International Journal of
Radiation Oncology biology
physics
www.redjournal.org
Physics Contribution
Optimizing 4-Dimensional Magnetic Resonance Imaging Data Sampling for Respiratory Motion Analysis of Pancreatic Tumors Bjorn Stemkens, MSc,* Rob H.N. Tijssen, DPhil,* Baudouin D. de Senneville, PhD,y,z Hanne D. Heerkens, MD,* Marco van Vulpen, MD, PhD,* Jan J.W. Lagendijk, PhD,* and Cornelis A.T. van den Berg, PhD* *Department of Radiotherapy and yImaging Division, University Medical Center Utrecht, Utrecht, The Netherlands; and zL’Institut de Mathe´matiques de Bordeaux, Unite´ Mixte de Recherche 5251, Centre National de la Recherche Scientifique/University of Bordeaux, Bordeaux, France Received May 8, 2014, and in revised form Sep 22, 2014. Accepted for publication Sep 22, 2014.
Summary A novel MRI method for simultaneous 4D motion analysis of abdominal tumors and organs at risk is demonstrated. Eleven healthy subjects were scanned for protocol optimization. The optimized protocol was then used to scan 6 patients for whom motion trajectories were calculated for both the tumor and duodenum. The results demonstrated that the proposed technique was able to calculate 4D motion trajectories within a clinically acceptable time frame.
Purpose: To determine the optimum sampling strategy for retrospective reconstruction of 4-dimensional (4D) MR data for nonrigid motion characterization of tumor and organs at risk for radiation therapy purposes. Methods and Materials: For optimization, we compared 2 surrogate signals (external respiratory bellows and internal MRI navigators) and 2 MR sampling strategies (Cartesian and radial) in terms of image quality and robustness. Using the optimized protocol, 6 pancreatic cancer patients were scanned to calculate the 4D motion. Region of interest analysis was performed to characterize the respiratory-induced motion of the tumor and organs at risk simultaneously. Results: The MRI navigator was found to be a more reliable surrogate for pancreatic motion than the respiratory bellows signal. Radial sampling is most benign for undersampling artifacts and intraview motion. Motion characterization revealed interorgan and interpatient variation, as well as heterogeneity within the tumor. Conclusions: A robust 4D-MRI method, based on clinically available protocols, is presented and successfully applied to characterize the abdominal motion in a small number of pancreatic cancer patients. Ó 2015 Elsevier Inc.
Reprint requests to: Bjorn Stemkens, MSc, Department of Radiotherapy, University Medical Center Utrecht, Q.00.118, Heidelberglaan 100, 3584CX Utrecht, The Netherlands. Tel: (þ31) (0) 88-7553035; E-mail: [email protected] Int J Radiation Oncol Biol Phys, Vol. 91, No. 3, pp. 571e578, 2015 0360-3016/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2014.10.050
Conflict of interest: none. Supplementary material www.redjournal.org.
for
this
article
can
be
found
at
572
Stemkens et al.
Introduction Radiation therapy treatment of pancreatic cancer is challenging because of motion, primarily induced by respiration. Conventionally, margins are used to ensure full-dose coverage of the target at the cost of extra radiation to nearby organs at risk (OARs). Prior knowledge about patient-specific motion, however, may lead to more conformal radiation plans (1). Currently, 4-dimensional (4D)-CT is the method of choice for motion characterization in radiation therapy. Unfortunately, CT has limited soft-tissue contrast, and 4D-CT often suffers from volume inconsistencies (eg duplicating, incomplete, or overlapping structures) caused by discrepancies between the breathing frequency and the pitch (2, 3). Magnetic resonance imaging (MRI), on the other hand, has superior soft-tissue contrast, allowing better differentiation between tumor and surrounding tissue. Moreover, the flexibility of image acquisition in MRI could be used to overcome some of the image artifacts typical of 4D-CT. Previous authors have reported MRI-based motion characterization using 2D-cine MRI (4-7). However, these methods solely provide tumor motion characteristics and do not provide information of the OARs. To construct high-resolution 4D-MR data, both prospective gated imaging and retrospective binning have been explored in previous studies. In prospective gated imaging, 3-dimensional (3D) volumes are acquired at different breathing depths (8), determined for example by a 1-dimensional (1D)-MRI navigator that records the diaphragm position (9). However, prospective gating may prolong scan time, and the probabilistic information about the duration of each respiratory phase is not captured. Consequently, retrospective binning methods have gained interest. To bin the acquired data into appropriate respiratory phases, a number of respiratory surrogates have been investigated, such as respiratory bellows (10), image-based body area (11), or 2-dimensional (2D) navigator slices (12). Cai et al (11) used 2D axial slices for retrospective slice stacking, which is closely related to 4D-CT and may therefore suffer from volume inconsistencies similar to those with 4D-CT. To address these disadvantages, Von Siebenthal et al (12) and Tryggestad et al (10) used retrospective stacking of 2D sagittal slices. Whereas Von Siebenthal et al used a custom 2D acquisition, Tryggestad et al used clinical sequences and a novel 2-pass sorting approach. In lieu of these multislice 2D methods, volumetric 3Deacquired MRI scans, in which the entire volume is excited each time, have been used by Buerger et al (13) for retrospective binning because they inherently have a higher signal-to-noise ratio (SNR) and no through-plane motion. However, the last 3 methods either require a custom read-out or long acquisition/reconstruction times, which is undesirable for clinical practice. Ideally, one would like to benefit from the higher SNR of 3D acquisitions while still using clinically available scan protocols.
International Journal of Radiation Oncology Biology Physics
The goal of this study, therefore, is to develop a robust 4D-MRI method, based on 3D-acquired MR data, to characterize respiratory-induced motion of pancreatic tumors and surrounding OARs using clinically available scan protocols within a clinically acceptable time frame. Two motion surrogates, which serve as input for the binning process, are assessed for their agreement with respiratoryinduced organ motion. Moreover, 2 MRI sampling strategies are explored in terms of robustness and image quality in multiple volunteers. The final resulting technique is applied to 6 pancreatic cancer patients. To demonstrate the applicability of this technique to radiation therapy, motion of both tumor and OARs are characterized using a 3D optical flow (OF) algorithm, because of its short processing time and minimal user intervention (14).
Methods and Materials We performed 2 different experiments in which the surrogate signal and MRI sampling strategy were independently optimized. The optimized protocol was then used to scan 6 pancreatic cancer patients. Three-dimensional motion was calculated for both the tumor and the duodenum over the complete respiratory cycle. All experiments were conducted on a 1.5 T Philips Achieva MR scanner (Philips Healthcare, Best, The Netherlands). Processing was performed using MATLAB (The Mathworks, Natick, MA).
Surrogate signal comparison: Correlation with pancreatic motion Six healthy subjects were scanned using a fast, coronal, 2D transient balanced steady-state gradient echo (balanced turbo field echo) sequence with radial read-out (a Z 30+ , time to echo/time to repetition 1.3/2.6 ms, field of view [FOV] 294 294 mm2, pixel size 1.5 1.5 mm2, slice thickness 7 mm, bandwidth 1580 Hz, turbo-factor 128, 500 dynamics [ie number of consecutive repeats of the 2D sequence], temporal frequency 2.6 Hz). The imaging plane was angulated parallel to the spine to sample along the principal axis of pancreatic motion, thereby reducing through-plane motion. The image acquisition was interleaved with a 1D MRI navigator, which was positioned at the dome of the right hemidiaphragm and processed online using the vendor’s reconstruction method. The respiratory bellows, provided by the MRI vendor, were placed on the upper abdomen. While the navigator is inherently synchronized with the motion in the 2D images, the bellows signal required minor manual synchronization, owing to imperfect temporal synchronization with the MR acquisition. Pixel-wise nonrigid displacement was calculated on the 2D time series using a 2D OF algorithm (15-17). In short, the OF algorithm calculates nonrigid displacement between 2 imaging frames according to intensity gradients, with an additional constraint on motion smoothness to model
Volume 91 Number 3 2015
Optimizing 4D-MRI for motion analysis
Table 1 Acquisition parameters for the in vivo experiments for 2 different in-plane sampling strategies, with the number of healthy volunteers scanned using the corresponding protocol Parameter
Cartesian
Radial
Time to echo (ms) Time to repetition (ms) a (o) FOV (mm3) Matrix size Dynamics Tacq (min) Bandwidth (Hz) No. of subjects
1.45 3.0 30 330 250 64 152 126 16 50 10:13.7 1780 5
1.45 3.0 30 330 330 96 152 152 24 25 8:35.1 1780 10
elastic organ deformation. To assess the agreement of the 2 surrogates with the pancreatic motion, we performed a general linear model (GLM) analysis, which is a leastsquares fit to the data described by YZX b þ ε. Here, Y is the pixel-based cranio-caudal (CC) motion, obtained from the pixel-by-pixel vector field calculated by the 2D OF algorithm, b the parameter estimates, X the surrogate signal, and ε represents the residuals. This analysis was performed on each pixel independently. The root mean square of the residuals was calculated and plotted in a spatial residual error map to find the surrogate with the lowest residuals. Furthermore, power spectrum analysis was performed on the surrogate signals and the pancreatic motion to gain insight into the agreement in frequency range of the surrogates with the pancreatic motion.
In vivo experiments Eleven healthy volunteers provided written informed consent to be scanned according to institutional rules to test the
A
573
feasibility and robustness of the proposed 4D-MRI technique. A clinically available 3D balanced turbo field echo (bTFE) with Cartesian or radial in-plane sampling was used, utilizing the parameters described in Table 1. Turbo direction was along kz and the shot length equal to the amount of slices, thus acquiring all kz lines within a single shot (Fig. 1a). Before each shot, a 1D MR navigator, placed at the dome of the right hemidiaphragm, was acquired. Twenty-five dynamics (ie the number of repeats of a 3D volume acquisition) were acquired for the radial protocol as a concession between imaging time and filling. For the Cartesian protocol 50 dynamics were acquired to assure sufficient data were acquired at the center of k-space. Four volunteers were scanned with both radial and Cartesian protocols. Because of time constraints, 6 volunteers were scanned using solely the radial protocol (from which 2 had a smaller FOV of 330 330 64 mm3), whereas 1 subject was scanned using only the Cartesian protocol. Reconstruction was performed offline on an 8-core CPU computer using ReconFrame (Gyrotools, Zurich, Switzerland), which is a commercial MATLAB-based MRI reconstruction software package with the ability to operate on the MRI scanner. Using the respiratory waveform derived from the navigator, k-space data were sorted into 10 respiratory phases using phase binning (ie equal-timeebased binning), similar to 4D-CT phase binning. Complex averaging was performed to increase SNR, when certain k-space segments were acquired multiple times. In case no k-lines were assigned, the lines were left blank. Image quality was assessed both visually and using the structural similarity (SSIM) index (18), which calculates the similarity between 2 images as a ratio between 0 (no similarity) and 1 (perfect similarity). The SSIM values were calculated between undersampled images and the full reference (ie 50 dynamics for Cartesian, 25 dynamics for radial).
B
C Registered images
3D images
r2,1 r1,1 I2 kz
D2
Reg 2 I1
D3
Reg 3
I3
r1,n
r2,1
r 1,1
D10
K1 r2,n
r1,n
K 10 kz
kz
10 respiratory phases (k-spaces) Respiratory phases
k-space read-out
Graphical overview phase binning method
I10
Reg 10
Calculate 4D deformation field and registered images
Fig. 1. Workflow of the in vivo experiments. (a) Cartesian and radial trajectories with turbo direction along kz. First, readout lines r1,1 . r1,n are acquired within 1 turbo field echo shot. In the next shot, r2,1 . r2,n are acquired, etc. (b) Graphic overview of the phase binning method for a respiratory signal. (c) After reconstruction, 3-dimensional images are registered to the reference volume, which results in a 4-dimensional deformation field.
574
International Journal of Radiation Oncology Biology Physics
Stemkens et al.
4D nonrigid motion analysis of tumor and OARs Six patients with an unresectable pancreas tumor were scanned using the radial protocol described in Table 1 according to institutional rules after written informed consent. After phase binning and image reconstruction, nonrigid displacement between the respiratory phases was estimated using a multithreaded optimized 3D OF algorithm (14). All respiratory phases were registered to 1 reference phase (the first, exhale phase), as can be seen in Figure 1c. The tumor and OARs were delineated on the reconstructed reference volume, and the mean motion within each delineated structure was calculated for all respiratory phases to construct motion trajectories of the tumor and duodenum over 10 respiratory phases.
The residual error within the pancreas over the 6 subjects is 2.4 1.1 times higher for the bellows compared with when the navigator is used. Moreover, the respiratory bellows signal required manual synchronization. Figure 2b displays the power spectra of the pancreatic motion, and the spectra of the 2 surrogates for the first 2 subjects shown in Figure 2a. Aside from large respiratory peaks between 0.2 and 0.32 Hz in all spectra, the power spectrum of the respiratory bellows of subject 2 displays additional peaks close to 1 Hz, owing to cardiac motion. Moreover, low frequencies (
Radiation Oncology biology
physics
www.redjournal.org
Physics Contribution
Optimizing 4-Dimensional Magnetic Resonance Imaging Data Sampling for Respiratory Motion Analysis of Pancreatic Tumors Bjorn Stemkens, MSc,* Rob H.N. Tijssen, DPhil,* Baudouin D. de Senneville, PhD,y,z Hanne D. Heerkens, MD,* Marco van Vulpen, MD, PhD,* Jan J.W. Lagendijk, PhD,* and Cornelis A.T. van den Berg, PhD* *Department of Radiotherapy and yImaging Division, University Medical Center Utrecht, Utrecht, The Netherlands; and zL’Institut de Mathe´matiques de Bordeaux, Unite´ Mixte de Recherche 5251, Centre National de la Recherche Scientifique/University of Bordeaux, Bordeaux, France Received May 8, 2014, and in revised form Sep 22, 2014. Accepted for publication Sep 22, 2014.
Summary A novel MRI method for simultaneous 4D motion analysis of abdominal tumors and organs at risk is demonstrated. Eleven healthy subjects were scanned for protocol optimization. The optimized protocol was then used to scan 6 patients for whom motion trajectories were calculated for both the tumor and duodenum. The results demonstrated that the proposed technique was able to calculate 4D motion trajectories within a clinically acceptable time frame.
Purpose: To determine the optimum sampling strategy for retrospective reconstruction of 4-dimensional (4D) MR data for nonrigid motion characterization of tumor and organs at risk for radiation therapy purposes. Methods and Materials: For optimization, we compared 2 surrogate signals (external respiratory bellows and internal MRI navigators) and 2 MR sampling strategies (Cartesian and radial) in terms of image quality and robustness. Using the optimized protocol, 6 pancreatic cancer patients were scanned to calculate the 4D motion. Region of interest analysis was performed to characterize the respiratory-induced motion of the tumor and organs at risk simultaneously. Results: The MRI navigator was found to be a more reliable surrogate for pancreatic motion than the respiratory bellows signal. Radial sampling is most benign for undersampling artifacts and intraview motion. Motion characterization revealed interorgan and interpatient variation, as well as heterogeneity within the tumor. Conclusions: A robust 4D-MRI method, based on clinically available protocols, is presented and successfully applied to characterize the abdominal motion in a small number of pancreatic cancer patients. Ó 2015 Elsevier Inc.
Reprint requests to: Bjorn Stemkens, MSc, Department of Radiotherapy, University Medical Center Utrecht, Q.00.118, Heidelberglaan 100, 3584CX Utrecht, The Netherlands. Tel: (þ31) (0) 88-7553035; E-mail: [email protected] Int J Radiation Oncol Biol Phys, Vol. 91, No. 3, pp. 571e578, 2015 0360-3016/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2014.10.050
Conflict of interest: none. Supplementary material www.redjournal.org.
for
this
article
can
be
found
at
572
Stemkens et al.
Introduction Radiation therapy treatment of pancreatic cancer is challenging because of motion, primarily induced by respiration. Conventionally, margins are used to ensure full-dose coverage of the target at the cost of extra radiation to nearby organs at risk (OARs). Prior knowledge about patient-specific motion, however, may lead to more conformal radiation plans (1). Currently, 4-dimensional (4D)-CT is the method of choice for motion characterization in radiation therapy. Unfortunately, CT has limited soft-tissue contrast, and 4D-CT often suffers from volume inconsistencies (eg duplicating, incomplete, or overlapping structures) caused by discrepancies between the breathing frequency and the pitch (2, 3). Magnetic resonance imaging (MRI), on the other hand, has superior soft-tissue contrast, allowing better differentiation between tumor and surrounding tissue. Moreover, the flexibility of image acquisition in MRI could be used to overcome some of the image artifacts typical of 4D-CT. Previous authors have reported MRI-based motion characterization using 2D-cine MRI (4-7). However, these methods solely provide tumor motion characteristics and do not provide information of the OARs. To construct high-resolution 4D-MR data, both prospective gated imaging and retrospective binning have been explored in previous studies. In prospective gated imaging, 3-dimensional (3D) volumes are acquired at different breathing depths (8), determined for example by a 1-dimensional (1D)-MRI navigator that records the diaphragm position (9). However, prospective gating may prolong scan time, and the probabilistic information about the duration of each respiratory phase is not captured. Consequently, retrospective binning methods have gained interest. To bin the acquired data into appropriate respiratory phases, a number of respiratory surrogates have been investigated, such as respiratory bellows (10), image-based body area (11), or 2-dimensional (2D) navigator slices (12). Cai et al (11) used 2D axial slices for retrospective slice stacking, which is closely related to 4D-CT and may therefore suffer from volume inconsistencies similar to those with 4D-CT. To address these disadvantages, Von Siebenthal et al (12) and Tryggestad et al (10) used retrospective stacking of 2D sagittal slices. Whereas Von Siebenthal et al used a custom 2D acquisition, Tryggestad et al used clinical sequences and a novel 2-pass sorting approach. In lieu of these multislice 2D methods, volumetric 3Deacquired MRI scans, in which the entire volume is excited each time, have been used by Buerger et al (13) for retrospective binning because they inherently have a higher signal-to-noise ratio (SNR) and no through-plane motion. However, the last 3 methods either require a custom read-out or long acquisition/reconstruction times, which is undesirable for clinical practice. Ideally, one would like to benefit from the higher SNR of 3D acquisitions while still using clinically available scan protocols.
International Journal of Radiation Oncology Biology Physics
The goal of this study, therefore, is to develop a robust 4D-MRI method, based on 3D-acquired MR data, to characterize respiratory-induced motion of pancreatic tumors and surrounding OARs using clinically available scan protocols within a clinically acceptable time frame. Two motion surrogates, which serve as input for the binning process, are assessed for their agreement with respiratoryinduced organ motion. Moreover, 2 MRI sampling strategies are explored in terms of robustness and image quality in multiple volunteers. The final resulting technique is applied to 6 pancreatic cancer patients. To demonstrate the applicability of this technique to radiation therapy, motion of both tumor and OARs are characterized using a 3D optical flow (OF) algorithm, because of its short processing time and minimal user intervention (14).
Methods and Materials We performed 2 different experiments in which the surrogate signal and MRI sampling strategy were independently optimized. The optimized protocol was then used to scan 6 pancreatic cancer patients. Three-dimensional motion was calculated for both the tumor and the duodenum over the complete respiratory cycle. All experiments were conducted on a 1.5 T Philips Achieva MR scanner (Philips Healthcare, Best, The Netherlands). Processing was performed using MATLAB (The Mathworks, Natick, MA).
Surrogate signal comparison: Correlation with pancreatic motion Six healthy subjects were scanned using a fast, coronal, 2D transient balanced steady-state gradient echo (balanced turbo field echo) sequence with radial read-out (a Z 30+ , time to echo/time to repetition 1.3/2.6 ms, field of view [FOV] 294 294 mm2, pixel size 1.5 1.5 mm2, slice thickness 7 mm, bandwidth 1580 Hz, turbo-factor 128, 500 dynamics [ie number of consecutive repeats of the 2D sequence], temporal frequency 2.6 Hz). The imaging plane was angulated parallel to the spine to sample along the principal axis of pancreatic motion, thereby reducing through-plane motion. The image acquisition was interleaved with a 1D MRI navigator, which was positioned at the dome of the right hemidiaphragm and processed online using the vendor’s reconstruction method. The respiratory bellows, provided by the MRI vendor, were placed on the upper abdomen. While the navigator is inherently synchronized with the motion in the 2D images, the bellows signal required minor manual synchronization, owing to imperfect temporal synchronization with the MR acquisition. Pixel-wise nonrigid displacement was calculated on the 2D time series using a 2D OF algorithm (15-17). In short, the OF algorithm calculates nonrigid displacement between 2 imaging frames according to intensity gradients, with an additional constraint on motion smoothness to model
Volume 91 Number 3 2015
Optimizing 4D-MRI for motion analysis
Table 1 Acquisition parameters for the in vivo experiments for 2 different in-plane sampling strategies, with the number of healthy volunteers scanned using the corresponding protocol Parameter
Cartesian
Radial
Time to echo (ms) Time to repetition (ms) a (o) FOV (mm3) Matrix size Dynamics Tacq (min) Bandwidth (Hz) No. of subjects
1.45 3.0 30 330 250 64 152 126 16 50 10:13.7 1780 5
1.45 3.0 30 330 330 96 152 152 24 25 8:35.1 1780 10
elastic organ deformation. To assess the agreement of the 2 surrogates with the pancreatic motion, we performed a general linear model (GLM) analysis, which is a leastsquares fit to the data described by YZX b þ ε. Here, Y is the pixel-based cranio-caudal (CC) motion, obtained from the pixel-by-pixel vector field calculated by the 2D OF algorithm, b the parameter estimates, X the surrogate signal, and ε represents the residuals. This analysis was performed on each pixel independently. The root mean square of the residuals was calculated and plotted in a spatial residual error map to find the surrogate with the lowest residuals. Furthermore, power spectrum analysis was performed on the surrogate signals and the pancreatic motion to gain insight into the agreement in frequency range of the surrogates with the pancreatic motion.
In vivo experiments Eleven healthy volunteers provided written informed consent to be scanned according to institutional rules to test the
A
573
feasibility and robustness of the proposed 4D-MRI technique. A clinically available 3D balanced turbo field echo (bTFE) with Cartesian or radial in-plane sampling was used, utilizing the parameters described in Table 1. Turbo direction was along kz and the shot length equal to the amount of slices, thus acquiring all kz lines within a single shot (Fig. 1a). Before each shot, a 1D MR navigator, placed at the dome of the right hemidiaphragm, was acquired. Twenty-five dynamics (ie the number of repeats of a 3D volume acquisition) were acquired for the radial protocol as a concession between imaging time and filling. For the Cartesian protocol 50 dynamics were acquired to assure sufficient data were acquired at the center of k-space. Four volunteers were scanned with both radial and Cartesian protocols. Because of time constraints, 6 volunteers were scanned using solely the radial protocol (from which 2 had a smaller FOV of 330 330 64 mm3), whereas 1 subject was scanned using only the Cartesian protocol. Reconstruction was performed offline on an 8-core CPU computer using ReconFrame (Gyrotools, Zurich, Switzerland), which is a commercial MATLAB-based MRI reconstruction software package with the ability to operate on the MRI scanner. Using the respiratory waveform derived from the navigator, k-space data were sorted into 10 respiratory phases using phase binning (ie equal-timeebased binning), similar to 4D-CT phase binning. Complex averaging was performed to increase SNR, when certain k-space segments were acquired multiple times. In case no k-lines were assigned, the lines were left blank. Image quality was assessed both visually and using the structural similarity (SSIM) index (18), which calculates the similarity between 2 images as a ratio between 0 (no similarity) and 1 (perfect similarity). The SSIM values were calculated between undersampled images and the full reference (ie 50 dynamics for Cartesian, 25 dynamics for radial).
B
C Registered images
3D images
r2,1 r1,1 I2 kz
D2
Reg 2 I1
D3
Reg 3
I3
r1,n
r2,1
r 1,1
D10
K1 r2,n
r1,n
K 10 kz
kz
10 respiratory phases (k-spaces) Respiratory phases
k-space read-out
Graphical overview phase binning method
I10
Reg 10
Calculate 4D deformation field and registered images
Fig. 1. Workflow of the in vivo experiments. (a) Cartesian and radial trajectories with turbo direction along kz. First, readout lines r1,1 . r1,n are acquired within 1 turbo field echo shot. In the next shot, r2,1 . r2,n are acquired, etc. (b) Graphic overview of the phase binning method for a respiratory signal. (c) After reconstruction, 3-dimensional images are registered to the reference volume, which results in a 4-dimensional deformation field.
574
International Journal of Radiation Oncology Biology Physics
Stemkens et al.
4D nonrigid motion analysis of tumor and OARs Six patients with an unresectable pancreas tumor were scanned using the radial protocol described in Table 1 according to institutional rules after written informed consent. After phase binning and image reconstruction, nonrigid displacement between the respiratory phases was estimated using a multithreaded optimized 3D OF algorithm (14). All respiratory phases were registered to 1 reference phase (the first, exhale phase), as can be seen in Figure 1c. The tumor and OARs were delineated on the reconstructed reference volume, and the mean motion within each delineated structure was calculated for all respiratory phases to construct motion trajectories of the tumor and duodenum over 10 respiratory phases.
The residual error within the pancreas over the 6 subjects is 2.4 1.1 times higher for the bellows compared with when the navigator is used. Moreover, the respiratory bellows signal required manual synchronization. Figure 2b displays the power spectra of the pancreatic motion, and the spectra of the 2 surrogates for the first 2 subjects shown in Figure 2a. Aside from large respiratory peaks between 0.2 and 0.32 Hz in all spectra, the power spectrum of the respiratory bellows of subject 2 displays additional peaks close to 1 Hz, owing to cardiac motion. Moreover, low frequencies (
