φ-rotation (in addition to the θ-rotation that leads to the rosetting ellipse) of the orbital plane. Incorporation of the two above-mentioned concepts has led to three orders-of-magnitude accuracy-improvement of the computed (1) precession (compared to JPL's DE405) of Lunar orbit, as verified using three independent methods and (2) radial position (compared to JPL's DE430/431) of the Moon. LESMA will enable scientists to make efficient use of research-funds from NASA, etc., for production of new science results from APOLLO. LESMA will also be useful for getting better science results (than Folkner reported {in 2014} submeter accurate Position of the Moon) from the GRAIL (gravity recovery and Interior laboratory) mission (costing 500 million USD), by spending a little more for revisiting the computations, utilizing LESMA data.

Relativistic location algorithm in curved spacetime

In this article, we describe and numerically implement a method for relativistic location in slightly curved but otherwise generic spacetimes. For terrestrial positioning in the context of Global Navigation Satellite Systems, our algorithm incorporates gravitational as well as tropospheric and ionospheric effects modeled by the Gordon metric. The algorithm is implemented in the squirrel.jl code, which employs a quasi-Newton Broyden algorithm in conjunction with automatic differentiation of numerical geodesics. Our work provides a practical solution to the relativistic location problem in a generic spacetime and consolidates relativistic and atmospheric effects in a single framework. Though optimization is not our primary focus, our implementation is already fast enough for practical use, establishing a position from five emission points in $<1\text{ }\text{ }\mathrm{s}$ on a desktop computer for reasonably simple spacetime geometries. In vacuum, our implementation can achieve submillimeter accuracy considering the Kerr metric with terrestrial parameters and submeter accuracy including tropospheric and ionospheric effects.

Measuring Railway Track Irregularities at High Accuracy and Efficiency Based on GNSS/INS/TS Integration

Measuring the railway track irregularities (abbr. Irr) with sub-millimeter accuracy is vital for the travel safety of high-speed trains. In comparison to the traditional total station (TS)-based track inspection trolley (TIT) for track irregularity examination, the integration of global navigation satellite system (GNSS) and inertial navigation system (INS) is more time-efficient but less accurate in longwave irregularity detection. To achieve high accuracy and high survey efficiency, a GNSS/INS/TS integrated irregularity measuring (GIT-Irr) system is proposed based on the attitude-inferred irregularity measuring method utilizing TS constraint. In addition, to reduce the conversion error of TS coordinates from local engineering frame to local geo-frame, TS-aided calibration in the deviation domain is recommended in the GIT-Irr system. A field test was carried out near the Baotou East railway station in China in order to validate the performance of the proposed system. Four sets of independent measurements are collected together with an external reference from a TS-based TIT. The results demonstrate that the proposed system increases the surveying ability of shortwave and longwave alignment irregularities by about 5% and 75%, respectively. Under a 3-sigma confidence level, shortwave alignment irregularities are 0.50 mm and 0.57 mm in horizontal and vertical directions, respectively, while those of the longwaves are 1.78 mm and 2.06 mm. Therefore, the proposed GIT-Irr system provides great potential for highly efficient sub-millimeter track irregularity measurement.

Validity and reliability of NOTCH® inertial sensors for measuring elbow joint angle during tennis forehand at different sampling frequencies

Portable and low-cost motion capture systems are gaining importance for biomechanical analysis. The aim was to determine the concurrent validity and reliability of the NOTCH® inertial sensors to measure the elbow angle during tennis forehand at different sampling frequencies (100, 250 and 500 Hz), using an optical capture system with sub-millimetre accuracy as a reference. 15 competitive players performed forehands wearing NOTCH and an upper body marker-set and the signals from both systems were adjusted and synchronized. The error magnitude was tolerable (5-10°) for all joint-axis and sampling frequencies, increasing significantly at 100 Hz for the flexion–extension and pronation-supination angles (p = 0.002 and 0.023; Cohen d > 0.8). Concordance correlation coefficient was very large (0.7–0.9) in all cases. The within-subject error variation between the test–retest did not show significant differences (p > 0.05). NOTCH® is a valid, reliable and portable alternative to measure elbow angles during tennis forehand.

A deep learning and Monte Carlo based framework for bioluminescence imaging center of mass-guided glioblastoma targeting

Objective. Bioluminescence imaging (BLI) is a valuable tool for non-invasive monitoring of glioblastoma multiforme (GBM) tumor-bearing small animals without incurring x-ray radiation burden. However, the use of this imaging modality is limited due to photon scattering and lack of spatial information. Attempts at reconstructing bioluminescence tomography (BLT) using mathematical models of light propagation show limited progress. Approach. This paper employed a different approach by using a deep convolutional neural network (CNN) to predict the tumor’s center of mass (CoM). Transfer-learning with a sizeable artificial database is employed to facilitate the training process for, the much smaller, target database including Monte Carlo (MC) simulations of real orthotopic glioblastoma models. Predicted CoM was then used to estimate a BLI-based planning target volume (bPTV), by using the CoM as the center of a sphere, encompassing the tumor. The volume of the encompassing target sphere was estimated based on the total number of photons reaching the skin surface. Main results. Results show sub-millimeter accuracy for CoM prediction with a median error of 0.59 mm. The proposed method also provides promising performance for BLI-based tumor targeting with on average 94% of the tumor inside the bPTV while keeping the average healthy tissue coverage below 10%. Significance. This work introduced a framework for developing and using a CNN for targeted radiation studies for GBM based on BLI. The framework will enable biologists to use BLI as their main image-guidance tool to target GBM tumors in rat models, avoiding delivery of high x-ray imaging dose to the animals.

Experimental Research of the Structure Condition Using Geodetic Methods and Crackmeter

The article presents an approach to monitoring the structure’s condition with two measurement methods: the SHM-X crackmeter and the classic geodetic method of determining displacements, supplemented with additional information on the condition of the external environment obtained from thermal images. The study aimed to propose an approach combining geodetic and non-geodetic methods of assessing the condition of a structure and its effectiveness in practical application. The research facility is a public utility building of the Bydgoszcz University of Technology with a reinforced concrete structure. Objects of this type require periodic tests of their constancy. Interpreting the test results and identifying possible dangerous states that may indicate the risk of a construction failure is extremely important. The results presented in the article are an extension of the previous ones, in which several factors that could have a destructive effect on the structure were excluded. Observation of the object showed that only the reinforced construction plate is deformed. The only factor influencing the change in structure geometry is thermal changes. As part of the tests in places where cracks were noticed, the SHM-X crackmeter was used to measure the cracks’ opening. In the geodetic research, measurements of the measurement and control network displacement were carried out, in which the TDRA6000 laser station measurement technology was used. The control points were also placed in places where the width of the cracks was directly observed. The proposed approach, with the applied calculation scheme and supplementing the information with the temperature measurement with thermal images, showed the submillimeter accuracy of the determined 3D displacements of the controlled points. Additionally, the parallel application of these methods gives a complete picture of changes in the structure elements, in which signs of destruction appear under the influence of stress.

Discrete-Time Optimal Control of Miniature Helical Swimmers in Horizontal Plane

Microswimmer and miniswimmer toward precision-targeted medicine have attracted extensive attention recently. We have developed an autonomous manipulation approach for magnetic-driven helical miniswimmer at low Reynolds number in the horizontal plane (<inline-formula> <tex-math notation="LaTeX">$H$ </tex-math></inline-formula>-plane). Different from our previous work which just makes the barycenter of miniswimmer on the reference path as well as takes the swimming direction not into consideration in planar path following, our control policy in this article can make the miniswimmer to follow the reference path and, at the same time, its swimming direction is also along the reference path. A robust tracking method is employed to locate the helical miniswimmer in real time. Due to different external disturbances, an angle compensating model in the global coordinate frame is developed by radial basis function (RBF) networks trained by backpropagation algorithms, which is used to express the swimming model of the helical miniswimmer facing the gravity and lateral disturbances. A discrete-time optimal controller is formulated based on the linear-quadratic feedback control. Simulations and experiments are conducted to quantitatively validate the autonomous manipulation, and the results show the control performance with submillimeter accuracy in the <inline-formula> <tex-math notation="LaTeX">$H$ </tex-math></inline-formula>-plane. <i>Note to Practitioners:</i> This article is motivated by the potential application of precision-targeted medicine using magnetic-driven microswimmer/miniswimmer. The formulated controller employs the error model in the horizontal plane to design the control law. Simulations and experiments validate the effectiveness of the proposed discrete-time optimal control scheme using magnetic-driven miniswimmers.

Preparation of photosensitive SiO2/SiC ceramic slurry with high solid content for stereolithography

Stereolithography is a popular three-dimensional (3D) printing technology, which is widely used for manufacturing ceramic components owing to its high efficiency and precision. However, it is a big challenge to prepare SiC ceramic slurry with high solid content for stereolithography due to the strong light absorption and high refractive index of dark SiC powders. Here, we propose a novel strategy to develop photosensitive SiO2/SiC ceramic slurry with high solid content of 50–65 vol% by adding spherical silica with low light absorbance and applying a stacking flow model to improve the solid content of the slurry. The as-prepared slurry exhibits excellent stereolithography properties with a dynamic viscosity lower than 20 Pa s and curing thickness more than 120 μm. Therefore, it can be successfully applied for stereolithography-based additive manufacturing of SiC green bodies with large size (100 mm), sub-millimeter accuracy (0.2 mm), and complex structure. The stacking flow model also shows immense potential for the stereolithography of other dark-color ceramics with high solid content.

Design and development of a sensor assembly based on 1-D photodiode arrays for distance offset measurements

This paper proposes a sensing device to accurately measure lateral offsets for applications in calibration and alignment activities. The proposed sensor assembly is primarily designed and developed for use in volumetric calibration procedures of custody tanks in the energy industry. In order to achieve sub-millimeter accuracy levels for offset measurements, the sensor device utilizes multiple 1-D photodiode arrays in a specific configuration that enables improvement upon currently utilized procedures and devices. The presented design is also modular such that it can be easily modified to address applications involving robotic vehicles, and other systems that require tracking, measurement, alignment and calibration activities. The experimental evaluation of the proposed sensor assembly showed that it can measure distance offset measurements with a resolution of 125μm and a measured accuracy within 730μm. The promising results show that the proposed sensing device can be integrated into existing procedures that require high accuracy measurements.

High-speed multi-camera 3D DIC measurement of the deformation of cassette structure with large shaking table

Vision-based deformation measurement methods have been proposed to overcome the disadvantages of laborious, time-consuming measurement data that are difficult to accurately obtain using contact sensors in three directions. However, such methods remain a challenge for large-scale applications owing to difficulties in full-field imaging, stereo calibration, and measurement resolution, among other factors. To the best of our knowledge, there are no reports on the three-dimensional (3D) full-field measurement of a large shaking table structure with multiple cameras. In this paper, a multi-camera 3D DIC method using four high-speed cameras to measure the multi-layer structure model in a shaking table is proposed. To calibrate the stereo cameras with a large field of view, a speckle-based calibration method and a coordinate system unification method are proposed. The implementation of the proposed method is illustrated with a shaking table test for a 1/8 scaled 15-story high-rise cassette concrete structure with several different intensity levels of ground motions. The results of displacement, velocity, and acceleration of the structure measured using the proposed method correspond well with those of a laser displacement meter and accelerometer. The experiment results show that this method can guarantee full-field measurements in the range of tens of square meters and realize a sub-millimeter displacement measurement accuracy. They also indicate that the proposed multi-camera 3D DIC method has high potential for understanding the failure mechanism of structures and improving the structural design.

Extended Intraoperative Longitudinal 3-Dimensional Cone Beam Computed Tomography Imaging With a Continuous Multi-Turn Reverse Helical Scan.

Cone beam computed tomography (CBCT) imaging is becoming an indispensable intraoperative tool; however, the current field of view prevents visualization of long anatomical sites, limiting clinical utility. Here, we demonstrate the longitudinal extension of the intraoperative CBCT field of view using a multi-turn reverse helical scan and assess potential clinical utility in interventional procedures. A fixed-room robotic CBCT imaging system, with additional real-time control, was used to implement a multi-turn reverse helical scan. The scan consists of C-arm rotation, through a series of clockwise and anticlockwise rotations, combined with simultaneous programmed table translation. The motion properties and geometric accuracy of the multi-turn reverse helical imaging trajectory were examined using a simple geometric phantom. To assess potential clinical utility, a pedicle screw posterior fixation procedure in the thoracic spine from T1 to T12 was performed on an ovine cadaver. The multi-turn reverse helical scan was used to provide postoperative assessment of the screw insertion via cortical breach grading and mean screw angle error measurements (axial and sagittal) from 2 observers. For all screw angle measurements, the intraclass correlation coefficient was calculated to determine observer reliability. The multi-turn reverse helical scans took 100 seconds to complete and increased the longitudinal coverage by 370% from 17 cm to 80 cm. Geometric accuracy was examined by comparing the measured to actual dimensions (0.2 ± 0.1 mm) and angles (0.2 ± 0.1 degrees) of a simple geometric phantom, indicating that the multi-turn reverse helical scan provided submillimeter and degree accuracy with no distortion. During the pedicle screw procedure in an ovine cadaver, the multi-turn reverse helical scan identified 4 cortical breaches, confirmed via the postoperative CT scan. Directly comparing the screw insertion angles (n = 22) measured in the postoperative multi-turn reverse helical and CT scans revealed an average difference of 3.3 ± 2.6 degrees in axial angle and 1.9 ± 1.5 degrees in the sagittal angle from 2 expert observers. The intraclass correlation coefficient was above 0.900 for all measurements (axial and sagittal) across all scan types (conventional CT, multi-turn reverse helical, and conventional CBCT), indicating excellent reliability between observers. Extended longitudinal field-of-view intraoperative 3-dimensional imaging with a multi-turn reverse helical scan is feasible on a clinical robotic CBCT imaging system, enabling long anatomical sites to be visualized in a single image, including in the presence of metal hardware.

Frameless neuronavigation with computer vision and real-time tracking for bedside external ventricular drain placement: a cadaveric study.

A major obstacle to improving bedside neurosurgical procedure safety and accuracy with image guidance technologies is the lack of a rapidly deployable, real-time registration and tracking system for a moving patient. This deficiency explains the persistence of freehand placement of external ventricular drains, which has an inherent risk of inaccurate positioning, multiple passes, tract hemorrhage, and injury to adjacent brain parenchyma. Here, the authors introduce and validate a novel image registration and real-time tracking system for frameless stereotactic neuronavigation and catheter placement in the nonimmobilized patient. Computer vision technology was used to develop an algorithm that performed near-continuous, automatic, and marker-less image registration. The program fuses a subject's preprocedure CT scans to live 3D camera images (Snap-Surface), and patient movement is incorporated by artificial intelligence-driven recalibration (Real-Track). The surface registration error (SRE) and target registration error (TRE) were calculated for 5 cadaveric heads that underwent serial movements (fast and slow velocity roll, pitch, and yaw motions) and several test conditions, such as surgical draping with limited anatomical exposure and differential subject lighting. Six catheters were placed in each cadaveric head (30 total placements) with a simulated sterile technique. Postprocedure CT scans allowed comparison of planned and actual catheter positions for user error calculation. Registration was successful for all 5 cadaveric specimens, with an overall mean (± standard deviation) SRE of 0.429 ± 0.108 mm for the catheter placements. Accuracy of TRE was maintained under 1.2 mm throughout specimen movements of low and high velocities of roll, pitch, and yaw, with the slowest recalibration time of 0.23 seconds. There were no statistically significant differences in SRE when the specimens were draped or fully undraped (p = 0.336). Performing registration in a bright versus a dimly lit environment had no statistically significant effect on SRE (p = 0.742 and 0.859, respectively). For the catheter placements, mean TRE was 0.862 ± 0.322 mm and mean user error (difference between target and actual catheter tip) was 1.674 ± 1.195 mm. This computer vision-based registration system provided real-time tracking of cadaveric heads with a recalibration time of less than one-quarter of a second with submillimetric accuracy and enabled catheter placements with millimetric accuracy. Using this approach to guide bedside ventriculostomy could reduce complications, improve safety, and be extrapolated to other frameless stereotactic applications in awake, nonimmobilized patients.

Parameter Estimation and Tracking of an Expandable Cable-Driven Parallel Mechanism for Minimally Invasive Interventions

Conventional catheter-based procedures suffer from inaccurate device steering, and navigation limitations as the long and flexible devices are manipulated remotely from outside the patient body. A recently developed approach aims to overcome these limitations by using an expandable cable-driven parallel mechanism for local device manipulation and tracking relative to the anatomy. However, as the frame of this system is flexible and may be constrained by unknown anatomical constraints inside the patient body, accurate system identification is a key and critical step which would ultimately determine the accuracy in device steering and tracking. In this paper, we present an optimization approach to identify the parameters of an expandable cable-driven parallel mechanism deployed in an unknown environment, merely based on measurements of the cable length variations that would be available outside the patient body. We have verified and validated the proposed approach in terms of accuracy in the prediction of the frame shape and cable anchor locations, as well as its tracking accuracy. Both simulations and experimental studies have been performed, and we have demonstrated that the proposed method allows for rapid frame shape and anchor position estimation with unknown anatomical constraints and permits localized device tracking and steering with submillimeter accuracy.

At-Home Pupillometry using Smartphone Facial Identification Cameras.

With recent developments in medical and psychiatric research surrounding pupillary response, cheap and accessible pupillometers could enable medical benefits from early neurological disease detection to measurements of cognitive load. In this paper, we introduce a novel smartphone-based pupillometer to allow for future development in clinical research surrounding at-home pupil measurements. Our solution utilizes a NIR front-facing camera for facial recognition paired with the RGB selfie camera to perform tracking of absolute pupil dilation with sub-millimeter accuracy. In comparison to a gold standard pupillometer during a pupillary light reflex test, the smartphone-based system achieves a median MAE of 0.27mm for absolute pupil dilation tracking and a median error of 3.52% for pupil dilation change tracking. Additionally, we remotely deployed the system to older adults as part of a usability study that demonstrates promise for future smartphone deployments to remotely collect data in older, inexperienced adult users operating the system themselves.

Documenting Cultural Heritage in very hostile fruition contexts: the synoptic visualization of Giottesque frescoes by Multispectral and 3D Close-range Imaging

Abstract The paper reports the results obtained from the digital survey of Giottesque frescoes in the Archiepiscopal Palace of Milan. Multispectral Imaging and Image Based Modelling techniques have allowed the study, the high-resolution documentation and the virtual reassembly of the fragmented frescoes: the integration of spectral data with RGB ortho-images and 3D models has supported the interpretation of the pictorial cycle by virtually relocating the fresco fragments and evaluating their consistency from a technical, iconographic and art-historical point of view. The experimental results show that the proposed approach has great potential for the documentation of wall paintings preserved in highly hostile contexts of fruition. Due to its versatility and portability in the field, its sub-millimetric accuracy, and the different outputs that can be generated, Close-Range Digital Photogrammetry has proven to be the ideal tool in case of scattered and difficult-to-access data.

Magnetic Resonance-Guided Ultrasound Hyperthermia for Prostate Cancer Radiotherapy: An Immobilization Device Embedding the Ultrasound Applicator

Aim: To test 3D-printed immobilization devices for future use in magnetic resonance-guided focused ultrasound hyperthermia. Material &amp; methods: Using a surface scanner, patient-specific pelvic immobilization devices were 3D printed. The setup reproducibility was measured both on linear accelerator (LINAC) and magnetic resonance. An ultrasound imaging probe was used to acquire reference images and later to acquire images once attached into the embedded holder. Results: Prepositioning accuracy was tested at LINAC using an optical surface monitoring and MRI and showed submillimeter accuracy and small angular rotations. Agreement was high between the ultrasound reference images versus the immobilized probe. Conclusion: Reported results are considered as a promising step toward a fast and precise positioning of patients and an easier integration of radiotherapy and magnetic resonance-guided focused ultrasound hyperthermia.

Reducing 4DCBCT imaging dose and time: exploring the limits of adaptive acquisition and motion compensated reconstruction

This study investigates the dose and time limits of adaptive 4DCBCT acquisitions (adaptive-acquisition) compared with current conventional 4DCBCT acquisition (conventional-acquisition). We investigate adaptive-acquisitions as low as 60 projections (∼25 s scan, 6 projections per respiratory phase) in conjunction with emerging image reconstruction methods. 4DCBCT images from 20 patients recruited into the adaptive CT acquisition for personalized thoracic imaging clinical study (NCT04070586) were resampled to simulate faster and lower imaging dose acquisitions. All acquisitions were reconstructed using Feldkamp–Davis–Kress (FDK), McKinnon–Bates (MKB), motion compensated FDK (MCFDK), motion compensated MKB (MCMKB) and simultaneous motion estimation and image reconstruction (SMEIR) algorithms. All reconstructions were compared against conventional-acquisition 4DFDK-reconstruction using Structural SIMilarity Index (SSIM), signal-to-noise ratio (SNR), contrast-to-noise-ratio (CNR), tissue interface sharpness diaphragm (TIS-D), tissue interface sharpness tumor (TIS-T) and center of mass trajectory (COMT) for difference in diaphragm and tumor motion. All reconstruction methods using 110-projection adaptive-acquisition (11 projections per respiratory phase) had a SSIM of greater than 0.92 relative to conventional-acquisition 4DFDK-reconstruction. Relative to conventional-acquisition 4DFDK-reconstruction, 110-projection adaptive-acquisition MCFDK-reconstructions images had 60% higher SNR, 10% higher CNR, 30% higher TIS-T and 45% higher TIS-D on average. The 110-projection adaptive-acquisition SMEIR-reconstruction images had 123% higher SNR, 90% higher CNR, 96% higher TIS-T and 60% higher TIS-D on average. The difference in diaphragm and tumor motion compared to conventional-acquisition 4DFDK-reconstruction was within submillimeter accuracy for all acquisition reconstruction methods. Adaptive-acquisitions resulted in faster scans with lower imaging dose and equivalent or improved image quality compared to conventional-acquisition. Adaptive-acquisition with motion compensated-reconstruction enabled scans with as low as 110 projections to deliver acceptable image quality. This translates into 92% lower imaging dose and 80% less scan time than conventional-acquisition.

181 Frameless Neuronavigation With Computer Vision and Real Time Tracking for Bedside External Ventricular Drain Placement: A Cadaveric Study

INTRODUCTION: A major obstacle to improving ventriculostomy safety and accuracy with image guidance technologies is the lack of a rapidly deployable, real-time registration and tracking system for a moving patient. METHODS: Computer vision technology was used to develop an algorithm for near continuous, automatic, and marker-less image registration. The program fuses a subject’s pre-procedure CT scan to live 3D camera images (Snap-Surface), and movement is incorporated by artificial intelligence driven recalibration (Real-Track). Surface registration error (SRE) and target registration error (TRE) were calculated for five cadaver heads that underwent serial movements (fast and slow velocity roll, pitch, and yaw motions), and several test conditions. Six catheters were placed in each cadaver (30 placements). Post-procedure CT scans allowed comparison of planned and actual catheter position for user error calculation. RESULTS: Registration was successful for all five cadaveric specimens (average SRE 0.429 mm, ±0.108). Accuracy of TRE was under 1.2mm throughout specimen movements of low and high velocities of roll, pitch, and yaw movements, with the slowest recalibration time of 0.23 seconds. There were no statistically significant differences in SRE when the specimens were draped or fully undraped (p-value = 0.336). Addition of bright light or performance under dim light did not significantly affect SRE (p-value = 0.742 and 0.859, respectively). Average TRE for catheter placements was 0.862 mm (±0.322), and average user error for the was 1.674 mm (±1.195). CONCLUSION: This computer vision-based registration system provides real-time tracking of cadaveric heads with recalibration time of less than one-quarter of a second with sub-millimetric accuracy, and enabled catheter placements with millimetric accuracy. Using this approach to guide bedside ventriculostomy could reduce complications, improve safety, and could be extrapolated to other frameless stereotactic applications in awake, non-immobilized patients.