S Bao

Investigation of motion artifacts associated with fat saturation technique in 3D flash imaging

PURPOSE Fast low-angle shot (FLASH) imaging is widely used in dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) because it permits fast and accurate T1 measurement in vivo. Suppression of the fat signal is necessary for most FLASH applications; otherwise, fat will appear hyperintense. The fat saturation technique is one popular method to reduce fat images on clinical scanners. However, fat saturation combined with the 3D FLASH sequence in breast DCE-MRI scans results in heavy ghosting artifacts caused by heartbeat. We used simulation and experimental scans to determine the cause of these artifact-enhancement phenomena. METHODS We simulated imaging of motion in the x, y, and z directions, with and without fat saturation, to investigate the origin of artifacts. Fourier transform (FT) of the whole field of view was used in the simulation, and we assumed that the uniform phantom was static during one TR. The amplitude of each echo was considered a factor in the FT data. Images were reconstructed using FT data from different phantom positions multiplied by the amplitude factor. Phantom experiments and volunteer studies were implemented to verify the conclusion. RESULTS Both phantom and volunteer results showed artifacts similar to those in simulation images. We found that FLASH sequence without fat saturation is insensitive to motion. Fat saturation radiofrequency pulses placed before each group of echoes disrupted the steady state of the signal amplitude and produced a low-pass filter effect that enhanced the motion artifacts. CONCLUSIONS We conclude that the low-pass filter effect associated with the fat saturation technique is responsible for dramatically increased motion artifacts.

Flow compensation for the fast spin echo triple-echo Dixon sequence

The fast spin echo (FSE) triple-echo Dixon (FTED) sequence uses bipolar triple-echo readout during each echo-spacing period of FSE to collect all the images necessary for Dixon water and fat separation in a single scan. In comparison to other FSE implementations of the Dixon technique, the triple echo readout used in FTED incurs minimal deadtime in the pulse sequence design and thus greatly enhances the overall scan efficiency. A potential drawback of FTED is that the time dependence of the gradient moment along the frequency encode direction becomes more complicated than in FSE and flow compensation based on the gradient moment (GM) nulling is difficult to achieve. In this work, the first order GM along the frequency encode direction of FTED was examined and two different methods to minimize the GM were proposed. The first method nulls the GM at all the locations of the refocusing radiofrequency pulses so that the Carr-Purcell Meiboom-Gill condition is always maintained. The second method minimizes the GM of the spin echo component of the FSE signal at the echo locations. The efficacy of both methods in reducing the first order GM and flow-related artefacts was demonstrated both in phantom and in images in vivo.

SU‐FF‐I‐132: Simultaneous Estimation of Perfusion and Permeability Parameters: Validation Study On Animal Model

Purpose: To demonstrate the feasibility of simultaneously obtaining permeability and perfusion parameters using one MR pulse sequence with one administration of contrast agent (CA). Method and material: A 3D dual‐echo pulse sequence was used to simultaneously acquire T1‐weighted and T 2 >* ‐weighted images. Data were acquired from a stroke mouse model using the pulse sequence with six pre‐contrast and eighty four post‐contrast dynamic contrast‐enhanced series. An analysis method was developed to estimate perfusion and permeability parameters. Shortening in T1 by CA enhances signal intensity, while shortening in T2* reduces signal intensity. This competitive process by T1 and T2* was taken into account in our analysis method. To minimize subjectivity and increase consistency, AIF was automatically estimated using fuzzy C‐means clustered based techniques. Result: Three‐dimensional pharmacokinetic parameter maps of volume transfer constant (Ktrans), extracellular extravascular space (Ve) and reflux rate (Kep), as well as perfusion parameter maps of the relative cerebral blood volume (rCBV), relative cerebral blood flow (rCBF) and relative mean transit time (rMTT) were successfully obtained. Conclusion: It is feasible to simultaneously estimate permeability and perfusion parameters using data acquired by one pulse sequence with only one administration of CA. This new technique should prove to be a useful tool for both animal and human DCE‐MRI studies.

Dose point kernel for boron-11 decay and the cellular S values in boron neutron capture therapy

The study of the radiobiology of boron neutron capture therapy is based on the cellular level dosimetry of boron-10s thermal neutron capture reaction 10B(n,alpha)7Li, in which one 1.47 MeV helium-4 ion and one 0.84 MeV lithium-7 ion are spawned. Because of the chemical preference of boron-10 carrier molecules, the dose is heterogeneously distributed in cells. In the present work, the (scaled) dose point kernel of boron-11 decay, called 11B-DPK, was calculated by GEANT4 Monte Carlo simulation code. The DPK curve drops suddenly at the radius of 4.26 microm, the continuous slowing down approximation (CSDA) range of a lithium-7 ion. Then, after a slight ascending, the curve decreases to near zero when the radius goes beyond 8.20 microm, which is the CSDA range of a 1.47 MeV helium-4 ion. With the DPK data, S values for nuclei and cells with the boron-10 on the cell surface are calculated for different combinations of cell and nucleus sizes. The S value for a cell radius of 10 microm and a nucleus radius of 5 microm is slightly larger than the value published by Tung et al. [Appl. Radiat. Isot. 61, 739-743 (2004)]. This result is potentially more accurate than the published value since it includes the contribution of a lithium-7 ion as well as the alpha particle.

TU‐C‐BRB‐02: Personalized Assessment of KV‐CBCT Doses in Image Guided Radiotherapy of Pediatric Cancer Patients

Purpose: To develop an empirical method for estimation of doses to critical structures of children undergoing kV‐CBCT imaging for image‐guidedradiotherapy. Methods: Forty pediatric patients were retrospectively assorted into four groups by gender and scanned sites with approval of IRB. Critical structures in CTimages were delineated on Eclipse treatment planning system and were subsequently packaged into patient CT phantoms for Monte Carlo simulations. A benchmarked EGS4 Monte Carlo code was used to calculate 3D dose distributions produced by High‐quality head and Pelvic protocols that were pre‐defined by manufacturer. The ratios of irradiated volume over longitudinal length (V/L) and the equivalent sphere diameters (ESDs) were calculated by Eclipse. In accordance with the scanned region, either occipital‐frontal circumferences (OFCs) or hip circumferences (HIPs) were calculated by DICOMan software. Regression was performed with SigmaPlot software suite where average standard errors of estimate (SEEs) were compared to evaluate the accuracy of regression and reliability of various parameters. Results: For both sexes, mean doses of all critical structures were observed to decrease linearly with the increase of all physical parameters including age, weight, V/L, ESD, OFC, and HIP. As suggested by smaller standard errors of estimates, OFC and weight are recommended for head dose estimation while ESD and weight are more appropriate for pelvic/abdominal dose regression. Unique empirical functions for the estimation of doses to critical organs for male or female patients are presented which make a personalized dose assessment readily feasible. Conclusions: Using Monte Carlo simulations in a population of 40 selected children, kV‐CBCT doses were found highly correlated with body sizes of children. Considering uneven body development levels among children of same age, OFC and weight are recommended primarily for dose estimation in head, while weight and ESD are preferred for pelvic dose assessment. This work was partially supported by National Basic Research Subject (973) of China (No.2011CB707701), China Scholarship Council (No. 2010601107), and Joint Research Foundation of Beijing Education Committee, China (No. JD100010607)

TH-E-BRC-06: Optimizing KVCBCT Scan Protocol for Individual Patient Undergoing IGRT

Purpose: To develop an automatic tool for optimal mAs and kVp settings in kVCBCT based on patient CTanatomy and user‐defined priorities on normal tissue sparing and image quality. Methods: An EGS4 Monte Carlo code was employed to calculate 3D dose distributions in patients scanned with kVCBCT at both half‐fan and full‐fan modes. Absorbed doses to various organs were analyzed. The relationship between absorbed doses, mAs and kVp was studied with phantom measurements and fit with empirical functions. An optimizer tool based on conjugated gradient searching algorithm in multi‐dimensions was developed to generate the optimal settings of mAs and kVp based on 3D dose depositions to various organs of individual patient, with consideration of user‐defined priorities on normal tissue sparing and image quality. The effectiveness of the optimizer was tested on Catphan and various patient CTanatomies. Results: Our phantom study indicates that the dose contributions from OBI CBCT half‐ fan and full‐fan modes depend almost linearly on mAs while exponentially on kVp. Empirical functions can be used to reproduce the kVCBCT doses accurately at any allowed mAs and kVp settings. In general, our optimizer recommends much less mAs and kVp than default settings without compromising kVCBCT image quality. Compared to default modes, the optimized modes deposit much less doses, with reduction of 61–66% to the adult and 45–68% to the pediatric patients, respectively. Conclusions: A novel kVCBCT scan protocol optimizer has been developed to help clinicians choose appropriate mAs and kVp settings for individual patient while maintaining a balance between normal tissue sparing and image quality of kVCBCT scans. With optimal settings, the dose depositions to the patients could reduce by 45–68% on average in comparison to the default without compromising image quality. The developed optimizer could potentially improve radiation safety in medical imaging of both adult and pediatric patients.

TH‐D‐201C‐11: Reference Image Compressed Sensing Based Dynamic Contrast Enhanced MRI

Purpose: The purpose of this work is to implement the CS‐based MRI in DCE‐MRI to develop a new method that can exhibit both high temporal and high spatial resolution results which are significant for DCE‐MRI and related diagnosis. Method and Materials: By exploring the DCE specific characteristic that administration of contrast agent (CA) induces change only in image signal intensity in certain areas such as vessels and lesions no significant change in anatomical structure we have developed a novel approach Reference imAge based Compressed sEnsing (RACE) to capitalize the sparsity and compressibility of DCE‐MRI. Phantom experiments have been performed on an MAGNETOM ESSANZA 1.5T MRI scanner (Siemens Erlangen Germany)focusing on the study of temporal and spatial resolution respectively. Spatial finite difference (SPD) is chosen to be the sparse transformation. 3D radial sampling is implemented as under sampling scheme. The reconstruction is based on solving a constrained total variation (TV)‐norm minimization problem. Results: According to the phantom experiments during a same time course much more (up to 10 times) time frames can be obtained with RACE which means higher temporal resolution. The comparisons demonstrate that the details of the dynamic curves can be detected using RACE especially when the intensity varying rate is high (Fig.3 c). Meanwhile the spatial resolution is not degraded. Conclusion: This work proved that RACE which properly explored the features of DCE‐MRI can significantly improve the temporal resolution of DCE‐MRI without degrading its spatial resolution.Conflict of Interest (only if applicable): Phantom experiments are supported by Siemens Mindit Magnetic Resonance Ltd. Co.