Joyoni Dey

Surface normal overlap: a computer-aided detection algorithm with application to colonic polyps and lung nodules in helical CT

We developed a novel computer-aided detection (CAD) algorithm called the surface normal overlap method that we applied to colonic polyp detection and lung nodule detection in helical computed tomography (CT) images. We demonstrate some of the theoretical aspects of this algorithm using a statistical shape model. The algorithm was then optimized on simulated CT data and evaluated using a per-lesion cross-validation on 8 CT colonography datasets and on 8 chest CT datasets. It is able to achieve 100% sensitivity for colonic polyps 10 mm and larger at 7.0 false positives (FPs)/dataset and 90% sensitivity for solid lung nodules 6 mm and larger at 5.6 FP/dataset.

A flexible multicamera visual‐tracking system for detecting and correcting motion‐induced artifacts in cardiac SPECT slices

Patient motion is inevitable in SPECT and PET due to the lengthy period of time patients are imaged. The authors hypothesized that the use of external-tracking devices which provide additional information on patient motion independent of SPECT data could be employed to provide a more robust correction than obtainable from data-driven methods. Therefore, the authors investigated the Vicon MX visual-tracking system (VTS) which utilizes near-infrared (NIR) cameras to stereo-image small retroreflective markers on stretchy bands wrapped about the chest and abdomen of patients during cardiac SPECT. The chest markers are used to provide an estimate of the rigid-body (RB) motion of the heart. The abdomen markers are used to provide a signal used to bin list-mode acquisitions as part of correction of respiratory motion of the heart. The system is flexible in that the layout of the cameras can be designed to facilitate marker viewing. The system also automatically adapts marker tracking to employ all of the cameras visualizing a marker at any instant, with visualization by any two being sufficient for stereo-tracking. Herein the ability of this VTS to track motion with submillimeter and subdegree accuracy is established through studies comparing the motion of Tc-99m containing markers as assessed via stereo-tracking and from SPECT reconstructions. The temporal synchronization between motion-tracking data and timing marks embedded in list-mode SPECT acquisitions is shown to agree within 100 ms. In addition, motion artifacts were considerably reduced in reconstructed SPECT slices of an anthropomorphic phantom by employing within iterative reconstruction the motion-tracking information from markers attached to the phantom. The authors assessed the number and placement of NIR cameras required for robust motion tracking of markers during clinical imaging in 77 SPECT patients. They determined that they were able to track without loss during the entire period of SPECT and transmission imaging at least three of the four markers on the chest and one on the abdomen bands 94% and 92% of the time, respectively. The ability of the VTS to correct motion clinically is illustrated for ten patients who volunteered to undergo repeat-rest imaging with the original-rest SPECT study serving as the standard against which to compare the success of correction. Comparison of short-axis slices shows that VTS-based motion correction provides better agreement with the original-rest-imaging slices than either no correction or the vendor-supplied software for motion correction on, our SPECT system. Comparison of polar maps shows that VTS-based motion-correction results in less numerical difference on average in the segments of the polar maps between the original-rest study and the second-rest study than the other two strategies. The difference was statistically significant for the comparison between VTS-based and clinical vendor-supplied software correction. Taken together, these findings suggest that VTS-based motion correction is superior to either no-motion correction or the vendor-supplied software the authors investigated in clinical practice.

Estimation of Rigid-Body and Respiratory Motion of the Heart From Marker-Tracking Data for SPECT Motion Correction

Motion of patients undergoing cardiac SPECT perfusion imaging causes artifacts in the acquired images which may lead to difficulty in interpretation. Our work investigates a technique of obtaining patient motion estimates from retro-reflective markers on stretchy bands wrapped around the chest and abdomen of patients being imaged clinically. Motion signals obtained from the markers consist of at least two components, body motion (BM) and periodic motion (PM) due to respiration. We present a method for separating these components from the motion-tracking data of each marker, and then report a method for combining the BM estimated from chest markers to estimate the 6-degree-of-freedom (6-DOF) rigid-body motion (RBM) of the heart. Motion studies of volunteers and patients are used to evaluate the methods. Illustrative examples of the motion of the heart due to patient body movement and respiration (upward creep) are presented and compared to estimates of the motion of the heart obtained directly from SPECT data. Our motion-tracking method is seen to give reasonable agreement with the motion-estimates from the SPECT data while being considerably less noisy.

4D reconstruction for low-dose cardiac gated SPECT

PURPOSE Due to the combination of high-frequency use and relatively high diagnostic radiation dose (>9 mSv for one scan), there is a need to lower the radiation dose used in myocardial perfusion imaging (MPI) studies in cardiac gated single photon emission computed tomography (GSPECT) in order to reduce its population based cancer risk. The aim of this study is to assess quantitatively the potential utility of advanced 4D reconstruction for GSPECT for significantly lowered imaging dose. METHODS For quantitative evaluation, Monte Carlo simulation with the 4D NURBS-based cardiac-torso (NCAT) phantom is used for GSPECT imaging at half and quarter count levels in the projections emulating lower injected activity (dose) levels. Both 4D and 3D reconstruction methods are applied at these lowered dose levels, and compared with standard clinical spatiotemporal reconstruction (ST121) at full dose using a number of metrics on the reconstructed images: (1) overall reconstruction accuracy of the myocardium, (2) regional bias-variance analysis of the left ventricle (LV) wall, (3) uniformity of the LV wall, (4) accuracy of the time activity curve (TAC) of the LV wall, and (5) detectability of perfusion defects using channelized Hotelling observer. As a preliminary demonstration, two sets of patient data acquired in list-mode are used to illustrate the conservation of both image quality and LV ejection fraction (LVEF) by 4D reconstruction where only a portion of the acquired counts at each projection angle are used to mimic low-dose acquisitions. RESULTS Compared to ST121 at standard dose, even at quarter dose 4D achieved better performance on overall reconstruction accuracy of the myocardium (28.7% improvement on relative root mean square error: standard vs 4D quarter p-value < 10(-30)), regional bias-variance analysis, and similar performance on accuracy of the TAC of the LV wall and detectability of perfusion defects. A slight degradation in uniformity of the LV wall was observed in 4D at quarter dose due to use of scatter correction which increases reconstruction variance. The reconstructed images from simulated and patient data show that 4D at quarter dose is visually comparable to ST121 at standard dose, if not better. Compared to ST121 and 3D, 4D images exhibit less noise artifacts and better definition of the LV wall. The 4D images are also observed to be more consistent between half dose and quarter dose. 4D also yields more consistent LVEF results at different count levels on the patient data. CONCLUSIONS With various quantitative metrics 4D reconstruction is demonstrated to achieve better or comparable performance at quarter dose (∼2.25 mSv, 75% dose reduction) compared to conventional clinical reconstruction at standard dose (∼9 mSv). Preliminary results from two patient datasets also show that 4D at an equivalent quarter dose can achieve better performance than clinical and 3D methods at higher dose levels. Such a significant dose reduction (75%) has not been demonstrated quantitatively in previous studies in GSPECT. These promising results warrant further investigations on the lower bound of dose reduction with different reconstruction strategies and more comprehensive clinical studies with greater patient variability.

Theoretical and numerical study of MLEM and OSEM reconstruction algorithms for motion correction in emission tomography

Patient body-motion and respiratory-motion impacts the image quality of cardiac PET or SPECT perfusion images. Several algorithms exist in the literature to correct for motion within the iterative maximum-likelihood reconstruction framework. In this work, three algorithms are derived using Poisson statistics to correct for patient motion. The first one is a motion compensated MLEM algorithm (MC-MLEM). The next two algorithms called MGEM-1 and MGEM-2 (short for Motion Gated EM, 1 and 2) use the motion states as subsets, in two different ways. Experiments were performed with NCAT phantoms with exactly known motion as the source and attenuation distributions. The SIMIND Monte Carlo simulation software was used to create SPECT projection images of the NCAT phantoms. The projection images were then modified to have Poisson noise levels equivalent to that of clinical acquisition. We investigated application of these algorithms to correction of (1) a large body-motion of 2 cm in Superior-Inferior (SI) and Anterior-Posterior (AP) directions each and (2) respiratory motion of 2 cm in SI and 0.6 cm in AP. We obtained the bias with respect to the NCAT phantom activity for noiseless reconstructions as well as the bias-variance for noisy reconstructions. The MGEM-1 advanced along the biasvariance curve faster than the MC-MLEM with iterations. The MGEM-1 also lowered the noiseless bias (with respect to NCAT truth) faster with iterations, compared to the MC-MLEM algorithms, as expected with subset algorithms. For the body motion correction with two motion states, after the 9th iteration the bias was close to that of MC-MLEM at iteration 17, reducing the number of iterations by a factor of 1.89. For the respiratory motion correction with 9 motion states, based on the noiseless bias, the iteration reduction factor was approximately 7. For the MGEM-2, however, bias-plot or the bias-variance-plot saturates with iteration because of successive interpolation error.

Theoretical and Numerical Study of MLEM and OSEM Reconstruction Algorithms for Motion Correction in Emission Tomography

Patient body-motion and respiratory-motion impacts the image quality of cardiac SPECT and PET perfusion images. Several algorithms exist in the literature to correct for motion within the iterative maximum-likelihood reconstruction framework. In this work, three algorithms are derived starting with Poisson statistics to correct for patient motion. The first one is a motion compensated MLEM algorithm (MC-MLEM). The next two algorithms called MGEM-1 and MGEM-2 (short for Motion Gated OSEM, 1 and 2) use the motion states as subsets, in two different ways. Experiments were performed with NCAT phantoms (with exactly known motion) as the source and attenuation distributions. Experiments were also performed on an anthropomorphic phantom and a patient study. The SIMIND Monte Carlo simulation software was used to create SPECT projection images of the NCAT phantoms. The projection images were then modified to have Poisson noise levels equivalent to that of clinical acquisition. We investigated application of these algorithms to correction of (1) a large body-motion of 2 cm in Superior-Inferior (SI) and Anterior-Posterior (AP) directions each and (2) respiratory motion of 2 cm in SI and 0.6 cm in AP. We determined the bias with respect to the NCAT phantom activity for noiseless reconstructions as well as the bias-variance for noisy reconstructions. The MGEM-1 advanced along the bias-variance curve faster than the MC-MLEM with iterations. The MGEM-1 also lowered the noiseless bias (with respect to NCAT truth) faster with iterations, compared to the MC-MLEM algorithms, as expected with subset algorithms. For the body motion correction with two motion states, after the 9th iteration the bias was close to that of MC-MLEM at iteration 17, reducing the number of iterations by a factor of 1.89. For the respiratory motion correction with 9 motion states, based on the noiseless bias, the iteration reduction factor was approximately 7. For the MGEM-2, however, bias-plot or the bias-variance-plot saturated with iteration because of successive interpolation error. SPECT data was acquired simulating respiratory motion of 2 cm amplitude with an anthropomorphic phantom. A patient study acquired with body motion in a second rest was also acquired. The motion correction was applied to these acquisitions with the anthropomorphic phantom and the patient study, showing marked improvements of image quality with the estimated motion correction.

Targeted 2D/3D registration using ray normalization and a hybrid optimizer

X-ray images are often used to guide minimally invasive procedures in interventional radiology. The use of a preoperatively obtained 3D volume can enhance the visualization needed for guiding catheters and other surgical devices. However, for intraoperative usefulness, the 3D dataset needs to be registered to the 2D x-ray images of the patient. We investigated the effect of targeting subvolumes of interest in the 3D datasets and registering the projections with C-arm x-ray images. We developed an intensity-based 2D/3D rigid-body registration using a Monte Carlo-based hybrid algorithm as the optimizer, using a single view for registration. Pattern intensity (PI) and mutual information (MI) were two metrics tested. We used normalization of the rays to address the problems due to truncation in 3D necessary for targeting. We tested the algorithm on a C-arm x-ray image of a pigs head and a 3D dataset reconstructed from multiple views of the C-arm. PI and MI were comparable in performance. For two subvolumes starting with a set of initial poses from +/-15 mm in x, from +/-3 mm (random), in y and z and +/-4 deg in the three angles, the robustness was 94% for PI and 91% for MI, with accuracy of 2.4 mm (PI) and 2.6 mm (MI), using the hybrid algorithm. The hybrid optimizer, when compared with a standard Powells direction set method, increased the robustness from 59% (Powell) to 94% (hybrid). Another set of 50 random initial conditions from [+/-20] mm in x,y,z and [+/-10] deg in the three angles, yielded robustness of 84% (hybrid) versus 38% (Powell) using PI as metric, with accuracies 2.1 mm (hybrid) versus 2.0 mm (Powell).

Segmenting and tracking diaphragm and heart regions in Gated-CT datasets as an aid to developing a predictive model for respiratory motion-correction

Diaphragm motion during respiration induces motion of the heart, which is known to cause artifacts in cardiac PET-CT imaging. A possible method of correcting for this is to build a predictive model, which requires knowledge of diaphragm and heart motion interactions. The purpose of this work was to segment and track the diaphragm in 6 respiratory- gated CT datasets for comparison with the heart motion, previously determined from affine registration. Diaphragm segmentation was performed using an algorithm developed for this purpose, based on the extraction of features from an edge- enhanced image. Three regions on the diaphragm surface were then selected for tracking, corresponding to the maxima in the left and right domes and the trough between them. The average diaphragm height in each region was calculated and tracked over the respiratory cycle, as were ratios in the height, reflecting shape. The curves generated from tracking were compared with 6 affine registration parameters (3 translations and 3 rotations) and correlations between them investigated. Correlations were derived for pairs of motion-curves for each patient and for parameter amplitudes across all patients. A number of significant correlations were found (p<0.05). The strongest relationship, as expected, was that between the diaphragm motion and the z translation of the heart, correlated in both the motion-curves and across patients. Other correlations include the amplitude of the diaphragm (right-hand dome) with the amplitudes of the x and y rotations and the x translation of the heart, indicating that these parameters may have useful predictive value. Additionally, diaphragm ratios showed correlations with certain heart motions, confirming that the changing shape of the diaphragm is linked with the type of heart motion. It can be concluded that strong links exist between certain aspects of diaphragm motion and shape change that should have predictive value in building a suitable motion-correction model.

MRI investigation of the linkage between respiratory motion of the heart and markers on patient's abdomen and chest: Implications for respiratory amplitude binning list-mode PET and SPECT studies

Respiratory motion of the heart impacts the diagnostic accuracy of myocardial-perfusion emission-imaging studies. Amplitude binning has come to be the method of choice for binning list-mode based acquisitions for correction of respiratory motion in PET and SPECT. In some subjects respiratory motion exhibits hysteretic behavior similar to damped non-linear cyclic systems. The detection and correction of hysteresis between the signals from surface movement of the patients body used in binning and the motion of the heart within the chest remains an open area for investigation. This study reports our investigation in nine volunteers of the combined MRI tracking of the internal respiratory motion of the heart using Navigators with stereo-tracking of markers on the volunteers chest and abdomen by a visual-tracking system (VTS). The respiratory motion signals from the internal organs and the external markers were evaluated for hysteretic behavior analyzing the temporal correspondence of the signals. In general, a strong, positive correlation between the external marker motion (AP direction) and the internal heart motion (SI direction) during respiration was observed. The average ± standard deviation in the Spearmans ranked correlation coefficient (ρ) over the nine volunteer studied was 0.92±0.1 between the external abdomen marker and the internal heart, and 0.87±0.2 between the external chest marker and the internal heart. However despite the good correlation on average for the nine volunteers, in three studies a poor correlation was observed due to hysteretic behavior between inspiration and expiration for either the chest marker and the internal motion of the heart, or the abdominal marker and the motion of the heart. In all cases we observed a good correlation of at least either the abdomen or the chest with the heart. Based on this result, we propose the use of marker motion from both the chest and abdomen regions when estimating the internal heart motion to detect and address hysteresis when binning list-mode emission data.

Digital anthropomorphic phantoms of non-rigid human respiratory and voluntary body motion for investigating motion correction in emission imaging

The development of methods for correcting patient motion in emission tomography has been receiving increased attention. Often the performance of these methods is evaluated through simulations using digital anthropomorphic phantoms, such as the commonly used extended cardiac torso (XCAT) phantom, which models both respiratory and cardiac motion based on human studies. However, non-rigid body motion, which is frequently seen in clinical studies, is not present in the standard XCAT phantom. In addition, respiratory motion in the standard phantom is limited to a single generic trend. In this work, to obtain a more realistic representation of motion, we developed a series of individual-specific XCAT phantoms, modeling non-rigid respiratory and non-rigid body motions derived from the magnetic resonance imaging (MRI) acquisitions of volunteers. Acquisitions were performed in the sagittal orientation using the Navigator methodology. Baseline (no motion) acquisitions at end-expiration were obtained at the beginning of each imaging session for each volunteer. For the body motion studies, MRI was again acquired only at end-expiration for five body motion poses (shoulder stretch, shoulder twist, lateral bend, side roll, and axial slide). For the respiratory motion studies, an MRI was acquired during free/regular breathing. The magnetic resonance slices were then retrospectively sorted into 14 amplitude-binned respiratory states, end-expiration, end-inspiration, six intermediary states during inspiration, and six during expiration using the recorded Navigator signal. XCAT phantoms were then generated based on these MRI data by interactive alignment of the organ contours of the XCAT with the MRI slices using a graphical user interface. Thus far we have created five body motion and five respiratory motion XCAT phantoms from the MRI acquisitions of six healthy volunteers (three males and three females). Non-rigid motion exhibited by the volunteers was reflected in both respiratory and body motion phantoms with a varying extent and character for each individual. In addition to these phantoms, we recorded the position of markers placed on the chest of the volunteers for the body motion studies, which could be used as external motion measurement. Using these phantoms and external motion data, investigators will be able to test their motion correction approaches for realistic motion obtained from different individuals. The non-uniform rational B-spline data and the parameter files for these phantoms are freely available for downloading and can be used with the XCAT license.