Browsing by Subject "PET"

Purpose: (1) To verify the accuracy and linearity of the ThermoScientific Radeye G Personal Rate Meter with respect to exposure rate across the full dynamic range of the instrument. (2) Use a combination of empirical data and Monte Carlo methods to estimate dose distribution in a GE Discovery 690 PET/CT scanner room and adjacent hallway. (3) Quantify components of occupational dose to PET technologists.

Materials & Methods: (Project 1) The Radeye unit and a calibrated ion chamber were placed in the beam of a Cesium 137 calibrator. They were exposed from 46 μR/hr to 1 R/hr with the pulse of each beam lasting for 90 seconds. The Radeye made 15 exposure rate measurements during each pulse. The ion chamber was read in the mid-point of each pulse's duration. (Project 2) Six Radeye units were placed at key points within the Discovery 690 scan room and two were placed in the adjacent hallway. 1600 exposure rate measurements were made over eleven hours during each day of operation. Data was collected for seven days. The total integrated data from the detectors inside the room was used to develop a Monte Carlo model of the room using FLUKA software. This model was then able to estimate the contribution from radiation escaping the scan room to the detectors in the hallway. (Project 3) Three PET technologists wore Radeye units while performing their daily tasks. The detectors recorded a mean exposure rate over each 25 second sampling period. The technologists were also asked to maintain a written log of all their interaction with radioactive material as well as their interactions with injected patients. Each day the Radeye unit produced a plot of radiation exposure with respect to time. Each interaction with radioactivity from the logs was highlighted on the plot and integrated to obtain the exposure received while performing that task.

Results: (Project 1) The Radeye deviated from the known value of exposure by up to 9.3% and deviated from the ion chamber measurement by up to 8.6% for exposure rates of 1 mR/hr and greater. The Radeye measured up to 29.6% higher than the known rate and up to 33.6% higher than the ion chamber measurement for exposure rates less than 1 mR/hr. The variance in the Radeye measurements decreased as exposure rate increased. The standard deviation of the Radeye measurements were less than 4% of their respective mean values for exposure rates less than 1 mR/hr. This value increased for lower exposure rates, up to 14% at 0.046 mR/hr. (Project 2) Mean daily exposures to five points in the PET/CT scan room were measured for CT and PET emissions separately. A Monte Carlo model of the scan room was created to model the distribution, including an initial approximation for the scanner gantry. The simulations showed that the virtual scanner should be thinner (i.e. less attenuating), especially for the 511KeV PET photons. (Project 3) The mean exposure received per dose draw and accompanying injection was 0.70±0.23mR for the 113 injections recorded over the course of the study. No correlation was observed between the dosage injected and the exposure received. The percent contributed to the total exposure by each category and participant was as follows. Technologist #1: 68% from Dose Draw, 6% from Patient Positioning, 4% from Patient Transport, 1% from General Patient Care, 21% from nonspecific sources. Technologist #2: 34%, 32%, 14%, 6%, and 14%. Technologist #3: 32%, 32%, 16%, <1%, and 20%. The dose draws and accompanying injections account for between one and two thirds of daily exposure. This indicates it is likely a 30% daily dose reduction could be achieved with use of automated injection equipment.

In radiation oncology, 18F-FDG Positron Emission Tomography (PET) is used for determining metabolic activity of cancers as well as delineating gross tumor volumes (GTV) for treatment planning. More recently, PET is being utilized for adaptive therapies for gynecological malignancies in which tumor response may be estimated and treatments adjusted during the course of radiation. In addition to treatment assessment, 18F-FDG PET has become a tool in the prediction of tumor response because of the derived Standard Uptake Value (SUV), a measure of the metabolic activity of a tumor. In this study, we seek to establish texture analysis as complimentary to SUV for predicting tumor response as well as understanding temporal changes during treatment in gynecological cancers. An additional experiment was performed studying the variability of texture features from baseline and intra-treatment PET scans due to reconstruction parameters in order to identify features that show statistically significant changes during treatment and that are independent of reconstruction parameters.

In this IRB approved clinical research study, 29 women with node positive gynecological malignancies visible on PET including cervical, endometrial, vulvar, and vaginal cancers are treated with radiation therapy. Prescribed dose varied between 45-50.4Gy, with a 55-70Gy boost to the PET positive nodes. A baseline, intra-treatment (between 30-36Gy), and post-treatment PET-CT were obtained with tumor response determined by a physician according to post-treatment RECIST. All volumes were re-contoured on the intra-treatment PET-CT. Primary GTVs were segmented both with the 40% SUVmax threshold method and a validated gradient-based contouring tool, PET Edge (MIM Software Inc., Cleveland, OH). A MATLAB Graphical User Interface (GUI) called Duke FIRE (Functional Imaging Research Environment) was developed for this study in order to calculate four mathematical algorithms representing the spatial distribution of pixels in an image: gray level co-occurrence matrix (GLCM), gray level run length matrix (GLRLM), gray level size zone matrix (GLSZM), and the neighborhood gray level difference matrix (NGLDM). Features representing characteristics of the image are derived from these texture matrices: 12 local features from the GLCM, 11 regional features from the GLRLM, 11 regional features from the GLSZM, and 5 local features from the NGLDM. Additionally, 6 global SUV histogram features including SUVmean, SUVmedian, SUVmax, skewness, kurtosis, and variance as well as metabolic volume (MV) and total lesion glycolysis (TLG) are extracted. The prognostic power of each baseline feature derived from both gradient-based and threshold segmentation methods was determined using the Wilcoxon rank-sum test. Receiver operating characteristic (ROC) curves were calculated to understand the sensitivity and specificity of baseline texture features compared to SUV metrics. Changes in features from baseline to intra-treatment PET-CT were determined using the Wilcoxon signed-rank test. A subset of 7 patient baseline and intra-treatment raw PET data was reconstructed 6 times using a TrueX+TOF algorithm on a Siemens Biograph mCT with varying iterations and Gaussian filter widths. Texture features were derived from the GTV as before. Texture features per patient were normalized to the respective clinical baseline value in order to limit variability to reconstruction parameters. Mean percent ranges of each feature at baseline and intra-treatment were determined and the change in features was compared using the Wilcoxon signed-rank test.

Of the 29 patients, there were 16 complete responders, 7 partial responders, and 6 non-responders. Comparing CR/PR vs. NR for the gradient-based GTVs, 7 texture values had a p < 0.05. The threshold GTVs yielded 4 texture features with p < 0.05. ROC and logistic regression was performed and texture features from both PET Edge and thresholding yielding a higher area under the curve (AUC) than SUV metrics. Features derived from PET Edge GTVs also showed higher AUCs than the threshold GTVs. From baseline to intra-treatment, 16 texture features changed with p < 0.05. Texture analysis of PET imaged gynecological tumors is considerably more powerful than SUV in early prognosis of tumor response, especially when using a gradient based method.

We then took the 16 texture features showing significant changes (p < 0.05) between baseline and intra-treatment PET scans in 29 patients and tested these against the subset of reconstructed features to determine if these changes were dependent upon the method in which the scans were reconstructed. A total of 13 features (including entropy, zone non-uniformity, and complexity) were found to be consistently different even when subjected to different means of reconstruction, however 3 of the 16 (inverse variance, run percentage, and zone percentage) were found to be dependent upon these reconstruction parameters. Texture features such as entropy, zone non-uniformity, and complexity are excellent candidates for future investigations of changes in texture analysis during radiation therapy of gynecological cancers. Caution should be taken with inverse variance, run percentage, and zone percentage due to their dependence upon reconstruction parameters.

This comprehensive work characterizes gynecological cancers using texture analysis in order to identify texture features that may be used for predicting tumor response as well as reflecting changes during treatment. It is the first study to our knowledge that utilizes all 4 texture matrices (GLCM, GLRLM, GLSZM, and NGLDM) and found 7 statistically significant features classifying responding and non-responding gynecological tumors: energy, entropy, max probability, zone gray level non uniformity, zone size non uniformity, contrast (NGLDM), and complexity. A novel method was implemented extending the NGLDM and its respective features to 3D space for this study. It is also the first study concluding that a semi-automatic gradient-based segmentation method results in more, stronger predictors than using a 40% SUVmax threshold method. Finally, this is the first study to examine variability of texture features with reconstruction parameters and to identify texture features as reliable and independent of reconstruction. In conclusion, texture analysis is a promising method of characterizing tumors visible on PET and should be considered for future studies.

Head motion during a Positron Emission Tomography (PET) brain scan can considerably degrade image quality. External motion-tracking devices have proven successful in minimizing this effect, but the associated time, maintenance, and workflow changes inhibit their widespread clinical use. List-mode PET acquisition allows for the retroactive analysis of coincidence events on any time scale throughout a scan, and therefore potentially offers a data-driven motion detection and characterization technique. An algorithm was developed to parse list-mode data, divide the full acquisition into short scan intervals, and calculate the line-of-response (LOR) midpoint average for each interval. These LOR midpoint averages, known as “radioactivity centroids,” were presumed to represent the center of the radioactivity distribution in the scanner, and it was thought that changes in this metric over time would correspond to intra-scan motion.

Several scans were taken of the 3D Hoffman brain phantom on a GE Discovery IQ PET/CT scanner to test the ability of the radioactivity to indicate intra-scan motion. Each scan incrementally surveyed motion in a different degree of freedom (2 translational and 2 rotational). The radioactivity centroids calculated from these scans correlated linearly to phantom positions/orientations. Centroid measurements over 1-second intervals performed on scans with ~1mCi of activity in the center of the field of view had standard deviations of 0.026 cm in the x- and y-dimensions and 0.020 cm in the z-dimension, which demonstrates high precision and repeatability in this metric. Radioactivity centroids are thus shown to successfully represent discrete motions on the submillimeter scale. It is also shown that while the radioactivity centroid can precisely indicate the amount of motion during an acquisition, it fails to distinguish what type of motion occurred.

Purpose: To characterize the effect of different discretization techniques (methods and values) on radiomic features extracted from positron emission tomography (PET) images of patients undergoing definitive radiotherapy for oropharyngeal cancer (OPC) and determine if there are optimal binning techniques associated with the computed texture and histogram measurements.Methods: 71 patients were enrolled in a prospective clinical trial to receive definitive radiotherapy (70Gy) for OPC. PET/CT images were acquired both prior to treatment and two weeks into treatment (i.e., after 20 Gy). All patients were scanned on the same PET/CT imaging system. The gross tumor volume at the primary tumor site was manually segmented on CT and transferred to PET, from which 74 quantitative radiomic features were extracted as potential imaging biomarkers. The sensitivity of feature extraction to common discretization techniques (fixed bin number vs. fixed bin size) was systematically evaluated by measuring radiomic feature values at monotonically increasing bin numbers (32, 64, 128, 256) and bin sizes (0.1, 0.5, 1.0, 5.0). Disparities in radiomics data parameterized by these different discretization settings were quantified based on t-tests of individual features and cross-correlation of matrix-level feature spaces. A discretization invariance score (DIS) was defined as an aggregation of each unique probability of rejecting the null hypothesis that any two discretization techniques produce the same feature value. To evaluate the generalization of these characteristics during treatment, DIS values were compared between pre- and intra-treatment imaging. Results: Only 50% of radiomic features were robust (DIS > 0.7) to changes in bin number, compared to 66% of features when varying bin size. Regardless of discretization technique, grey level variance (DIS=0.0) and high grey level size emphasis (DIS=0.21) were the most sensitive to binning perturbations, while skewness (DIS=1.0) and kurtosis (DIS=1.0) were nearly invariant. The cross-correlation between discretization-specific feature spaces was maximized for fixed bin number and minimized for fixed bin size. Ranked DIS measurements were comparable between pre-treatment and intra-treatment imaging, implying that feature sensitivity is invariant to changes in the absolute feature value over time. Conclusion: The impact of discretization is largely feature-dependent. Individual features demonstrated a non-linear response to systematic changes in discretization parameters, which was captured by our DIS metric. DIS values can be used to optimize downstream radiomic biomarkers, where the prognostic value of individual features may depend on feature-specific discretization.

Coincidence Timing Calibration (CTC) is an essential part of ensuring proper PET scanner function. The purpose of CTC is to account for timing differences in detector modules. The importance and precision in which this calibration needs to work is even more stringent for Time-of-Flight (TOF) PET. In this work, we looked to investigate the CTC process by which the TOF capable GE PET/CT Discovery-690 (D690) operates. Currently, it uses a 68Ge rotating pin source (RPS) to perform the calibration. The purpose of this work was to investigate the use of a centrally located source to perform the calibration. The timing resolution of the D690 was determined and used as a metric to evaluate both methods.

Two cylindrical 18F filled phantoms of 7.5 and 10 cm diameter were used to perform the CTC. The RPS and system table motion had to be disabled in order to use the centrally located sources in the CTC. All CTCs started with the default calibration file in place. Iterations of the CTC were performed until convergence of the calibration was observed on the review screen. Even after convergence, more iterations were performed for further analysis. At the end of the CTC with the centrally located sources, a follow-up iteration with the RPS was performed to see what adjustments would be made. Next, the timing resolution of the system was measured using a 68Ge line source. An apparatus with known locations to support the source allowed for the evaluation of the timing resolution off the central axis. The importance of this was that it allowed for non-centrally located lines of response to be evaluated. Furthermore, the timing resolution was measured with specific calibration files enabled that corresponded to particular iterations. In addition, a novel way of measuring the timing resolution (propagated method) for a particular calibration result without an actual measurement with that calibration enabled was developed and implemented. This greatly reduced the number of resolution measurements needed, which was particularly helpful for evaluating the improvement for each iteration.

The timing resolution of the system improved as more iterations were done. The difference between the propagated and measured timing resolution was under 2% most of the time. The cases in which the discrepancy was larger than 2% corresponded to one of the first iterations performed. After 15 iterations were performed for both centrally located scanners, the timing resolution of the system was measured through propagation to be 610 ps. The 15 iterations amounts to 15 minutes of acquisition time. After one iteration for the RPS, the timing resolution was measured to be 585 ps (587 ps in the propagated measurement). The single iteration of the RPS corresponded to 8 minutes of acquisition time. When following up the final iterations of the centrally located sources with the RPS, there was a change observed that improved the timing resolution to that measured after only one iteration of the RPS.

Conclusively, trends in the data showed that the centrally located sources did bring opposing detectors into good timing alignment with one another. These trends also indicated that the current CTC algorithm is not optimized for centrally located sources for the diameters tested. Finally, the method of propagating the change in calibration files illustrated a new CTC process method.

Problem: PET system performance, particularly the count rate-related effects, depends on a variety of effects including the patient size and the amount and distribution of radioactivity in the patient. The performance also depends on the particular PET system. This is primarily due to differences in detector material and detector size. This leads to a difference in image quality for the same activity level for different detectors. The current activity dosing protocol in Duke University’s Cancer Center is weight-based and system-independent, even though the systems vary greatly in count rate capability. This protocol might not be the most optimal protocol given that patients of the same weight are given the same dose but would produce different image qualities depending on the system they were scanned on. The work done in this thesis explores the components of the dosing protocol in an effort to reconsider the patient- and system- specific dosing needs for optimal image quality. This study uses Noise Equivalent Count (NEC) curves to simulate image quality for data that has been acquired using different systems, body sizes and shapes, and activity levels.Methods: This study investigates the behavior of three different hybrid PET/CT systems: the GE Discovery 690 (D690), the GE Discovery IQ (DIQ), and the GE Discovery MI (DMI). Phantom data were used to understand the performance of the three systems, and existing patient data were used to further evaluate the effects that different body characteristics have on each system. Two phantoms were used in this study: a whole-body phantom that simulates a medium- large patient and a smaller cylindrical phantom that simulates an extreme case of a small object. Both phantoms were filled with a large amount of activity (about 18 mCi for the whole-body phantom and about 12 mCi for the smaller cylindrical phantom) and thoroughly mixed before being scanned repeatedly for a long duration on all three systems to test each system’s behavior with different-sized phantoms. A Bash script was run to collect information from the phantoms’ DICOM headers so that NEC formulas could be calculated, and NEC curves could be analyzed. The dead-time model was adjusted to best fit the simulated data to the actual data to potentially improve accuracy with patient data. The phantoms were used to analyze the systems’ general behaviors without any human factors such as different uptakes for different organs, a larger variety of shapes and sizes, and different compositions. Once the general behaviors were understood and the models were adjusted, a large selection of patient data (500 for the DIQ and 500 for the DMI) was obtained. This was accomplished through the creation of multiple Bash and Python scripts that ran through patient data, retrieved the desired patients and patient scans based on specific criteria determined by the scripts, and collected anonymized data used to form NEC curves and experiment with body metrics. A few anonymized CT and PET images were saved for each patient as well so that body diameter measurements could be made. Results: It was determined that the NEC curves produced by the two different detector materials (BGO and LYSO) peaked at different activity levels for the same phantom. Also, to obtain the same NEC rate, the smaller cylindrical phantom required less activity than the whole- body phantom for each of the three systems. Dead-time data found in the image header was analyzed using Stearns’ NEC model, and his model appeared to consistently deviate from actual measurements. To improve the model, adjustments were made to parameters in the dead-time model to create a best fit to the phantom data, considering the three different systems and two phantom sizes. It was determined that a single dosing protocol may not be optimal for all systems since the NEC curves peaked at very different activities for each system and peaked at different counts per second for each system. Furthermore, the dosing protocol may not be benefiting patients of all sizes, as heavier patients may be receiving higher doses than needed for good image quality. Various body metrics were tested to compare which is the best to implement into an improved dosing protocol. These included body weight, BMI, and a pseudodiameter calculated from a cylindrical body approximation. This pseudodiameter was formed as an effort to approximate body diameters from patient weights and heights. The relationship between optimal dose (the dose at which peak image quality occurs) and the three body metrics was tested to determine whether a new dosing protocol can be formed based off of optimal doses depending on a certain body metric. It was determined that there is no correlation between optimal dose and the three body metrics. Conclusion: Body weight was concluded to be the most meaningful metric for calculating patient doses due to the ease of the measurement and the consistent relationship between image quality and patient weight for each system. Since patients with similar weights tend to produce similar image qualities, body weight can be used as a fairly reliable predictor of image quality when injected with a specific dose. Due to the differences in detection between the Discovery IQ and Discovery MI, the NEC curves produced by either system are very different, so the current dosing protocol would work best if it were system-dependent. Patients scanned on the DIQ could especially be receiving lower doses while still producing near-optimal image quality. If the goal of scanning patients is to produce the same image quality for every patient, then the dose for some patients could be significantly decreased. This can be concluded due to the NEC curves of patients at different body weights peaking at different count rates (where the lightest patients peak at higher count rates, and the heaviest patients peak at lower count rates). In this case, the current system-independent dosing calculation may not be optimal. A new dosing protocol was proposed. For the DIQ, patients would all be injected with 5.84 mCi. For the DMI, patients would be injected with a dose calculated by multiplying patients’ body weights by 0.06 mCi/kg, with a maximum injected dose of 11 mCi.

Background

Methods

Results

Conclusion

AbstractIntroduction: SPECT and PET are long established methods for functional imaging of bone lesions, including lesions in bone marrow and bone metastasis. These imaging modalities are however limited by poor spatial resolution which degrades quantitative accuracy and precise localization. This limitation in quantitative accuracy corresponds to the partial volume effect (PVE), in which a portion of the radiotracer activity truly in one structure appears, in the image, to be in nearby image voxels. To some extent PVEs can be corrected by iterative image reconstruction algorithms, such as ordered-subsets expectation maximization (OSEM), that model spatial resolution. This approach is however limited by noise, which is amplified as spatial resolution is recovered and PVEs are reduced. SPECT and PET imaging often involves CT as well. CT provides very high-resolution anatomical information which can be used to correct PVEs in SPECT and PET. One approach to PVE correction is using Markov Random Fields (MRFs) that incorporate anatomical information. However, there has been relatively little investigation on MRF-based PVE correction for SPECT/PET bone imaging using CT information. In this work, two types of CT anatomical information are considered: (i) partial anatomical information (pAI) which distinguishes, for example, compact bone from bone marrow but does not otherwise distinguish the tumor from surrounding tissue and (ii) full anatomical information (fAI), which fully distinguishes tumor from surrounding tissue. Image reconstructions involving pAI and fAI are referred to as RpAI and RfAI, respectively. RfAI is expected to provide the best correction of tumor PVEs, but RpAI may be more often available from CT images. The objective of the work is to assess the effectiveness of RpAI as compared to RfAI and OSEM.

Methods: Radiotracer (SPECT/PET) and attenuation coefficient (CT) phantoms were generated using XCAT software. Tumor lesions with high activity were added to the bone marrow in the radiotracer phantom. Two CT phantoms, pAI and fAI, were generated, with the fAI CT phantom including reduced CT number in the tumor-lesion locations. Projection data were simulated, and images were reconstructed using the computer code SPECT-MAP, with modeled spatial resolutions of 12mm (SPECT-like data) and 6mm (PET-like data). The RpAI and RfAI image reconstructions were performed using the iterative coordinate descent (ICD) algorithm and the Bowsher prior. The reconstructions were performed with projection data at 4 noise levels: 5M-, 50M-, and 100M-counts and noise-free. Reconstructed images were evaluated by visual inspection and by root-mean-square (RMS) error across the entire image and in 2 small ROIs (ROI-1 and ROI-2) surrounding the tumor lesions.

Results: The estimated rmsemin calculated from ROI-1 and ROI-2 reconstructed images of noisy (5M counts) projection data with res-12mm using OSEM, RpAI and RfAI were 0.92E-5 & 0.82E-5 (both at iteration 5, subset 9), 0.76E-5 & 0.65E-5 (both at OPS of 1.0E+4), and 0.44E-5 & 0.44E-5 (both at OPS of 1.0E+4), respectively; while for res-6mm, the rmsemin were 0.85E-5 & 0.81E-5 (both at iteration 10, subset 9), 0.57E-5 & 0.53E-5 (both at OPS of 1.0E+4), 0.37E-5 & 0.37E-5 (both at OPS of 1.0E+4), respectively. At both spatial resolutions, the RpAI reconstructions, using partial anatomical information only, provided reduced RMS errors compared to OSEM. Conclusions: At spatial resolutions characteristic of SPECT and PET, the partial anatomical information available from normal bone structures such as marrow and compact bone can improve estimation of hot-spot lesions, as measured by visual inspection and RMS error.

Gallium-68 labeled compounds have shown an important role in PET imaging of detecting neuroendocrine tumors and prostate tumors. As the prostate exists near the bladder, image quality of prostate tumors can be challenged by radioactivity concentration accumulated in the bladder. Previous studies have shown that activity measurements within 4 cm distal to the bladder were more affected by its higher 〖^18〗F concentration and improved with time of flight (TOF) technique. The aim of this research is to understand and compare the effects of 〖^68〗Ga and 〖^18〗F activity concentration on evaluation of structures closer to the bladder, and compare it for different scanners, acquisition modes and reconstruction techniques.

Methods: A bladder insert was placed in the center of an oval phantom with radioactivity at three different bladder-to-background ratios: 1:1, 40:1 and 80:1. Twelve 1-cm spheres representing 8:1 small lesions were symmetrically located in two axial planes, all with 1.2 cm distance to the bladder surface. The whole phantom was scanned on both the GE Discovery 690 and GE Discovery IQ, filled with 〖^18〗F and 〖^68〗Ga separately. In addition to default reconstruction settings, images were also reconstructed with nTOF and TOF modes on D690, and OSEM and REG modes on DIQ. The same ROIs were applied to all these images.

Results: Radioactivity in the bladder resulted in worse visualization and quantitation of the small spheres. The value difference among the 12 spheres became greater, and the lesion-to-background ratio were also affected by high bladder activity. The lesion value could be overestimated to be about three time as much as, or underestimated by 50% at most of its true measurement. The greatest difference was seen when the bladder was filled with 〖^68〗Ga. Both TOF and REG modes provided better lesion value variations compared with nTOF/OSEM reconstruction. Image contrast was significantly improved with more iterations.

Conclusions: Radioactivity of 〖^68〗Ga in the bladder has a similar but more serious effect than 〖^18〗F in the value measurements of surrounding structures. Greater measurement variations occur with higher bladder activity concentration. Iterative reconstruction with TOF information or REG reconstruction can be used to improve image quality, otherwise more iterations are recommended.

The use of Positron Emission Tomography (PET) has been suggested as a tool for quantitative biological measurement to determine outcomes of therapy, diagnosis, and novel drugs, through measures of tumor change from repeated PET scans. However, inherent variability (due to technical and biological effects) results in different measurements even where there is no change in the tumor. This study evaluates the random component of variability due to the limited number of counts acquired in the scans. We have proposed using PET raw list mode data as a way to determine the variability associated with the scanner, including the nonlinear processes like the max standard uptake value (SUVmax) and iterative reconstruction processes. PET simulation (digitally simulated oval phantom), PET list mode whole body (WB) and tapering phantom (TP) data in addition to clinical data (prostate cancer patients) were used to divide a larger acquisition into sequentially smaller half scan durations to compare their variabilities with the ideal Poisson variability in a PET system. Poisson statistics predicts that variability decreases as 1/sqrt(n) of the number of counts (or scan duration).

The WB phantom contained 21 spheres (six 3 cm, six 2 cm, nine 1 cm), the TP contained spheres (thirty-two 1 cm spheres and twenty-four 2 cm spheres) distributed over four levels. Simulated data was used as an ideal scenario with larger statistical power, and showed excellent agreement (<10%) with its Poisson calculation using 4 mm

of smoothing. Through the use of simulation and phantom data variability among measurements using SUVmax have been shown. This data has demonstrated that maximum ROI methodology on iteratively reconstructed images retains the Poisson nature of PET coincidence counts in spite of the potential nonlinearities of both the reconstruction method and the ROI methodology.

The purpose of this work was to develop a new radiomics paradigm for sparse, time-series imaging data, where features are extracted from a spatial-temporal manifold modeling the time evolution between images, and to assess the prognostic value on patients with oropharyngeal cancer (OPC).To accomplish this, we developed an algorithm to mathematically describe the relationship between two images acquired at time t=0 and t>0. These images serve as boundary conditions of a partial differential equation describing the transition from one image to the other. To solve this equation, we propagate the position and momentum of each voxel according to Fokker-Planck dynamics (i.e., a technique common in statistical mechanics). This transformation is driven by an underlying potential force uniquely determined by the equilibrium image. The solution generates a spatial-temporal manifold (3 spatial dimensions + time) from which we define dynamic radiomic features. First, our approach was numerically verified by stochastically sampling dynamic Gaussian processes of monotonically decreasing noise. The transformation from high to low noise was compared between our Fokker-Planck estimation and simulated ground-truth. To demonstrate feasibility and clinical impact, we applied our approach to 18F-FDG-PET images to estimate early metabolic response of patients (n=57) undergoing definitive (chemo)radiation for OPC. Images were acquired pre-treatment and two-weeks intra-treatment (after 20 Gy). Dynamic radiomic features capturing changes in texture and morphology were then extracted. Patients were partitioned into two groups based on similar dynamic radiomic feature expression via k-means clustering and compared by Kaplan-Meier analyses with log-rank tests (p<0.05). These results were compared to conventional delta radiomics to test the added value of our approach. Numerical results confirmed our technique can recover image noise characteristics given sparse input data as boundary conditions. Our technique was able to model tumor shrinkage and metabolic response. While no delta radiomics features proved prognostic, Kaplan-Meier analyses identified nine significant dynamic radiomic features. The most significant feature was Gray-Level-Size-Zone-Matrix gray-level variance (p=0.011), which demonstrated prognostic improvement over its corresponding delta radiomic feature (p=0.722). We developed, verified, and demonstrated the prognostic value of a novel, physics-based radiomics approach over conventional delta radiomics via data assimilation of quantitative imaging and differential equations.

The Standardized Uptake Value (SUV) is a method for semiquantitative evaluation of radiotracer accumulation on PET scans. Changes in SUV can be used to determine treatment response. However, SUV measurements are influenced by a variety of biological and technological factors, including image reconstruction parameters.

There are other semiquantitative metrics used in PET that relate to the total metabolic activity of a tumor. Current metrics of this type (e.g., Total Lesion Glycolysis) use a combination of SUV and an object volume. Such concentration-based metrics may not capture all radioactivity of an object. We propose a more direct method to assess total radiotracer uptake (TRU): the total radioactivity in a large VOI is measured and background is subtracted.

Phantom studies were performed to assess the effect of image reconstruction parameters on SUV, and to compare the TRU with concentration-based metrics. Patient images were evaluated to estimate the percent error of the TRU metric in imaging of humans.

Methods:

A whole body phantom with 1 cm hot spheres was scanned with a GE Discovery 690 PET/CT scanner, with time of flight (TOF) capability. Data were reconstructed several different ways to examine the effect of image matrix size, amount of smoothing, field of view (FOV) size, TOF vs. non-TOF reconstruction, iterations of reconstruction algorithm, and image matrix shift on SUV.

An additional whole body phantom was scanned on the same system to compare the accuracy and variability of the new TRU metric with existing measures.

Results:

Reconstruction parameters had substantial effects on SUV for 1 cm spheres. Varying the FOV from 35 to 70 cm produced an 11% change in average normalized SUV. Changing the image matrix size from 128x128 pixels to 256x256 pixels produced an 5.3% difference. Shifting the image matrix produced up to a 12% change in SUV. TOF vs. non-TOF reconstruction resulted in up to a 29% difference in SUV for two iterations.

The TRU method was more accurate than TLG and SUV for all sphere types in images with 0 mm to 10 mm of smoothing. Mean errors of TRU were between 1-12%. The TRU method was less variable than TLG in unsmoothed images with acquisition lengths of 1, 2, and 4 minutes. Coefficients of variation were between from 2-17% for TRU measurements, compared to 5-19% for TLG measurements. Simulation of TRU applied to human images shows potential error from 10-18% for 10:1 lesions 1-4 cm in diameter.

Conclusions:

Changes in image reconstruction parameters could significantly influence the SUV for small, 1 cm lesions. These effects are reduced for larger, 2.5 cm lesions.

TRU can accurately quantify small lesions in a phantom study. In some cases, TRU is less variable than TLG and SUV. Computer simulations of error in TRU when applied to human studies show low percentage errors for realistic tumor contrasts and volumes.

Quantitative measurements in PET imaging have recently become more widespread as a way to diagnose and stage many types of malignant cancer. Currently patients need to have follow-up scans performed on the same PET system due to technical factors. Multi-clinic studies using quantitative PET measurements are also confounded by these technological factors. This work aims to evaluate the use of commonly available phantoms to cross-calibrate processing parameters to equalize small lesion quantitation. The method was verified using an abdomen phantom with small hot sphere inserts, as well as a smaller phantom with small hot sphere inserts.

Methods: A GE Discovery 690 and STE were used. Both time-of-flight (TOF) and non-TOF images were used from the D690. Jaszczak phantoms with hot rod and cold rod inserts were scanned on both systems consecutively for 20 minutes. Images were reconstructed with a range of iterations and post-smoothed (PS) with 2-10 mm of smoothing. Automated analysis of the images used the CT images to find rods and then calculate a rod to background ratio for each rod sector, PET image variant, and scanner. A target rod contrast could then be chosen and parameters determined for both systems separately to equalize rod contrast. Iteration-based resolution control and PS were both evaluated. To verify, an abdomen phantom was filled with a low background activity and ten 10-mm diameter spheres filled with FDG and CT contrast. In order to evaluate any size dependence, six 10-mm diameter spheres filled with FDG and CT contrast were placed inside a Jaszczak container filled with low background activity. An automatic CT-based analysis of the spheres was performed, obtaining mean and maximum values across the spheres.

Results: Small sphere quantitation differed substantially for similar processing between systems. However, sphere quantitation matched well when cross-calibrating the DSTE and non-TOF D690 Jaszczak phantom images by independently limiting iterations. Doing the same process with post-smoothing yielded similar results, with high iteration PS performing slightly better than PS at iterations used clinically at Duke for twenty-minute scans. Equalizing TOF images from the D690 with DSTE images with spheres placed in an abdomen phantom resulted in relatively poor correlation, but correlated well with spheres placed inside the Jaszczak phantom. Shorter scan durations behaved similarly to the twenty-minute scans.

Conclusions: Both Jaszczak phantoms worked well for cross-calibrating processing parameters to equalize quantitation in small lesions for non-TOF imaging. Iterations and PS could both be used to control resolution. It appears the best method is to use PS to fine-tune the resolution. The size dependence of TOF, and PET in general, seems to be an issue.

Positron Emission Tomography with 18F-FDG can be used as a predictor for post-therapeutic tumor response in rectal and gynecologic cancers. An issue with assessing lesions in the vicinity of the bladder is the radioactivity that accumulates in the bladder due to constant filling. SUV analysis is used to discriminate between therapy responders and non-responders based on percentage change/threshold, but it is not yet known how much variability can be attributed to the bladder radioactivity. The purpose of this research is to understand the effects of bladder radioactivity on surrounding concentration measurements with different scanning and image reconstruction techniques.

Methods: ROI analysis was performed on 67 PET scans from DUMC. Typical values of bladder volume, radioactivity, radioactivity concentration and bladder-to-background uptake ratio were determined and incorporated into phantom studies. A bladder phantom insert was created for a 25 L torso phantom to explore effects of a bladder in the center of the FOV, on the edge of the FOV, and outside the FOV on the GE Discovery STE. The bladder insert was also used in a phantom study to assess the effects of different, realistic bladder radioactivity levels on surrounding 1-cm lesions on both the GE Discovery STE and GE Discovery 690. Background activity and in-air environments were explored, as well 2D, 3D and time-of-flight PET acquisitions with different image reconstruction techniques.

Results: The DSTE 2D PET data suffered large void artifacts that worsened as radioactivity in the bladder increased. The DSTE 3D PET data over-estimated background measurements up to 30 slices with the bladder outside the FOV. Concentration measurements within 4 cm distal to the bladder showed great variability and were generally recovered best with TOF PET or 3D PET with an increased number of iterations. Lesions greater than 4 cm distal to the bladder showed consistent recovery in both the 3D and TOF PET data.

Discussion: Radioactivity within the bladder has substantial effects on surrounding radioactivity concentration measurements. There are limits to the measurements of the radioactivity concentration values used in SUV analysis in the vicinity of the bladder.

A general theme is that more accurate results are produced with smaller amounts of radioactivity in the bladder. This validates the DUMC protocol that begins acquisition just below the pelvic area to reduce as much filling as possible and highlights the usefulness of patient voiding prior to scanning.

X-ray computed tomography (CT) forms the basis for attenuation corrected positron emission tomography (PET) using combined PET/CT scanners. With concerns of high radiation exposure to patients through widespread use of CT, the lowest photon flux that will provide uniform attenuation correction for PET to within 5% over a range of body sizes was investigated. Additionally, clinical uniformity measurements are performed on a uniform phantom, but their results may not be applicable as an estimate of error of hot lesions. PET simulations of variability and localized error were performed with and without hot lesions using a tapering phantom. Images were reconstructed using a variety of fixed and modulated tube-current CT scans and various levels of scatter correction. A physical phantom was designed and scanned to augment the simulation results. Attenuation correction of uniform images was within 5% error when using 120 kVp using a noise index of 50 and 140 kVp using a noise index of 50 for all phantom sizes. Variability with hot lesions was within 5% for scans using 120 kVp and greater than 24 mAs for 21.9 cm and 31.7 cm effective diameters and greater than 48 mAs for 38.5 cm effective diameter. Variability was worse in the background than on hot lesions for poor attenuation correction and poor scatter correction cases. Background error overestimates the error in hot lesions when attenuation correction is biased. Variability was within 5% when estimation of scatter magnitude was within 20% of its true value both with and without hot lesions. Errors in background due to under and overcorrected scatter lead to an over and underestimate of hot lesion errors, respectively. Physical phantom uniformity was within 5% when using 120 kVp and 10 mAs, albeit with a much smaller phantom size. The background error and its underestimation of lesion error was also measured in the physical phantom.

Purpose: Our primary goal was to examine the effect of PET image reconstruction parameters on baseline and early-treatment FDG-PET/CT quantitative imaging. Early-treatment changes in tumor metabolism in primary tumor and nodes can potentially determine if the patient is responding to therapy, but this assessment can change based on the reconstruction parameters. We investigated the effect of the following reconstruction parameters: number of Ordered-Subset-Expectation-Maximization (OSEM) iterations, post-reconstruction smoothing, and quantitative metrics (SUV-max, SUV-mean, SUV-peak).

A concurrent investigation explored in detail the sampling parameters of SUV-peak by way of a Monte Carlo digital phantom study. SUV-peak was proposed as a compromise between SUV-max and SUV-mean, in hope to retain key attractive features of these two metrics (inter-physician independence of SUV-max, noise-averaging of SUV-mean) but reduce unwanted errors (noise dependence of SUV-max, contour-dependence of SUV-mean). Sampling parameters have vaguely been defined, in particular, the scanning resolution (i.e. 1 voxel, 1/2 voxel, 1/4 voxel, etc.) of the SUV-peak spherical ROI . We examined the role that partial-voxel scanning plays in tumor SUV recovery in both noise-free and realistic OS-EM noise environments.

Materials and Methods: The response assessment investigation involved 19 patients on an IRB-approved study who underwent 2 baseline PET scans (mean-separation = 11 days) prior to chemoradiotherapy (70 Gy, 2 Gy/fraction). An intra-treatment PET scan was performed early in the course of therapy (10-20 Gy, mean = 14 Gy). The images were reconstructed with varying OS-EM iterations (1-12) and Gaussian post-smoothing (0-7 mm). Patients were analyzed in two separate groups, distinguished by the PET/CT scanner used to acquire data: (1) GE Discovery STE; and (2) Siemens Biograph mCT. For each combination of iterations and smoothing, Bland-Altman analysis was applied to quantitative metrics (SUV-max, SUV-mean, SUV-peak) from the baseline scans to evaluate metabolic variability (repeatability, R = 1.96&sigma). The number and extent of early treatment changes that were significant, i.e., exceeding repeatability, was assessed.

An original SUV-peak algorithm was developed, which measures SUV-max and SUV-peak for as small as 1/32 voxel scanning. Two rounds of digital phantoms were generated for the SUV-peak investigation. First, 10,000 spherical tumors were generated at a random matrix location for each diameter 1-4 cm and smoothed with an isotropic Gaussian, FWHM = 0.8 cm, then evaluated using the SUV-peak algorithm. Next, realistic body-sized phantoms were generated with background activity, and 1,000 spherical tumors of activity 4 time the background for each diameter (1-4cm) were placed inside (8 tumors per phantom, location randomized within certain constraints). These images received realistic corrections in projection space for attenuation, spatial resolution, and noise, were reconstructed with an in-house OS-EM algorithm, and then assessed using the SUV-peak algorithm. The mean recovered activity above background and its coefficient of variation were calculated for all metrics for each tumor size, for both simulations. For the realistic noise simulation, various levels of Gaussian smoothing was applied post-reconstruction, the effects summarized in plots showing coefficient of variation vs. mean recovered activity above background - a comparison of the effectiveness of SUV-max and SUV-peak.

Results: For the GE Discovery STE 2D cases averaged over all metrics (SUV-max, SUV-mean, SUV-peak) and structures (GTV, LN), repeatability, R, improved with increasing smoothing and decreasing iterations. Individually, SUV-mean repeatability was less affected by the number of iterations, but demonstrated the same relationship with smoothing. SUV-mean outperformed SUV-max and SUV-peak with regards to the number of cases exceeding repeatability, N. Considering R, N, and the sum of relative metric change outside repeatability, &Omega, averaged over all metrics and all structures, and normalized, several combinations of reconstruction parameters produced five optimal combinations above set thresholds: 1 iteration with 0.1-3.0 mm smoothing; and 2 iterations with 2.0-3.0 mm smoothing. Current GE 2D reconstruction protocol for HN cases uses 2 iterations and 3.0 mm post-smoothing, which lies on the edge, but within these recommendations.

The relationship between repeatability and number of iterations for the 3D cases was more complex; SUV-max demonstrated the best repeatability with 2 iterations, with both SUV-mean and SUV-peak reaching the best repeatability with 4 iterations. The same dependence on smoothing was noted, i.e. increased smoothing gives lover (desirable) repeatability. SUV-mean once again outperformed SUV-max and SUV-peak with regards to the number of cases exceeding repeatability, N. The calculations of N and &Omega averaged over all metrics were limited severely by the low number of cases, damaging the statistical significance of the following recommendation. Three optimal combinations with averaged and normalized R, N, &Omega, above a set threshold are recommended as most effective reconstruction parameter combinations: 4 iterations with 2.0-4.0 mm smoothing. Current Siemens 3D reconstruction protocol for HN cases uses 4 iterations and 3.0 mm post-smoothing, which lies within these recommended parameters. However, no statistically significant conclusions could be drawn from this analysis for this scanner, and performing similar data analysis on a larger patient pool is proposed.

The minimum spherical tumor diameter required for full recovery was 3.0-3.5 cm for SUV-peak, and 2.5-3.0 cm for SUV-max. SUV-max was found to overestimate the recovered value of tumors by up to 46% (vs. 10% for SUV-peak); above the minimum diameter for full recovery, SUV-peak values were significantly closer to actual tumor activity. Considering only the realistic noise tumors, the coefficient of variation for SUV-max ranged from 5.5-17.7%, whereas for SUV-peak these values were lower, 2.7-13.2%. Partial-voxel scanning did not substantially affect the coefficient of variation (<0.2%). Comparison of coefficient of variation vs. mean recovered value demonstrated that SUV-max with additional Gaussian smoothing outperforms SUV-peak by up to 0.8% for 1 cm tumors and 0.2% for 4 cm tumors. Other tumor sizes showed little difference between the two metrics.

Conclusion: For patients scanned on the GE Discovery STE using the HN protocol (2D acquisition mode), images reconstructed for quantitative analysis may benefit from a low number of OS-EM iterations (≤ 2). Some post-reconstruction smoothing proved to be beneficial (1.0 mm ≤ FWHM ≤ 3.0 mm), however, over-smoothing for the sake of more qualitatively appealing images or improved image quality metric (e.g. SNR, CNR) may prove detrimental to quantitative response assessment analysis. Our results for the Siemens Biograph mCT using the HN protocol (3D acquisition mode) demonstrated favor towards 4 iterations and limited range of smoothing (2.0 mm ≤ FWHM ≤ 4.0 mm). These results are statistically limited, further cases are necessary for any conclusive recommendations on reconstruction parameters.

SUV-peak was shown to reduce uncertainties associated with quantitative PET image analysis when compared directly to SUV-max. Above the minimum tumor diameter required for full recovery, SUV-peak also provides a better estimate of the actual tumor activity. However, initial comparisons of SUV-peak and SUV-max over various levels of additional Gaussian smoothing found SUV-max more favorable. Partial-voxel scanning of SUV-peak did not reduce the metric's coefficient of variation in images with realistic noise. Therefore, a phantom investigation is proposed to compare SUV-peak and SUV-max of real scanned images with various levels of post-smoothing, which may conclusively eliminate the need for SUV-peak.

Positron Emission Tomography is a coincidence-detection-based nuclear imaging modality that has increased in clinical prevalence over the last two decades. Measures have recently been taken to improve the practice, specifically the synergistic combination with CT, and implementation of iterative reconstruction. The time-of- flight (TOF) technique is another improvement theorized early in PET development, which reduces image noise by measuring the difference in coincident photon detection times. It was difficult to implement at the time of inception because of limited technologies, but better detectors and electronics have recently made TOF feasible for clinical use. Its gain in image quality has been measured by various methods, but is difficult to quantify because of tradeoffs inherent in count-based imaging. This work set out to investigate the image quality gained with TOF imaging by determining the effective non-TOF scan time required to achieve equivalent image quality as TOF.

Methods: We used the TOF-capable GE Discovery 690 PET/CT scanner with ~600 ps timing resolution to acquire high-count list-mode data of hot spheres, cold bottles, and a novel low-contrast bead insert housed in three phantoms of increasing diameters. These data were reconstructed with and without TOF information into shorter images of 30 sec, 1, 2, 4 and 8 min, using the OS-EM reconstruction algorithm with 16 subsets and 1, 2, 3, 5 and 10 iterations each. Up to 16 replicates of each image were produced. Regions of interest were drawn on the high-count images and subsequently applied to all images in each set. These data were averaged across the replicate image sets for statistical power and were used to calculate contrast, background variability and replicate variability for regions within each phantom scan. Background variability was measured as the standard deviation of 1 cm ROI means spread throughout the background, while replicate noise was measured as the pixel deviation across replicated images. The contrast for each unique phantom region and scan time were plotted versus the two noise measures, and a unique quantification method was devised to calculate the scan time equivalent for images reconstructed with TOF versus those without.

Results: Visual evaluation showed universal improvement in image quality. Hot spheres were more easily resolved, cold regions were colder, and the low-contrast phantom became clearer overall. Gains were also higher as a function of phantom size. Plotting contrast versus the two variability measures demonstrated greater gains for larger phantoms than small.

The quantification method delivered easily interpretable results that correlated with visual and graphical evaluation. Hot spheres showed between 1.6× and 2.5× scan time gain factor, while cold bottles showed between 3.8× and 4.3× gain, when measuring background variability as the noise component. Three areas of the low- contrast insert were considered, and showed results generally lying between those of the cold and hot inserts, with one exception demonstrating 9.15× and 10.35× gains for the background and replicate variability measures, respectively.

Measuring gains using the replicate noise demonstrated similar quality gain as the background variability.

Conclusions: The results of this work agree with previous studies stating that TOF information contributes significantly to PET image quality when utilized during reconstruction, specifically for hot lesions and cold regions. This was shown visually, graphically, and quantitatively. The unique quantification method devised, which uses image quality plots to generate gain factors in terms of equivalent non-TOF scan time, was successfully implemented and yielded relatively consistent results. The new phantom insert developed to mimic lower-contrast regions present in human abdominal images was successfully imaged, showing a 1.3× to 4.2× overall gain in equivalent scan time across all phantom sizes.

Trends were observed in several aspects of these results that may subjugate TOF quality gain even further. Cold areas recover better than hot lesions, as expected, but low-contrast areas show varying levels of TOF improvement, and tend to lie between those demonstrated for hot and cold regions.

Finally, similar results were found when considering background variability and replicate variability noise measures, which can be considered further validation of the image quality results.