Guoping Chang

Implementation of an automated respiratory amplitude gating technique for PET/CT: Clinical evaluation

Amplitude gating techniques have recently been shown to be better at suppressing respiratory motion artifacts than phase gating. However, most commercial PET/CT scanners are equipped with phase gating capabilities only. The objective of this article was to propose and evaluate using patient studies an automated respiratory amplitude gating technique that could be implemented on current whole-body PET/CT scanners. A primary design feature of the proposed technique is to automatically match the respiratory amplitude captured during the CT scan with a corresponding amplitude during the PET scan. Methods: The proposed amplitude gating technique consists of a CT scan, followed by a list-mode PET scan. The CT scan was acquired while the patients respiratory motion was recorded by a monitoring device that determined the respiratory motion amplitude captured during the CT scan. A program was designed to inject triggers into the PET list stream whenever the patients respiration crossed a preset amplitude range determined by the captured amplitude during CT. To implement this proposed amplitude gating technique in whole-body PET/CT, a PET-first protocol was necessary to minimize the respiratory baseline drift between the CT and PET scans. In this implementation, a regular PET scan was first acquired over the patients whole body but excluding the bed position that covered the lesion of interest. The whole-body CT scan was then acquired, followed by a list-mode PET acquisition over the bed position that covered the area of interest (lesion). The proposed amplitude gating technique was tested using 13 patients with 21 lung or thoracic tumors. Results: In the patient studies, the gated images—when compared with the ungated images—showed statistically significant improvements, with an average 27% and 28% increase in maximum and mean standardized uptake value, respectively, for all lesions. Furthermore, the tumors in the gated images showed better contrast using visual inspection and line profiles. Conclusion: The implementation of the proposed respiratory amplitude gating technique on current PET/CT scanners is feasible, and amplitude-matched CT and PET data can be automatically generated using our proposed procedures without requiring patients to hold their breath or increase their radiation exposure.

Reliability of predicting image signal-to-noise ratio using noise equivalent count rate in PET imaging

PURPOSE Several investigators have shown that noise equivalent count rate (NECR) is linearly proportional to the square of image signal-to-noise ratio (SNR) when PET images are reconstructed using filtered back-projection. However, to our knowledge, none have shown a similar relationship in fully 3D ordered-subset expectation maximization (OSEM) reconstruction. This paper has two aims. The first is to investigate the NECR-SNR relationship for 3D-OSEM reconstruction using phantom studies while the second aim is to evaluate the NECR-SNR relationship using patient data. METHODS An anthropomorphic phantom was scanned on a GE Discovery-STE (DSTE) PET∕CT scanner in 3D mode with an initial activity concentration of 66.34 kBq∕cc. PET data were acquired over the lower chest∕upper abdomen region in dynamic mode. The experiment was repeated with the same activity concentration on a GE Discovery-RX (DRX) scanner. Care was taken to place the phantom at identical positions in both scanners. PET data were then reconstructed using 3D Reprojection (3D-RP) and 3D-OSEM with different reconstruction parameters and the NECR and SNR for each frame∕image were calculated. SNR(2) was then plotted versus the NECR for each scanner, reconstruction method and parameters. In addition, 40 clinical PET∕CT studies from the two scanners (20 patients∕scanner) were evaluated retrospectively. The patient studies from each scanner were further divided into two subgroups of body mass indices (BMI). Each PET study was acquired in 3D mode and reconstructed using both 3D-OSEM and 3D-RP. The NECR and SNR of the bed position covering the patient liver were calculated for each patient and averaged for each subgroup. Comparisons of the NECR and SNR between scanner types and BMIs were performed using a t-test and a p value less than 0.05 was considered significant. RESULTS Phantom results showed that SNR(2) versus NECR was linear for 3D-RP reconstruction across all activity concentration on both scanners, as expected. However, when 3D-OSEM was used, this relationship was nonlinear at activity concentrations beyond the peak NECR on both scanners. On the other hand, the plot of SNR(2) versus trues count rate was linear for 3D-OSEM across all activity concentrations on both scanners independent of reconstruction parameters used. In addition, for activity concentrations <30kBq∕cc, phantom results showed a higher SNR (by 12 ± 10%; p < 0.05) and NECR for the DRX scanner compared to DSTE for 3D-RP reconstruction. However, for 3D-OSEM reconstruction, these two scanners had similar SNRs (different by 2% ± 9%; p > 0.05), despite having different NECRs. Patient studies showed a statistically significant difference in NECR as well as the SNR for 3D-RP reconstruction between the two scanners. However, no statistically significant difference was found for 3D-OSEM. A statistically significant difference in both NECR and SNR were found between the different BMI subgroups for both 3D-RP and 3D-OSEM reconstructions. CONCLUSIONS For the scanners and reconstruction algorithm used in this study, our results suggest that the image SNR cannot be predicted by the NEC when using 3D-OSEM reconstruction particularly for those clinical applications requiring high activity concentration. Instead, our results suggest that image SNR varies with activity concentration and is dominated by the 3D-OSEM reconstruction algorithm and its associated parameters, while not being affected by the scanner type for the range of activity concentrations usually found in the clinic.

Effects of injected dose, BMI and scanner type on NECR and image noise in PET imaging

Noise equivalent count rate (NECR) and image noise are two different but related metrics that have been used to predict and assess image quality, respectively. The aim of this study is to investigate, using patient studies, the relationships between injected dose (ID), body mass index (BMI) and scanner type on NECR and image noise measurements in PET imaging. Two groups of 90 patients each were imaged on a GE DSTE and a DRX PET/CT scanner, respectively. The patients in each group were divided into nine subgroups according to three BMI (20-24.9, 25-29.9, 30-45 kg m(-2)) and three ID (296-444, 444-555, 555-740 MBq) ranges, resulting in ten patients/subgroup. All PET data were acquired in 3D mode and reconstructed using the VuePoint HD® fully 3D OSEM algorithm (2 iterations, 21(DRX) or 20 (DSTE) subsets). NECR and image noise measurements for bed positions covering the liver were calculated for each patient. NECR was calculated from the trues, randoms and scatter events recorded in the DICOM header of each patient study, while image noise was determined as the standard deviation of 50 non-neighboring voxels in the liver of each patient. A t-test compared the NECR and image noise for different scanners but with the same BMI and ID. An ANOVA test on the other hand was used to compare the results of patients with different BMI but the same ID and scanner type as well as different ID but the same BMI and scanner type. As expected the t-test showed a significant difference in NECR between the two scanners for all BMI and ID subgroups. However, contrary to what is expected no such findings were observed for image noise measurement. The ANOVA results showed a statistically significant difference in both NECR and image noise among the different BMI for each ID and scanner subgroup. However, there was no statistically significant difference in NECR and image noise across different ID for each BMI and scanner subgroup. Although the GE DRX PET/CT scanner has better count rate performance than the GE DSTE PET/CT scanner, this improvement does not translate to a lower image noise when using OSEM reconstruction. Our results show that patients with larger BMI consistently generate poorer image quality. Dose reduction from >555 to 296-444 MBq has minimal impact on image quality independent of the scanner used. A reduction in ID decreases patient and technologist exposure and can potentially reduce the overall cost of the study.

Optimization of super-resolution processing using incomplete image sets in PET imaging

Super-resolution (SR) techniques are used in PET imaging to generate a high-resolution image by combining multiple low-resolution images that have been acquired from different points of view (POVs). The number of low-resolution images used defines the processing time and memory storage necessary to generate the SR image. In this paper, the authors propose two optimized SR implementations (ISR-1 and ISR-2) that require only a subset of the low-resolution images (two sides and diagonal of the image matrix, respectively), thereby reducing the overall processing time and memory storage. In an N x N matrix of low-resolution images, ISR-1 would be generated using images from the two sides of the N x N matrix, while ISR-2 would be generated from images across the diagonal of the image matrix. The objective of this paper is to investigate whether the two proposed SR methods can achieve similar performance in contrast and signal-to-noise ratio (SNR) as the SR image generated from a complete set of low-resolution images (CSR) using simulation and experimental studies. A simulation, a point source, and a NEMA/IEC phantom study were conducted for this investigation. In each study, 4 (2 x 2) or 16 (4 x 4) low-resolution images were reconstructed from the same acquired data set while shifting the reconstruction grid to generate images from different POVs. SR processing was then applied in each study to combine all as well as two different subsets of the low-resolution images to generate the CSR, ISR-1, and ISR-2 images, respectively. For reference purpose, a native reconstruction (NR) image using the same matrix size as the three SR images was also generated. The resultant images (CSR, ISR-1, ISR-2, and NR) were then analyzed using visual inspection, line profiles, SNR plots, and background noise spectra. The simulation study showed that the contrast and the SNR difference between the two ISR images and the CSR image were on average 0.4% and 0.3%, respectively. Line profiles of the point source study showed that the three SR images exhibited similar signal amplitudes and FWHM. The NEMA/IEC study showed that the average difference in SNR among the three SR images was 2.1% with respect to one another and they contained similar noise structure. ISR-1 and ISR-2 can be used to replace CSR, thereby reducing the total SR processing time and memory storage while maintaining similar contrast, resolution, SNR, and noise structure.

Comparison between two super-resolution implementations in PET imaging.

Super-resolution (SR) techniques are used in PET imaging to generate a high-resolution image by combining multiple low-resolution images that have been acquired from different points of view (POV). In this article, the authors propose a novel implementation of the SR technique whereby the required multiple low-resolution images are generated by shifting the reconstruction pixel grid during the image reconstruction process rather than being acquired from different POVs. The objective of this article is to compare the performances of the two SR implementations using theoretical and experimental studies. A mathematical framework is first provided to support the hypothesis that the two SR implementations have similar performance in current PET/CT scanners that use block detectors. Based on this framework, a simulation study, a point source study, and a NEMA/IEC phantom study were conducted to compare the performance of these two SR implementations with respect to contrast, resolution, noise, and SNR. For reference purposes, a comparison with a native reconstruction (NR) image using a high-resolution pixel grid was also performed. The mathematical framework showed that the two SR implementations are expected to achieve similar contrast and resolution but different noise contents. These results were confirmed by the simulation and experimental studies. The simulation study showed that the two SR implementations have an average contrast difference of 2.3%, while the point source study showed that their average differences in contrast and resolution were 0.5% and 1.2%, respectively. Comparisons between the SR and NR images for the point source study showed that the NR image exhibited averages of 30% and 8% lower contrast and resolution, respectively. The NEMA/IEC phantom study showed that the three images (two SR and NR) exhibited different noise structures. The SNR of the new SR implementation was, on average, 21.5% lower than the original implementation largely due to an increase in background noise, while the NR image had averages of 18.5% and 8% lower SNR and contrast, respectively, versus the two SR images. The new SR implementation can potentially replace the original SR approach in current PET scanners that use block detectors while maintaining similar contrast and resolution, but at a relatively lower SNR. A major advantage of the new SR implementation is its shorter overall scan duration which results in an increase in scanner throughput and a reduction in patient motion.

Joint correction of respiratory motion artifact and partial volume effect in lung/thoracic PET/CT imaging.

PURPOSE Respiratory motion artifacts and partial volume effects (PVEs) are two degrading factors that affect the accuracy of image quantification in PET/CT imaging. In this article, the authors propose a joint motion and PVE correction approach (JMPC) to improve PET quantification by simultaneously correcting for respiratory motion artifacts and PVE in patients with lung/thoracic cancer. The objective of this article is to describe this approach and evaluate its performance using phantom and patient studies. METHODS The proposed joint correction approach incorporates a model of motion blurring, PVE, and object size/shape. A motion blurring kernel (MBK) is then estimated from the deconvolution of the joint model, while the activity concentration (AC) of the tumor is estimated from the normalization of the derived MBK. To evaluate the performance of this approach, two phantom studies and eight patient studies were performed. In the phantom studies, two motion waveforms-a linear sinusoidal and a circular motion-were used to control the motion of a sphere, while in the patient studies, all participants were instructed to breathe regularly. For the phantom studies, the resultant MBK was compared to the true MBK by measuring a correlation coefficient between the two kernels. The measured sphere AC derived from the proposed method was compared to the true AC as well as the ACs in images exhibiting PVE only and images exhibiting both PVE and motion blurring. For the patient studies, the resultant MBK was compared to the motion extent derived from a 4D-CT study, while the measured tumor AC was compared to the AC in images exhibiting both PVE and motion blurring. RESULTS For the phantom studies, the estimated MBK approximated the true MBK with an average correlation coefficient of 0.91. The tumor ACs following the joint correction technique were similar to the true AC with an average difference of 2%. Furthermore, the tumor ACs on the PVE only images and images with both motion blur and PVE effects were, on average, 75% and 47.5% (10%) of the true AC, respectively, for the linear (circular) motion phantom study. For the patient studies, the maximum and mean AC/SUV on the PET images following the joint correction are, on average, increased by 125.9% and 371.6%, respectively, when compared to the PET images with both PVE and motion. The motion extents measured from the derived MBK and 4D-CT exhibited an average difference of 1.9 mm. CONCLUSIONS The proposed joint correction approach can improve the accuracy of PET quantification by simultaneously compensating for the respiratory motion artifacts and PVE in lung/thoracic PET/CT imaging.

Design and performance of a respiratory amplitude gating device for PET/CT imaging.

PURPOSE Recently, the authors proposed a free-breathing amplitude gating (FBAG) technique for PET/CT scanners. The implementation of this technique required specialized hardware and software components that were specifically designed to interface with commercial respiratory gating devices to generate the necessary triggers required for the FBAG technique. The objective of this technical note is to introduce an in-house device that integrates all the necessary hardware and software components as well as tracks the patients respiratory motion to realize amplitude gating on PET/CT scanners. METHODS The in-house device is composed of a piezoelectric transducer coupled to a data-acquisition system in order to monitor the respiratory waveform. A LABVIEW program was designed to control the data-acquisition device and inject triggers into the PET list stream whenever the detected respiratory amplitude crossed a predetermined amplitude range. A timer was also programmed to stop the scan when the accumulated time within the selected amplitude range REACHED a user-set interval. This device was tested using a volunteer and a phantom study. RESULTS The results from the volunteer and phantom studies showed that the in-house device can detect similar respiratory signals as commercially available respiratory gating systems and is able to generate the necessary triggers to suppress respiratory motion artifacts. CONCLUSIONS The proposed in-house device can be used to implement the FBAG technique in current PET/CT scanners.

Determination of Internal Target Volume From a Single Positron Emission Tomography/Computed Tomography Scan in Lung Cancer

PURPOSE The use of four-dimensional computed tomography (4D-CT) to determine the tumor internal target volume (ITV) is usually characterized by high patient radiation exposure. The objective of this study was to propose and evaluate an approach that relies on a single static positron emission tomography (PET)/CT scan to determine the ITV, thereby eliminating the need for 4D-CT and thus reduce patient radiation dose. METHODS AND MATERIALS The proposed approach is based on the concept that the observed PET image is the result of a joint convolution of an ideal PET image (free from motion and partial volume effect) with a motion-blurring kernel (MBK) and partial volume effect. In this regard, the MBK and tumor ITV are then estimated from the deconvolution of this joint model. To test this technique, phantom and patient studies were performed using different sphere/tumor sizes and motion trajectories. In all studies, a 4D-CT and a PET/CT image of the sphere/tumor were acquired. The ITV from the proposed technique was then compared to the maximum intensity projection (MIP) volume of the 4D-CT images. A Dice coefficient of the two volumes was calculated to represent the similarity between the two ITVs. RESULTS The average ITVs of the proposed technique were 97.2% ± 0.3% and 81.0% ± 16.7% similar to the MIP volume in the phantom and patient studies, respectively. The average dice coefficients were 0.87 ± 0.05 and 0.73 ± 0.16, respectively, for the two studies. CONCLUSION Using the proposed approach, a single static PET/CT scan has the potential to replace a 4D-CT to determine the tumor ITV. This approach has the added advantage of reducing patient radiation exposure and determining the tumor MBK compared to 4D-CT/MIP-CT.

Implementation and optimization of a new Super-Resolution technique in PET imaging

Super-Resolution (SR) techniques are used in PET imaging to generate a high-resolution image by combining multiple low-resolution images that have been acquired from different points of view (POV). In this paper, we propose a new implementation of the SR technique (NSR) whereby the required multiple low-resolution images are generated by shifting the reconstruction pixel grid during the image-reconstruction process rather than being acquired from different POV. In order to reduce the overall processing time and memory storage, we further propose two optimized SR implementations (NSR-O1 & NSR-O2) that require only a subset of the low resolution images (two sides & diagonal of the image matrix, respectively). The objective of this paper is to test the performances of the NSR, NSR-O1 & NSR-O2 implementations and compare them to the original SR implementation (OSR) using experimental studies. Materials and Methods A point source and a NEMA/IEC phantom study were conducted for this investigation. In each study, an OSR image (256×256) was generated by combining 16 (4×4) low-resolution images (64×64) that were reconstructed from 16 different data sets acquired from different POV. Furthermore, another set of 16 low-resolution images were reconstructed from the same (first) data set using different reconstruction POV to generate a NSR image (256×256). In addition, two different subsets of the second 16-image set (two sides & diagonal of the image matrix, respectively) were combined to generate the NSR-O1 and NSR-O2 images respectively. The 4 SR images (OSR, NSR, NSR-O1 & NSR-O2) were compared with one another with respect to contrast, resolution, noise and SNR. For reference purposes a comparison with a native reconstruction (NR) image using a high resolution pixel grid (256×256) was also performed. Results The point source study showed that the proposed NSR, NSR-O1 & NSR-O2 images exhibited identical contrast and resolution as the OSR image (0.5% and 1.2% difference on average, respectively). Comparisons between the SR and NR images for the point source study showed that the NR image exhibited an average 30% and 8% lower contrast and resolution respectively. The NEMA/IEC phantom study showed that the three NSR images exhibited similar noise structures as one another but different from the OSR image. The SNR of the three NSR images were similar (2.1% difference) but on average 22% lower than the OSR image largely due to an increase in background noise, while the NR image had an average of 14.5% lower SNR versus the three NSR images. Conclusion The NSR implementation can potentially replace the OSR approach in current PET scanners while maintaining similar contrast and resolution, but at a relatively lower SNR. This NSR implementation can be further optimized as NSR-O1 & NSR-O2 implementations by using only a subset of low-resolution images which can achieve similar image contrast, resolution and SNR but require less processing time and memory storage. A major advantage of the NSR versus OSR implementation is its shorter overall scan duration which results in an increase in scanner throughput and a reduction of patient motion.

SU‐FF‐I‐72: Reducing PET Scan Duration By Improving SNR Using Super‐Resolution Techniques

Objectives: Super resolution (SR) techniques reconstruct a high resolution image from a series of low resolution images taken from different points of view of the same object. The aim of this abstract is to compare the SNR of imagesreconstructed with and without SR processing versus total scan duration in an attempt to produce images of similar quality but with shorter scan duration. Methods: A NEMA/IEC phantom containing 6 spheres of varying diameters (1–3.7cm) was filled with F‐18 water and scanned on a DSTE PET/CT scanner. The phantom was scanned in LIST mode for 4 min in 2D and 3D using three different sphere‐to‐background ratios (SBR) (3, 5, and 8). The LIST data was then rebinned into 24 different scan durations in 10 sec increments and reconstructed into 128*128 matrix using OSEM (2 iterations, 21 subsets). In addition, for each scan duration, a 128*128 SR image was generated from 4 64*64 images by offsetting the pixel grid by a 4mm along the X and Y axes. SNR was then determined by drawing ROIs on all spheres and background on all images with and without SR processing for all SBRs and scan durations in both 2D and 3D. Results: For the same scan duration, the noise in images with SR processing is reduced by 19–35% while SNR is improved by 36–40% for all SBRs and scan modes. For the same SNR, scan duration with SR processing can be reduced by 31–41% for different SBRs and scan modes. Conclusion: SR processing produces images with superior SNR and noise content thereby allowing a reduction in PET scan duration while maintaining similar image quality to standard reconstruction. One specific application of SR processing is to improve the quality of gated PETimages.