A statistical study of the factors influencing the extent of respiratory motion blur in PET imaging
Introduction
Respiratory motion is a main source of degradation in thoracic positron emission tomography (PET) and single photon emission computed tomography (SPECT) imaging due to the long scanning time of imaging system. The respiratory motion blur in PET or SPECT images results from the dispersed radiotracer activity over an area proportional to the breathing motion magnitude [1]. Studies have shown that the respiratory motion blur is harmful both to diagnosis and PET/CT guided radiotherapy: it not only degrades PET images by reducing spatial resolution, tumor-to-background contrast and signal-to-noise ratio (SNR) level, distorting the shape and location of the tumor [2], [3], but also affects the accuracy of quantitation, reduces the measured standard uptake value (SUV), causes the tumor size to be overestimated and thus increases the planned target volume [4], [5].
The breathing magnitude is known as one of the major factors to affect the extent of the respiratory motion blur. The effective spatial resolution of PET imaging is depended on the intrinsic resolution full width at half maximum (FWHM) and the amplitude of breathing and can be computed as [6]. It was found that the thoraco-abdominal tumors present displacements in the range of 6–23 mm mainly in the super-inferior and posterior–anterior direction, depending on their locations [2]. Respiratory motion blur observed at the lung/diaphragm interface is normally more serious than other locations since the diaphragm moves with the largest breathing magnitude among all organs. In [7] the average magnitude of lung tumor motion in the lower lobes was found to be 12 mm. Studies in [8] showed that a 20 mm motion magnitude can result in a loss of average activity between 25% and 45%. Besides motion magnitude, lesion size is another major factor influencing the severity of respiratory motion blur, small lung tumors or pulmonary nodules are inclined to suffer from serious blurring artifacts thus is hard to be detected in PET imaging [9], [10].
Previous studies mostly revealed the factors influencing respiratory motion blur by visual comparing the original blurred images with motion corrected or deblurred images [4]. Respiration gated PET/CT acquisition in conjunction with reconstruction or registration based motion correction algorithms are normally used to correct the motion blur [3], [10], [11], [12]. In gated acquisition the breathing magnitude of internal lesions can be estimated from the displacement of external markers recorded by tracking devices [7], [13], [14]. A post-reconstruction deconvolution was also proposed in [15], [16] to reduce the motion blur directly in ungated PET image domain, where the respiratory motion and corresponding point spread function (PSF) was estimated from gated CT.
The purpose of this study is to statistically show the factors influencing the extent of respiratory motion blur in PET images by regression analysis. The main usage of obtained regression equations is to estimate the effect of breathing on thoracic PET or SPECT imaging. The loss of mean intensity (LMI) in the region-of-interest is used as an index to represent the extent of motion blur. The rest of this paper is organized as follows: in Section 2 we first propose the regression models linked with LMI, tumor motion magnitude, tumor size and location and patient gender and height, then we introduce the method of blur identification and image restoration, which is used to estimate the respiration motion in image domain and derive the LMI. Section 3 is the results and discussions; the regression equations obtained from simulation data was tested by phantom SPECT data degraded with known motion, then we extend the work to clinical PET data to obtain meaningful regression equations. Section 4 gives conclusions.
Section snippets
Regression models
In general, there exists certain quantitative correlation between the image motion blur, tumor motion magnitude and tumor size. In nuclear medical imaging, PET or SPECT images corrupted with blurring artifact represent a decreased intensity level (number of counts), spatial resolution, contrast and SNR. Since the variation of image intensity is easier to compute than other parameters, we adopt the percentage loss of mean intensity (LMI) in the region of interest (ROI) to describe the extent of
Simulation study
The simulation study is designed to theoretically validate the regression models (3), (4). The static synthetic image size is 128×128, the ROI is a centered square object with uniform intensity and the background is zero intensity (square object is selected since the LMI caused by simulated motion blur can be more accurately computed than ROI in other shape). To investigate the impact of blur extent L and ROI size on motion blur, the square size is changed from 1×1 to 24×24, and L varies from
Conclusions
We presented a statistical study to quantitatively show the factors influencing the extent of respiratory motion blur in PET imaging. We proposed regression models linked with the motion blur induced loss of mean intensity (LMI), tumor motion magnitude, tumor size, patient gender and patient height. The blur identification and image restoration technique was employed to estimate the tumor motion and compute the LMI. Regression results were shown consistent in simulation, phantom and clinical
Conflict of interest statement
We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work; there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be constructed as influencing the position presented.
Acknowledgment
This work was supported by the National Natural Science Foundation of China (Grant no. 30730036).
References (26)
- et al.
Respiratory motion in positron emission tomography for oncology applications: problems and solutions
Nucl. Instrum. Methods Phys. Res. Sect. A: Accel., Spectrom., Detectors Assoc. Equip.
(2006) - et al.
Precise and real-time measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy
Int. J. Radiat. Oncol. Biol. Phys.
(2002) - et al.
Reduction of motion artifacts for PET imaging by respiratory correlated dynamic scanning
Nucl. Instrum. Methods Phys. Res. Sect. A: Accel., Spectrom., Detectors Assoc. Equip.
(2006) - et al.
Improving 3D PET imaging by restoration: a phantom study
Comput. Med. Imaging Graphics
(2005) - et al.
Identification of blur parameters from motion blurred images
Graphical Models Image Process.
(1997) - et al.
The CT motion quantitation of lung lesions and its impact on PET-measured SUVs
J. Nucl. Med.
(2004) - et al.
Respiratory motion correction in 3-D PET data with advanced optical flow algorithms
IEEE Trans. Med. Imaging
(2008) - et al.
Effect of respiratory gating on reducing lung motion artifacts in PET imaging of lung cancer
Med. Phys.
(2002) - et al.
Quantitation of respiratory motion during 4D-PET/CT acquisition
Med. Phys.
(2004) Respiratory motion handling is mandatory to accomplish the high-resolution PET destiny
Eur. J. Nucl. Med. Mol. Imaging
(2008)
Respiratory gating for 3-dimensional PET of the thorax: feasibility and initial results
J. Nucl. Med.
Implementation of an automated respiratory amplitude gating technique for PET/CT: clinical evaluation
J. Nucl. Med.
Computer-assisted quantification of lung tumors in respiratory gated PET/CT images: phantom study
Med. Biol. Eng. Comput.
Cited by (0)
- 1
Contributed equally to this work.