Energy-based scatter estimation in clinical PET.

BACKGROUND: Scatter correction (SC) is essential in PET for accurate quantitative imaging. The state-of-the-art SC method is single-scatter simulation (SSS). Although this method is usually robust and accurate, it can fail in some situations, for example when there is motion between the CT and PET scans in PET/CT. Therefore, it is of interest to consider other SC methods. PURPOSE: In this work, an energy-based scatter estimation (EBS) method is described in detail, tested in phantoms and patients, and compared to SSS. METHODS: This version of EBS was developed for list-mode data from Biograph Vision-600 PET/CT scanner. EBS is based on digitized 2D energy histograms in each bin of a coarsely sampled PET sinogram, either with or without time of flight (TOF). The histograms are modeled as a noisy realization of a linear combination of nine basis functions whose parameters were derived from a measurement of the 511-keV photopeak spectrum as well as Monte-Carlo simulations of the scattering process. EBS uses an iterative expectation maximization approach to determine the coefficients in the linear combination, and from this estimates the scatter. The investigation was restricted to 18 F-based PET data in which the acquired number of counts was similar to the levels seen in oncological whole-body PET/CT scans. To evaluate the performance, phantom scans were used that involved the NEMA NU2-2018 protocol, a slab phantom, an NU 2-1994 phantom, a cardiac phantom in an anthropomorphic chest phantom, and a uniformly-filled torso phantom with a bladder phantom slightly outside the axial field of view. Contrast recovery (CR) and other parameters were evaluated in images reconstructed with SSS and EBS. Furthermore, FDG PET scans of seven lung cancer patients were used in the evaluation. Standardized uptake values (SUV) based on SSS and EBS were compared in 27 lesions. RESULTS: EBS and SSS images were visually similar in all cases except the torso + bladder phantom, where the EBS was much closer to the expected uniform image. The NU2-2018 analysis indicated a 2% scatter residual in EBS images compared to 3% with SSS, and 10% higher background variability, which is a surrogate for image noise. The cardiac phantom scan showed that CR was 98.2% with EBS and 99.6% with SSS, and that the SSS sinogram had values greater than the net-true emission sinogram, indicating a slight overcorrection in the case of SSS. In the lesion SUV comparison in patient scans, EBS correlated strongly (R2  = 0.9973) with SSS, and SUV based on EBS were systematically 0.1 SUV lower. In the case of the torso + bladder phantom portion, the SSS image of the torso + bladder phantom was 299% times hotter than expected in one area, due to scatter estimation error, compared to 16% colder with EBS. CONCLUSIONS: In evaluating clinically relevant parameters such as SUV in focal lesions, EBS and SSS give almost the same results. In phantoms, some scatter figures of merit were slightly improved by use of EBS, though an image variability figure of merit was slightly degraded. In typical oncological whole-body PET/CT, EBS may be a suitable replacement for SSS, especially when SSS fails due to technical problems during the scan.

Effect of Scan Time on Oncologic Lesion Detection in Whole-Body PET.

Lesion-detection performance in oncologic PET depends in part upon count statistics, with shorter scans having higher noise and reduced lesion detectability. However, advanced techniques such as time-of-flight (TOF) and point spread function (PSF) modeling can improve lesion detection. This work investigates the relationship between reducing count levels (as a surrogate for scan time) and reconstructing with PSF model and TOF. A series of twenty-four whole-body phantom scans was acquired on a Biograph mCT TOF PET/CT scanner using the experimental methodology prescribed for the Utah PET Lesion Detection Database. Six scans were acquired each day over four days, with up to 23 (68)Ge shell-less lesions (diam. 6, 8, 10, 12, 16 mm) distributed throughout the phantom thorax and pelvis. Each scan acquired 6 bed positions at 240 s/bed in listmode format. The listmode files were then statistically pruned, preserving Poisson statistics, to equivalent count levels for scan times of 180 s, 120 s, 90 s, 60 s, 45 s, 30 s, and 15 s per bed field-of-view, corresponding to whole-body scan times of 1.5-24 min. Each dataset was reconstructed using ordinary Poisson line-of-response (LOR) OSEM, with PSF model, with TOF, and with PSF+TOF. Localization receiver operating characteristics (LROC) analysis was then performed using the channelized non-prewhitened (CNPW) observer. The results were analyzed to delineate the relationship between scan time, reconstruction method, and strength of post-reconstruction filter. Lesion-detection performance degraded as scan time was reduced, and progressively stronger filters were required to maximize performance for the shorter scans. PSF modeling and TOF were found to improve detection performance, but the degree of improvement for TOF was much larger than for PSF for the large phantom used in this study. Notably, the images using TOF provided equivalent lesion-detection performance to the images without TOF for scan durations 40% shorter, suggesting that TOF may offset, at least in part, the need for longer scan times in larger patients.

Reproducibility of FDG PET/CT image-based cancer staging and standardized uptake values with simulated reduction of injected FDG dose or acquisition time.

18F-fluorodeoxyglucose (FDG) PET/CT is widely used for oncologic imaging. This study aimed to evaluate, using data simulation, if reduction of injected FDG dose or PET acquisition time could be technically feasible when utilizing a sensitive commercial PET/CT imaging system, without sacrificing image quality, image-based staging accuracy, or standardized uptake value (SUV) accuracy. De-identified, standard of care oncologic FDG PET/CT datasets from 83 adults with lymphoma, lung carcinoma or breast carcinoma were retrospectively analyzed. All images had been acquired using clinical standard dose and acquisition time on a single PET/CT system. The list mode datasets were retrospectively software reprocessed to achieve undersampling of counts, thus simulating the effect of shorter PET acquisition time or lower injected FDG dose. The simulated reduced-count images were reviewed and compared with full-count images to assess and compare qualitative (subjective image quality, stage stability) and semi-quantitative (image noise, SUVmax stability, signal-to-noise and contrast-to-noise ratios within index lesions driving cancer stage) parameters. While simulated reduced-count images had measurably greater noise, there appeared to be no significant loss of image-based staging accuracy nor SUVmax reproducibility down to simulated FDG dose of 0.05 mCi/kg at continuous bed motion rate of 1.1 mm/sec. This retrospective simulation study suggests that a modest reduction of either injected FDG dose or emission scan time might be feasible in this limited oncologic population scanned on a single PET/CT system. Verification of these results with prospectively acquired images using actual low injected FDG activity and/or short imaging time is recommended.

Improved Alignment of PET and CT Images in Whole-Body PET/CT in Cases of Respiratory Motion During CT.

Respiratory motion during the CT and PET parts of a PET/CT scan leads to imperfect alignment of anatomic features seen by the 2 modalities. In this work, we concentrate on the effects of motion during CT. We propose a novel approach for improving the alignment. Methods: Respiratory waveform data were gathered during the CT and PET parts of 28 PET/CT scans of cancer patients with 40 lesions up to 3 cm in size in the lung or upper abdomen. PET list-mode data were reconstructed by 3 reconstruction methods: PET/static (the standard method with no motion correction); PET/ex (a method that calculates a range of expiratory amplitudes from the lowest one to the highest one); and PET/matched (a novel method that uses both waveforms). The 3 methods were compared. The distance between tumor positions in PET and CT were characterized in visual interpretation by physicians as well as quantitatively. Tumor SUVs (SUVmax and SUVpeak) were determined relative to SUV based on the static method. Image noise was evaluated in the liver and compared with PET/static. Results: In visual interpretation, the rate of good alignment was 13 of 21, 13 of 23, and 18 of 21 for the PET/static, PET/ex, and PET/matched methods, respectively, and the mean PET/CT distances were 3.5, 5.1, and 2.8 mm. In visual comparison with PET/ex, the rate of good alignment was increased in 1 of 10 and 7 of 10 cases for PET/static and PET/matched, respectively. SUVmax was on average 21% higher than PET/static when either PET/ex or PET/matched was used. SUVpeak was 12% higher. Image noise in the liver was 15% higher than PET/static for the PET/ex method, and 40% higher for PET/matched; that is, noise was much lower than in gated PET. Conclusion: Acquiring respiratory waveforms both in PET (as in the current state of the art) and in CT (an unusual key step in this approach) has the potential to improve the alignment of PET and CT images. A proposed method for using this information was tested. Improved alignment was demonstrated.

Metal Artifact Reduction of CT Scans to Improve PET/CT.

In recent years, different metal artifact reduction methods have been developed for CT. These methods have only recently been introduced for PET/CT even though they could be beneficial for interpretation, segmentation, and quantification of the PET/CT images. In this study, phantom and patient scans were analyzed visually and quantitatively to measure the effect on PET images of iterative metal artifact reduction (iMAR) of CT data. Methods: The phantom consisted of 2 types of hip prostheses in a solution of 18F-FDG and water. 18F-FDG PET/CT scans of 14 patients with metal implants (either dental implants, hip prostheses, shoulder prostheses, or pedicle screws) and 68Ga-labeled prostate-specific membrane antigen (68Ga-PSMA) PET/CT scans of 7 patients with hip prostheses were scored by 2 experienced nuclear medicine physicians to analyze clinical relevance. For all patients, a lesion was located in the field of view of the metal implant. Phantom and patients were scanned in a PET/CT scanner. The standard low-dose CT scans were processed with the iMAR algorithm. The PET data were reconstructed using attenuation correction provided by both standard CT and iMAR-processed CT. Results: For the phantom scans, cold artifacts were visible on the PET image. There was a 30% deficit in 18F-FDG concentration, which was restored by iMAR processing, indicating that metal artifacts on CT images induce quantification errors in PET data. The iMAR algorithm was useful for most patients. When iMAR was used, the confidence in interpretation increased or stayed the same, with an average improvement of 28% ± 20% (scored on a scale of 0%-100% confidence). The SUV increase or decrease depended on the type of metal artifact. The mean difference in absolute values of SUVmean of the lesions was 3.5% ± 3.3%. Conclusion: The iMAR algorithm increases the confidence of the interpretation of the PET/CT scan and influences the SUV. The added value of iMAR depends on the indication for the PET/CT scan, location and size/type of the prosthesis, and location and extent of the disease.

A knowledge-based method for reducing attenuation artefacts caused by cardiac appliances in myocardial PET/CT.

Attenuation artefacts due to implanted cardiac defibrillator leads have previously been shown to adversely impact cardiac PET/CT imaging. In this study, the severity of the problem is characterized, and an image-based method is described which reduces the resulting artefact in PET. Automatic implantable cardioverter defibrillator (AICD) leads cause a moving-metal artefact in the CT sections from which the PET attenuation correction factors (ACFs) are derived. Fluoroscopic cine images were measured to demonstrate that the defibrillator's highly attenuating distal shocking coil moves rhythmically across distances on the order of 1 cm. Rhythmic motion of this magnitude was created in a phantom with a moving defibrillator lead. A CT study of the phantom showed that the artefact contained regions of incorrect, very high CT values and adjacent regions of incorrect, very low CT values. The study also showed that motion made the artefact more severe. A knowledge-based metal artefact reduction method (MAR) is described that reduces the magnitude of the error in the CT images, without use of the corrupted sinograms. The method modifies the corrupted image through a sequence of artefact detection procedures, morphological operations, adjustments of CT values and three-dimensional filtering. The method treats bone the same as metal. The artefact reduction method is shown to run in a few seconds, and is validated by applying it to a series of phantom studies in which reconstructed PET tracer distribution values are wrong by as much as 60% in regions near the CT artefact when MAR is not applied, but the errors are reduced to about 10% of expected values when MAR is applied. MAR changes PET image values by a few per cent in regions not close to the artefact. The changes can be larger in the vicinity of bone. In patient studies, the PET reconstruction without MAR sometimes results in anomalously high values in the infero-septal wall. Clinical performance of MAR is assessed by two physicians' inspection of images generated in 30 patients with and without MAR. Noticeable image differences are judged in 14 of 28 (50%) observations with AICD leads, and significant clinical impact is judged in 2 of 28 (7%) of those observations. A polar map analysis shows significant differences in 10 of 14 (71%) studies with AICD leads, and 0 of 16 (0%) studies without AICD leads. These results show that the MAR method is successful in reducing the magnitude of the metal artefact without incorrectly altering cases without metal artefact. In spite of profound changes to the CT image from the moving metal, the PET ACF in that study was changed by no more than 20%.

Improving the Spatial Alignment in PET/CT Using Amplitude-Based Respiration-Gated PET and Respiration-Triggered CT.

UNLABELLED: Respiratory motion during PET can cause inaccuracies in the quantification of radiotracer uptake, which negatively affects PET-guided radiotherapy planning. Quantitative accuracy can be improved by respiratory gating. However, additional miscalculation of standardized uptake value (SUV) in PET images can be caused by inappropriate attenuation correction due to a spatial mismatch between gated PET and CT. In this study, the effect of respiration-triggered CT on the spatial match between CT and amplitude-based respiration-gated PET images is investigated. METHODS: (18)F-FDG PET/CT was performed in 38 patients. Images were acquired on 2 PET/CT scanners, one without and one with continuous bed motion during PET acquisition. The amplitude limits of the amplitude-based respiration-gated PET were used for the respiration-triggered sequential low-dose CT. Both standard (spiral) and triggered CT scans were used to reconstruct the PET data. Spatial mismatch was quantified using the position difference between the lung-liver boundary in PET and CT images, the distance between PET and CT lung lesions' centroids, and the amount of overlap of lesions indicated by the Jaccard similarity coefficient. Furthermore, the effect of attenuation correction was quantified by measuring SUVs in lung lesions. RESULTS: For triggered CT, the average distance between the lung-liver boundary in PET and CT was significantly reduced (4.5 ± 6.7 mm) when compared with standard CT (9.2 ± 8.1 mm) (P < 0.001). The mean distance between the lesions' centroids in PET and CT images was 6.3 ± 4.0 and 5.6 ± 4.2 mm (P = 0.424), for the standard and triggered CT, respectively. Similarly, the Jaccard similarity coefficient was 0.30 ± 0.21 and 0.32 ± 0.20 (P = 0.609) for standard and triggered CT, respectively. For 6 lesions, there was no overlap of PET and CT when the standard CT was used; compared with the triggered CT, these lesions showed (partial) overlap. The maximum and mean SUV increase of the PET/CT compared with the PET/triggered CT was 5.7% ± 11.2% (P < 0.001) and 6.1% ± 10.2% (P = 0.001), respectively. CONCLUSION: Amplitude-based respiration-gated PET in combination with respiration-triggered CT resulted in a significantly improved match in the area of the liver dome and a significantly higher SUV for lung lesions. However, lesions in the lungs did not show a consistent improvement in spatial match.

Application of discrete data consistency conditions for selecting regularization parameters in PET attenuation map reconstruction.

Simultaneous emission and transmission measurement is appealing in PET due to the matching of geometrical conditions between emission and transmission and reduced acquisition time for the study. A potential problem remains: when transmission statistics are low, attenuation correction could be very noisy. Although noise in the attenuation map can be controlled through regularization during statistical reconstruction, the selection of regularization parameters is usually empirical. In this paper, we investigate the use of discrete data consistency conditions (DDCC) to optimally select one or two regularization parameters. The advantages of the method are that the reconstructed attenuation map is consistent with the emission data and that it accounts for particularity in the emission reconstruction algorithm and acquisition geometry. The methodology is validated using a computer-generated whole-body phantom for both emission and transmission, neglecting random events and scattered radiation. MAP-TR was used for attenuation map reconstruction, while 3D OS-EM is used for estimating the emission image. The estimation of regularization parameters depends on the resolution of the emission image controlled by the number of iterations in OS-EM. The computer simulation shows that, on one hand, DDCC regularized attenuation map reduces propagation of the transmission scan noise to the emission image, while on the other hand DDCC prevents excessive attenuation map smoothing that could result in resolution mismatch artefacts between emission and transmission.

Evaluation of CT-based lean-body SUV.

PURPOSE: The authors introduce a novel method for defining standardized uptake values (SUVs) in PET∕CT based on routinely collected CT data. The goal of the study is to reduce, if possible, the variability of SUV in a heterogeneous population. Two well established methods for defining SUV are based on body weight (BW) and lean body mass, calculated as a function of height, weight, and sex with an empirical formula (LBM). The authors investigate two novel models, CT1 and CT2, that estimate the lean mass from CT Hounsfield Unit values. The authors compare the four methods, assessing the variability of hepatic SUV in (18)F-FDG studies. METHODS: CT images from 252 cancer patients were segmented into regions representing lean tissues, fat, and bone. The fraction of lean tissue in the scanned region was extrapolated to the entire body with a naive method (CT1) and a method that modeled typical FDG uptake patterns (CT2). For each method, SUV-based measurements of the liver were calculated for all patients and dependence on body weight was assessed. Coefficients of variation (CVs) were evaluated. Several sub-cohorts were analyzed, including those with low and high body mass index (BMI). The extrapolation technique was tested in 19 melanoma patients who received head to toe PET∕CT scans. CT-based weight predictions were compared with actual patient weight in melanoma studies and in PET∕CT scans of pigs. RESULTS: Only the SUV based on BW method depended significantly on body weight. CVs for the BW, LBM, CT1, and CT2 methods were, respectively, 18.0%, 15.5%, 15.9%, and 14.9%. In the high-BMI cohort, CVs were 18.2%, 16.2%, 16.2%, and 15.1%. Mean SUV of the 14 most obese patients agreed most closely with mean SUV of 120 lean patients when the CT2 method was used. SUV based on truncated CT agreed with head to toe predictions within 5% for the CT1 method and 1% for the CT2 method. CT-based weight estimate recovered 97.4% of the weight in head to toe studies of humans and 99.7% in pig studies. CONCLUSIONS: The novel CT1 and CT2 methods were less variable than the BW method and were comparable to the LBM method. SUV were little affected by missing CT data.

Experimental comparison of lesion detectability for four fully-3D PET reconstruction schemes.

The objective of this work was to evaluate the lesion detection performance of four fully-3D positron emission tomography (PET) reconstruction schemes using experimentally acquired data. A multi-compartment anthropomorphic phantom was set up to mimic whole-body (18)F-fluorodeoxyglucose (FDG) cancer imaging and scanned 12 times in 3D mode, obtaining count levels typical of noisy clinical scans. Eight of the scans had 26 (68)Ge "shell-less" lesions (6, 8-, 10-, 12-, 16-mm diameter) placed throughout the phantom with various target:background ratios. This provided lesion-present and lesion-absent datasets with known truth appropriate for evaluating lesion detectability by localization receiver operating characteristic (LROC) methods. Four reconstruction schemes were studied: 1) Fourier rebinning (FORE) followed by 2D attenuation-weighted ordered-subsets expectation-maximization, 2) fully-3D AW-OSEM, 3) fully-3D ordinary-Poisson line-of-response (LOR-)OSEM; and 4) fully-3D LOR-OSEM with an accurate point-spread function (PSF) model. Two forms of LROC analysis were performed. First, a channelized nonprewhitened (CNPW) observer was used to optimize processing parameters (number of iterations, post-reconstruction filter) for the human observer study. Human observers then rated each image and selected the most-likely lesion location. The area under the LROC curve ( A(LROC)) and the probability of correct localization were used as figures-of-merit. The results of the human observer study found no statistically significant difference between FORE and AW-OSEM3D ( A(LROC)=0.41 and 0.36, respectively), an increase in lesion detection performance for LOR-OSEM3D ( A(LROC)=0.45, p=0.076), and additional improvement with the use of the PSF model ( A(LROC)=0.55, p=0.024). The numerical CNPW observer provided the same rankings among algorithms, but obtained different values of A(LROC). These results show improved lesion detection performance for the reconstruction algorithms with more sophisticated statistical and imaging models as compared to the previous-generation algorithms.