Keum Sil Lee

Single-Cell Tracking With PET Using a Novel Trajectory Reconstruction Algorithm

Virtually all biomedical applications of positron emission tomography (PET) use images to represent the distribution of a radiotracer. However, PET is increasingly used in cell tracking applications, for which the “imaging” paradigm may not be optimal. Here, we investigate an alternative approach, which consists in reconstructing the time-varying position of individual radiolabeled cells directly from PET measurements. As a proof of concept, we formulate a new algorithm for reconstructing the trajectory of one single moving cell directly from list-mode PET data. We model the trajectory as a 3-D B-spline function of the temporal variable and use nonlinear optimization to minimize the mean-square distance between the trajectory and the recorded list-mode coincidence events. Using Monte Carlo simulations (GATE), we show that this new algorithm can track a single source moving within a small-animal PET system with 3 mm accuracy provided that the activity of the cell [Bq] is greater than four times its velocity [mm/s]. The algorithm outperforms conventional ML-EM as well as the “minimum distance” method used for positron emission particle tracking (PEPT). The new method was also successfully validated using experimentally acquired PET data. In conclusion, we demonstrated the feasibility of a new method for tracking a single moving cell directly from PET list-mode data, at the whole-body level, for physiologically relevant activities and velocities.

GPU-based prompt gamma ray imaging from boron neutron capture therapy

PURPOSE The purpose of this research is to perform the fast reconstruction of a prompt gamma ray image using a graphics processing unit (GPU) computation from boron neutron capture therapy (BNCT) simulations. METHODS To evaluate the accuracy of the reconstructed image, a phantom including four boron uptake regions (BURs) was used in the simulation. After the Monte Carlo simulation of the BNCT, the modified ordered subset expectation maximization reconstruction algorithm using the GPU computation was used to reconstruct the images with fewer projections. The computation times for image reconstruction were compared between the GPU and the central processing unit (CPU). Also, the accuracy of the reconstructed image was evaluated by a receiver operating characteristic (ROC) curve analysis. RESULTS The image reconstruction time using the GPU was 196 times faster than the conventional reconstruction time using the CPU. For the four BURs, the area under curve values from the ROC curve were 0.6726 (A-region), 0.6890 (B-region), 0.7384 (C-region), and 0.8009 (D-region). CONCLUSIONS The tomographic image using the prompt gamma ray event from the BNCT simulation was acquired using the GPU computation in order to perform a fast reconstruction during treatment. The authors verified the feasibility of the prompt gamma ray image reconstruction using the GPU computation for BNCT simulations.

MO‐D‐141‐02: Evaluation of Optimal Gating Respiratory Motion Management Technique On PET Quantification

PURPOSE To evaluate the uncertainty of PET quantification resulting from irregular breathing patterns when optimal gating (HD-Chest) acquisition is performed on a Siemens Biograph PET-CT scanner. METHODS A NEMA phantom containing spherical VOIs (Volumes-of-Interest) with diameters of 28, 22, 17, 13, and 10 mm was scanned on a moving platform. Seven breathing traces that were previously recorded optically from lung PET patients and one synthetic regular trace were used to drive the platform. Motion amplitudes for individual traces were 5∼20 mm peak-to-peak with intra-traces amplitude variances ranging between 1.6 mm to 10.3 mm. For each trace, the average (SUVavg) and maximum (SUVmax) standard uptake values within the VOIs were calculated. VOIs volumes were also evaluated at a threshold equal to 10% of SUVmax. Relative percent deviations of these parameters from ground truth values derived from a static acquisition were calculated. RESULTS The average deviations for all patient breathing patterns were 10.5% (SD=5.9%), 11% (SD=7.9%), and 9% (SD=9.3%) for volume, SUVavg, and SUVmax, respectively. The average deviations for the regular breathing pattern were 3.8%, 5.2%, and 6.6% for volume, SUVavg, and SUVmax, respectively. VOI size significantly affected the uncertainty. For the 28 mm VOI the average deviations were 6.9%, 6.6%, and 2.7% for volume, SUVavg, and SUVmax, respectively while for the 10 mm VOI these were 11.8%, 15.1%, and 19.3% for volume, SUVavg, and SUVmax, respectively. Without any breathing motion management technique in place, the average deviations for the regular breathing pattern were 16.6%, 42.8%, and 42.4% for volume, SUVavg, and SUVmax, respectively. CONCLUSION The optimal gating (HD-Chest) technique improved the PET quantification. However, one needs to take the residual deviations for regular and for all irregular breathing patterns into account for PET quantification. M. Casey and R. Rajaram are the employees of Siemens Medical Solutions.

TU‐E‐141‐07: Clinical Evaluation of the Iterative Metal Artifact Reduction Algorithm for CT Simulation in Radiotherapy

PURPOSE The iterative metal artifact reduction (IMAR) algorithm has been proposed for commercial implementation in upcoming Siemens platforms. The purpose of this study is to evaluate the performance of this algorithm in radiation oncology settings. METHODS Mean CT numbers and noise (standard-deviation) within delineated regions of interest were compared before/after IMAR correction on standard electron-density phantom images. Patient IMAR-corrected images were evaluated by 4 observers and ranked based on conspicuity of structures near artifacts (0-5 scale, 5 best score). The dosimetric impact of utilizing IMAR-corrected patient images for planning was analyzed by comparing original dose distributions and those recalculated on IMAR-corrected images. All images were acquired on a Siemens Definition scanner. In order to reference the observations herein, all analyses were also conducted on images corrected with a second algorithm: metal deletion technique (MDT), available for public use. RESULTS IMAR accurately recovers CT numbers. CT number percent differences were reduced on average from 62% to 18%, while average noise percent differences were minimally reduced (146% before, 140% after). MDT performed worse retrieving mean CT numbers (62% to 27%), and better at reducing noise (146% to 24%). After visually inspecting the images, physicians agreed that IMAR-corrected images offered better confidence at reading patient anatomy than original images. The MDT-corrected images scored 4.3 on average while IMAR-corrected images scored 4 with reviewing physicians (p = 0.052). Local dose differences up to ±20-30cGy were noted, but γ-analysis (3%/3mm) did not indicate major overall differences between plans calculated on original images and those calculated on IMAR-corrected images. CONCLUSION The IMAR algorithm accurately recovered CT numbers (better than MDT), while minimally reducing noise values (worse than MDT). No clinically significant differences were detected between dose distributions calculated on original CT images and those planned on IMAR-corrected images. Initial analysis indicates that IMAR images could be used for treatment planning. Siemens Healthcare.

Performance evaluation of RF coils integrated with an RF-penetrable PET insert for simultaneous PET/MRI

An “RF‐penetrable” PET insert that allows the MR body coil to be used for RF transmission was developed to make it easier for an existing MR center to achieve simultaneous PET/MRI. This study focuses on experiments and analyses to study PET/RF coil configurations for simultaneous PET/MR studies.

Evaluation of Zero-TE-based attenuation correction methods on PET quantification of PET/MRI head and neck lesions

Quantitative PET image reconstruction requires an accurate map of photon attenuation coefficients (μ-map) in order to correct the PET emission data. Current PET/MR imaging systems use methods based on MR image segmentation with subsequent assignment of empirical attenuation coefficients. In this study we examine the differences in the quantification of 18F-FDG standardized uptake values (SUV) in head and neck cancer, using two different MR imaging sequences for MR-based attenuation correction (MRAC): a zero echo time (ZTE) sequence which can image bone directly (ZTE-MRAC), and a vendor-provided 2-point Dixon sequence that neglects bone (Dixon-MRAC). The μ-maps from each MRAC techniques were compared to CT-based attenuation correction (CTAC) maps. Percent SUV-mean and SUV-max differences in relevant regions of interest (ROIs) were calculated for three patients. Relative to Dixon-MRAC, we observed 15±7% and 14±8% increase of SUV-mean and SUV-max, respectively, when ZTE-based bone information was incorporated in the attenuation map and using Dixon-based attenuation map, respectively. We also observed that use of Dixon-MRAC led to 7±7% and 8±8% underestimation of SUV-mean and SUV-max, respectively, whereas with ZTE-MRAC led to 6±8% and 5±8% higher SUV-mean and SUV-max, respectively, compared to CTAC. This study is the first demonstration of ZTE-based attenuation correction in the head and neck region and compared with CTAC as a gold standard with the goal of improving PET quantitation. The study shows that incorporation of bone information on μ-maps has a significant impact on SUV quantitation in head and neck cancer lesions.

Motion modulated sensitivity gating (MMSG) using internal signal for quantitative PET

Nuclear imaging modalities such as PET and SPECT are used as quantification tools for tumor staging, planning, and response evaluation. The imaging findings of these modalities are largely affected by respiratory motion [1]. Common approach to motion management is to acquire and reconstruct the data in relation to externally recorded signal by motion detection systems. Currently available respiratory motion detection systems typically account for the surrogate structures. However, the surrogate structures can be an inaccurate representation of internal motion [2]. Therefore, a respiratory motion management technique that entirely relies on external breathing motion detection system could lead to potentially significant errors in quantification. We hypothesize that for a given PET bed position, the temporal pattern of cumulative counts within a predefined time bin can provide a direct measurement of internal organ motion and thus lead to more accurate quantification of images being acquired. Therefore, instead of using approximately one third of total counts for image reconstruction of all bed position in a whole body scan, different percentage of total counts depending on the magnitude of motion occurred within the FOV of each bed acquisition can be used for image reconstruction.

Count-based listmode respiratory motion detection for quantitative PET

Respiratory motion can cause the image quality degradation and inaccurate quantification for nuclear imaging. This study is to propose a respiratory motion detection method that the motion of internal organs can be directly detected from their own motion instead of using a surrogate structure such as diaphragm to represent the motion of every organ.

SU‐GG‐T‐112: A Concept of Pseudo‐4D IMRT Optimization Using Phase‐Dependent Apertures

Purpose: To present a concept of pseudo‐4D IMRT optimization based on 3D optimization of all the respiratory phases. Method and Materials: A digital phantom with an embedded target was created. A 10‐phase cyclic respiratory motion was designed. Five equally spaced beams encompassing the target were used. Individual 3D optimization was performed for each phase. After the optimization, a phase was chosen as a reference to which all doses were mapped through the use of a registration model. A new phase‐dependent leaf sequence (pd‐LS) file for each field was re‐constituted by selecting the first aperture from the first phase, the second aperture from the second phase, and so on until all the corresponding apertures were selected. For each field, the apertures in the pd‐LS file were then served as inputs for a re‐optimization of their shapes and dose weights to achieve the result of the same intensity map of the field in the reference phase. The resultant 4D dose from all the fields was calculated by mapping the dose deposited for each phase by the respective apertures onto the reference phase. The 4D plan and DVH were studied and compared with those of the 3D plan optimized for the same reference phase. Results: Phase‐dependent leaf sequence file has been reconstituted that ‘follows’ the target motion; the pseudo‐4D plan could be delivered more efficiently by utilizing the full respiratory cycle. The pseudo‐4D plan yielded a similar coverage for the target and a lower dose to the surrounding normal tissues when compared to the 3D plan. Conclusion: To a first approximation, a pseudo‐4D dose optimization can be accomplished with 3D optimization tool by (1) combining the apertures selectively according to the respiratory phases and order of the apertures they are executed and (2) re‐optimizing the shape and dose weight of the aperture.

SU-GG-T-109: Four Dimensional Inverse Planning for Intensity Modulated Radiation Therapy

Purpose: This work develops 4D inverse planning methods and demonstrates the potential benefit of 4D IMRT.Method and Materials: Two 4D planning strategies are proposed and compared. The first one treats all respiration phases as a system and optimizes the dose delivery collectively in space and phase. The method is referred to as collective optimization of all phases (COAP). In this approach, a deformable model is employed to establish a voxel‐to‐voxel correspondence and the goal is to maximize the accumulative dose to the tumor target while minimizing the dose to the organ‐at‐risk (OARs). The second one treats each phase as an independent 3D inverse planning problem and optimizes them separately. The final dose distribution is obtained by summing the dose of each phase after a deformable image registration. This method is called separate optimization of each phase (SOEP). In both approaches, the dose is optimized with a linear programming technique. Results: The resultant dose distribution of COAP is markedly better than that of SOEP in both target dose coverage and organ‐at‐risk sparing. The improvement of COAP is resulted from reallocation of dose among the phases to cater for anatomical changes during the breathing process. It is found that, for a phase with favorable geometry for dose delivery, more doses are allocated by COAP, and vise versa. COAP optimally assigns dose for all the involved phases. Because of the lack of this degree of freedom, SOEP yields almost identical intensity maps and dose distributions for all the phases. Conclusion: Simultaneous spatio‐temporal doseoptimization in 4D inverse planning allows one to take consideration of the spatial variation of the patient anatomy caused by respiration and yields the optimal accumulative dose distribution.