Performance of MAP reconstruction for hot lesion detection in whole-body PET/CT: an evaluation with human and numerical observers

Impact of γ factor in the penalty function of Bayesian penalized likelihood reconstruction (Q.Clear) to achieve high-resolution PET images

BACKGROUND

The Bayesian penalized likelihood PET reconstruction (BPL) algorithm, Q.Clear (GE Healthcare), has recently been clinically applied to clinical image reconstruction. The BPL includes a relative difference penalty (RDP) as a penalty function. The β value that controls the behavior of RDP determines the global strength of noise suppression, whereas the γ factor in RDP controls the degree of edge preservation. The present study aimed to assess the effects of various γ factors in RDP on the ability to detect sub-centimeter lesions.

METHODS

All PET data were acquired for 10 min using a Discovery MI PET/CT system (GE Healthcare). We used a NEMA IEC body phantom containing spheres with inner diameters of 10, 13, 17, 22, 28 and 37 mm and 4.0, 5.0, 6.2, 7.9, 10 and 13 mm. The target-to-background ratio of the phantom was 4:1, and the background activity concentration was 5.3 kBq/mL. We also evaluated cold spheres containing only non-radioactive water with the same background activity concentration. All images were reconstructed using BPL + time of flight (TOF). The ranges of β values and γ factors in BPL were 50-600 and 2-20, respectively. We reconstructed PET images using the Duetto toolbox for MATLAB software. We calculated the % hot contrast recovery coefficient (CRC_hot) of each hot sphere, the cold CRC (CRC_cold) of each cold sphere, the background variability (BV) and residual lung error (LE). We measured the full width at half maximum (FWHM) of the micro hollow hot spheres ≤ 13 mm to assess spatial resolution on the reconstructed PET images.

RESULTS

The CRC_hot and CRC_cold for different β values and γ factors depended on the size of the small spheres. The CRC_hot, CRC_cold and BV increased along with the γ factor. A 6.2-mm hot sphere was obvious in BPL as lower β values and higher γ factors, whereas γ factors ≥ 10 resulted in images with increased background noise. The FWHM became smaller when the γ factor increased.

CONCLUSION

High and low γ factors, respectively, preserved the edges of reconstructed PET images and promoted image smoothing. The BPL with a γ factor above the default value in Q.Clear (γ factor = 2) generated high-resolution PET images, although image noise slightly diverged. Optimizing the β value and the γ factor in BPL enabled the detection of lesions ≤ 6.2 mm.

Development and validation of the Lesion Synthesis Toolbox and the Perception Study Tool for quantifying observer limits of detection of lesions in positron emission tomography

Purpose: Accurate detection of cancer lesions in positron emission tomography (PET) is fundamental to achieving favorable clinical outcomes. Therefore, image reconstruction, processing, visualization, and interpretation techniques must be optimized for this task. The objective of this work was to (1) develop and validate an efficient method to generate well-characterized synthetic lesions in real patient data and (2) to apply these lesions in a human perception experiment to establish baseline measurements of the limits of lesion detection as a function of lesion size and contrast using current imaging technologies. Approach: A fully integrated software package for synthesizing well-characterized lesions in real patient PET was developed using a vendor provided PET image reconstruction toolbox (REGRECON5, General Electric Healthcare, Waukesha, Wisconsin). Lesion characteristics were validated experimentally for geometric accuracy, activity accuracy, and absence of artifacts. The Lesion Synthesis Toolbox was used to generate a library of 133 synthetic lesions of varying sizes ( n = 7 ) and contrast levels ( n = 19 ) in manually defined locations in the livers of 37 patient studies. A lesion-localization perception study was performed with seven observers to determine the limits of detection with regard to lesion size and contrast using our web-based perception study tool. Results: The Lesion Synthesis Toolbox was validated for accurate lesion placement and size. Lesion intensities were deemed accurate with slightly elevated activities (5% at 2:1 lesion-to-background contrast) in small lesions ( Ø = 15    mm spheres), and no bias in large lesions ( Ø = 22.5    mm ). Bed-stitching artifacts were not observed, and lesion attenuation correction bias was small ( - 1.6 ± 1.2 % ). The 133 liver lesions were synthesized in ∼ 50    h , and readers were able to complete the perception study of these lesions in 12 ± 3    min with consistent limits of detection amongst all readers. Conclusions: Our open-source utilities can be employed by nonexperts to generate well-characterized synthetic lesions in real patient PET images and for administering perception studies on clinical workstations without the need to install proprietary software.

Quantitative performance and optimal regularization parameter in block sequential regularized expectation maximization reconstructions in clinical 68Ga-PSMA PET/MR

BACKGROUND

In contrast to ordered subset expectation maximization (OSEM), block sequential regularized expectation maximization (BSREM) positron emission tomography (PET) reconstruction algorithms can run until full convergence while controlling image quality and noise. Recent studies with BSREM and 18F-FDG PET reported higher signal-to-noise ratios and higher standardized uptake values (SUV). In this study, we investigate the optimal regularization parameter (β) for clinical 68Ga-PSMA PET/MR reconstructions in the pelvic region applying time-of-flight (TOF) BSREM in comparison to TOF OSEM. Two-minute emission data from the pelvic region of 25 patients who underwent 68Ga-PSMA PET/MR were retrospectively reconstructed. Reference OSEM reconstructions had 28 subsets and 2 iterations. BSREM reconstructions were performed with 15 β values between 150 and 1200. Regions of interest (ROIs) were drawn around lesions and in uniform background. Background SUVmean (average) and SUVstd (standard deviation), and lesion SUVmax (average of 5 hottest voxels) were calculated. Differences were analyzed using the Wilcoxon matched pairs signed-rank test.

RESULTS

A total of 40 lesions were identified in the pelvic region. Background noise (SUVstd) and lesions SUVmax decreased with increasing β. Image reconstructions with β values lower than 400 have higher (p < 0.01) background noise, compared to the reference OSEM reconstructions, and are therefore less useful. Lesions with low activity on images reconstructed with β values higher than 600 have a lower (p < 0.05) SUVmax compared to the reference. These reconstructions are likely visually appealing due to the lower background noise, but the lower SUVmax could possibly render small low-uptake lesions invisible.

CONCLUSIONS

In our study, we showed that PET images reconstructed with TOF BSREM in combination with the 68Ga-PSMA tracer result in lower background noise and higher SUVmax values in lesions compared to TOF OSEM. Our study indicates that a β value between 400 and 550 might be the optimal compromise between high SUVmax and low background noise.

Phantom, clinical, and texture indices evaluation and optimization of a penalized-likelihood image reconstruction method (Q.Clear) on a BGO PET/CT scanner

INTRODUCTION

The aim of this study was to evaluate the behavior of a penalized-likelihood image reconstruction method (Q.Clear) under different count statistics and lesion-to-background ratios (LBR) on a BGO scanner, in order to obtain an optimum penalization factor (β value) to study and optimize for different acquisition protocols and clinical goals.

METHODS

Both phantom and patient images were evaluated. Data from an image quality phantom were acquired using different Lesion-to-Background ratios and acquisition times. Then, each series of the phantom was reconstructed using β values between 50 and 500, at intervals of 50. Hot and cold contrasts were obtained, as well as background variability and contrast-to-noise ratio (CNR). Fifteen ¹⁸ F-FDG patients (five brain scans and 10 torso acquisitions) were acquired and reconstructed using the same β values as in the phantom reconstructions. From each lesion in the torso acquisition, noise, contrast, and signal-to-noise ratio (SNR) were computed. Image quality was assessed by two different nuclear medicine physicians. Additionally, the behaviors of 12 different textural indices were studied over 20 different lesions.

RESULTS

Q.Clear quantification and optimization in patient studies depends on the activity concentration as well as on the lesion size. In the studied range, an increase on β is translated in a decrease in lesion contrast and noise. The net product is an overall increase in the SNR, presenting a tendency to a steady value similar to the CNR in phantom data. As the activity concentration or the sphere size increase the optimal β increases, similar results are obtained from clinical data. From the subjective quality assessment, the optimal β value for torso scans is in a range between 300 and 400, and from 100 to 200 for brain scans. For the recommended torso β values, texture indices present coefficients of variation below 10%.

CONCLUSIONS

Our phantom and patients demonstrate that improvement of CNR and SNR of Q.Clear algorithm which depends on the studied conditions and the penalization factor. Using the Q.Clear reconstruction algorithm in a BGO scanner, a β value of 350 and 200 appears to be the optimal value for 18F-FDG oncology and brain PET/CT, respectively.

Evaluation of lesion detectability in positron emission tomography when using a convergent penalized likelihood image reconstruction method

We have previously developed a convergent penalized likelihood (PL) image reconstruction algorithm using the relative difference prior (RDP) and showed that it achieves more accurate lesion quantitation compared to ordered subsets expectation maximization (OSEM). We evaluated the detectability of low-contrast liver and lung lesions using the PL-RDP algorithm compared to OSEM. We performed a two-alternative forced choice study using a channelized Hotelling observer model that was previously validated against human observers. Lesion detectability showed a stronger dependence on lesion size for PL-RDP than OSEM. Lesion detectability was improved using time-of-flight (TOF) reconstruction, with greater benefit for the liver compared to the lung and with increasing benefit for decreasing lesion size and contrast. PL detectability was statistically significantly higher than OSEM for 20 mm liver lesions when contrast was [Formula: see text] ([Formula: see text]), and TOF PL detectability was statistically significantly higher than TOF OSEM for 15 and 20 mm liver lesions with contrast [Formula: see text] and [Formula: see text], respectively. For all other cases, there was no statistically significant difference between PL and OSEM ([Formula: see text]). For the range of studied lesion properties, lesion detectability using PL-RDP was equivalent or improved compared to using OSEM.

Postreconstruction filtering of 3D PET images by using weighted higher-order singular value decomposition

Background

Positron emission tomography (PET) always suffers from high levels of noise due to the constraints of the injected dose and acquisition time, especially in the studies of dynamic PET imaging. To improve the quality of PET image, several approaches have been introduced to suppress noise. However, traditional filters often blur the image edges, or erase small detail, or rely on multiple parameters. In order to solve such problems, nonlocal denoising methods have been adapted to denoise PET images.

Methods

In this paper, we propose to use the weighted higher-order singular value decomposition for PET image denoising. We first modeled the noise in the PET image as Poisson distribution. Then, we transformed the noise to an additive Gaussian noise by use of the anscombe root transformation. Finally, we denoised the transformed image using the proposed higher-order singular value decomposition (HOSVD)-based algorithms. The denoised results were compared with results from some general filters by performing physical phantom and mice studies.

Results

Compared to other commonly used filters, HOSVD-based denoising algorithms can preserve boundaries and quantitative accuracy better. The spatial resolution and the low activity features in PET image also can be preserved by use of HOSVD-based methods. Comparing with the standard HOSVD-based algorithm, the proposed weighted HOSVD algorithm can suppress the stair-step artifact, and the time-consumption is about half of that needed by the Wiener-augmented HOSVD algorithm.

Conclusions

The proposed weighted HOSVD denoising algorithm can suppress noise while better preserving of boundary and quantity in PET images.

Electronic supplementary material

The online version of this article (doi:10.1186/s12938-016-0221-y) contains supplementary material, which is available to authorized users.

Theoretical Analysis of Penalized Maximum-Likelihood Patlak Parametric Image Reconstruction in Dynamic PET for Lesion Detection

Detecting cancerous lesions is a major clinical application of emission tomography. In a previous work, we studied penalized maximum-likelihood (PML) image reconstruction for lesion detection in static PET. Here we extend our theoretical analysis of static PET reconstruction to dynamic PET. We study both the conventional indirect reconstruction and direct reconstruction for Patlak parametric image estimation. In indirect reconstruction, Patlak parametric images are generated by first reconstructing a sequence of dynamic PET images, and then performing Patlak analysis on the time activity curves (TACs) pixel-by-pixel. In direct reconstruction, Patlak parametric images are estimated directly from raw sinogram data by incorporating the Patlak model into the image reconstruction procedure. PML reconstruction is used in both the indirect and direct reconstruction methods. We use a channelized Hotelling observer (CHO) to assess lesion detectability in Patlak parametric images. Simplified expressions for evaluating the lesion detectability have been derived and applied to the selection of the regularization parameter value to maximize detection performance. The proposed method is validated using computer-based Monte Carlo simulations. Good agreements between the theoretical predictions and the Monte Carlo results are observed. Both theoretical predictions and Monte Carlo simulation results show the benefit of the indirect and direct methods under optimized regularization parameters in dynamic PET reconstruction for lesion detection, when compared with the conventional static PET reconstruction.

Impact of respiratory motion correction and spatial resolution on lesion detection in PET: a simulation study based on real MR dynamic data

The aim of this study is to investigate the impact of respiratory motion correction and spatial resolution on lesion detectability in PET as a function of lesion size and tracer uptake. Real respiratory signals describing different breathing types are combined with a motion model formed from real dynamic MR data to simulate multiple dynamic PET datasets acquired from a continuously moving subject. Lung and liver lesions were simulated with diameters ranging from 6 to 12 mm and lesion to background ratio ranging from 3:1 to 6:1. Projection data for 6 and 3 mm PET scanner resolution were generated using analytic simulations and reconstructed without and with motion correction. Motion correction was achieved using motion compensated image reconstruction. The detectability performance was quantified by a receiver operating characteristic (ROC) analysis obtained using a channelized Hotelling observer and the area under the ROC curve (AUC) was calculated as the figure of merit. The results indicate that respiratory motion limits the detectability of lung and liver lesions, depending on the variation of the breathing cycle length and amplitude. Patients with large quiescent periods had a greater AUC than patients with regular breathing cycles and patients with long-term variability in respiratory cycle or higher motion amplitude. In addition, small (less than 10 mm diameter) or low contrast (3:1) lesions showed the greatest improvement in AUC as a result of applying motion correction. In particular, after applying motion correction the AUC is improved by up to 42% with current PET resolution (i.e. 6 mm) and up to 51% for higher PET resolution (i.e. 3 mm). Finally, the benefit of increasing the scanner resolution is small unless motion correction is applied. This investigation indicates high impact of respiratory motion correction on lesion detectability in PET and highlights the importance of motion correction in order to benefit from the increased resolution of future PET scanners.

Evaluation of penalty design in penalized maximum-likelihood image reconstruction for lesion detection

Detecting cancerous lesions is a major clinical application in emission tomography. Previously, we developed a method to design a shift-variant quadratic penalty function in penalized maximum-likelihood (PML) image reconstruction to improve the lesion detectability. We used a multiview channelized Hotelling observer (mvCHO) to assess the lesion detectability in three-dimensional images and validated the penalty design using computer simulations. In this study, we evaluate the benefit of the proposed penalty function for lesion detection using real patient data and artificial lesions. A high-count real patient dataset with no identifiable tumor inside the field of view is used as the background data. A Na-22 point source is scanned in air at variable locations and the point source data are superimposed onto the patient data as artificial lesions after being attenuated by the patient body. Independent Poisson noise is introduced to the high-count sinograms to generate 200 pairs of lesion-present and lesion-absent datasets, each mimicking a 5-min scan. Lesion detectability is assessed using a mvCHO and a human observer two-alternative forced choice (2AFC) experiment. The results show improvements in lesion detection by the proposed method compared with the conventional first-order quadratic penalty function and a total variation (TV) edge-preserving penalty function.

Postreconstruction nonlocal means filtering of whole-body PET with an anatomical prior

Positron emission tomography (PET) images usually suffer from poor signal-to-noise ratio (SNR) due to the high level of noise and low spatial resolution, which adversely affect its performance for lesion detection and quantification. The complementary information present in high-resolution anatomical images from multi-modality imaging systems could potentially be used to improve the ability to detect and/or quantify lesions. However, previous methods that use anatomical priors usually require matched organ/lesion boundaries. In this study, we investigated the use of anatomical information to suppress noise in PET images while preserving both quantitative accuracy and the amplitude of prominent signals that do not have corresponding boundaries on computerized tomography (CT). The proposed approach was realized through a postreconstruction filter based on the nonlocal means (NLM) filter, which reduces noise by computing the weighted average of voxels based on the similarity measurement between patches of voxels within the image. Anatomical knowledge obtained from CT was incorporated to constrain the similarity measurement within a subset of voxels. In contrast to other methods that use anatomical priors, the actual number of neighboring voxels and weights used for smoothing were determined from a robust measurement on PET images within the subset. Thus, the proposed approach can be robust to signal mismatches between PET and CT. A 3-D search scheme was also investigated for the volumetric PET/CT data. The proposed anatomically guided median nonlocal means filter (AMNLM) was first evaluated using a computer phantom and a physical phantom to simulate realistic but challenging situations where small lesions are located in homogeneous regions, which can be detected on PET but not on CT. The proposed method was further assessed with a clinical study of a patient with lung lesions. The performance of the proposed method was compared to Gaussian, edge-preserving bilateral and NLM filters, as well as median nonlocal means (MNLM) filtering without an anatomical prior. The proposed AMNLM method yielded improved lesion contrast and SNR compared with other methods even with imperfect anatomical knowledge, such as missing lesion boundaries and mismatched organ boundaries.

Regularization design in penalized maximum-likelihood image reconstruction for lesion detection in 3D PET

Detecting cancerous lesions is a major clinical application in emission tomography. In previous work, we have studied penalized maximum-likelihood (PML) image reconstruction for the detection task and proposed a method to design a shift-invariant quadratic penalty function to maximize detectability of a lesion at a known location in a two dimensional image. Here we extend the regularization design to maximize detectability of lesions at unknown locations in fully 3D PET. We used a multiview channelized Hotelling observer (mvCHO) to assess the lesion detectability in 3D images to mimic the condition where a human observer examines three orthogonal views of a 3D image for lesion detection. We derived simplified theoretical expressions that allow fast prediction of the detectability of a 3D lesion. The theoretical results were used to design the regularization in PML reconstruction to improve lesion detectability. We conducted computer-based Monte Carlo simulations to compare the optimized penalty with the conventional penalty for detecting lesions of various sizes. Only true coincidence events were simulated. Lesion detectability was also assessed by two human observers, whose performances agree well with that of the mvCHO. Both the numerical observer and human observer results showed a statistically significant improvement in lesion detection by using the proposed penalty function compared to using the conventional penalty function.

Model observers in medical imaging research

Model observers play an important role in the optimization and assessment of imaging devices. In this review paper, we first discuss the basic concepts of model observers, which include the mathematical foundations and psychophysical considerations in designing both optimal observers for optimizing imaging systems and anthropomorphic observers for modeling human observers. Second, we survey a few state-of-the-art computational techniques for estimating model observers and the principles of implementing these techniques. Finally, we review a few applications of model observers in medical imaging research.

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.

Evaluation of three MRI-based anatomical priors for quantitative PET brain imaging

In emission tomography, image reconstruction and therefore also tracer development and diagnosis may benefit from the use of anatomical side information obtained with other imaging modalities in the same subject, as it helps to correct for the partial volume effect. One way to implement this, is to use the anatomical image for defining the a priori distribution in a maximum-a-posteriori (MAP) reconstruction algorithm. In this contribution, we use the PET-SORTEO Monte Carlo simulator to evaluate the quantitative accuracy reached by three different anatomical priors when reconstructing positron emission tomography (PET) brain images, using volumetric magnetic resonance imaging (MRI) to provide the anatomical information. The priors are: 1) a prior especially developed for FDG PET brain imaging, which relies on a segmentation of the MR-image (Baete , 2004); 2) the joint entropy-prior (Nuyts, 2007); 3) a prior that encourages smoothness within a position dependent neighborhood, computed from the MR-image. The latter prior was recently proposed by our group in (Vunckx and Nuyts, 2010), and was based on the prior presented by Bowsher (2004). The two latter priors do not rely on an explicit segmentation, which makes them more generally applicable than a segmentation-based prior. All three priors produced a compromise between noise and bias that was clearly better than that obtained with postsmoothed maximum likelihood expectation maximization (MLEM) or MAP with a relative difference prior. The performance of the joint entropy prior was slightly worse than that of the other two priors. The performance of the segmentation-based prior is quite sensitive to the accuracy of the segmentation. In contrast to the joint entropy-prior, the Bowsher-prior is easily tuned and does not suffer from convergence problems.