Positron Emission Tomography Segmentation Research Articles

Prognostic model using 18F-FDG PET radiomics predicts progression-free survival in relapsed/refractory Hodgkin lymphoma

AbstractInvestigating prognostic factors in patients with relapsed or primary refractory classical Hodgkin lymphoma (R/R cHL) is essential to optimize risk-adapted treatment strategies. We built a prognostic model using baseline quantitative 18F-fluorodeoxyglucose positron emission tomography (PET) radiomics features and clinical characteristics to predict the progression-free survival (PFS) among patients with R/R cHL treated with salvage chemotherapy followed by autologous stem cell transplantation. Metabolic tumor volume and several novel radiomics dissemination features, representing interlesional differences in distance, volume, and standard uptake value, were extracted from the baseline PET. Machine learning using backward selection and logistic regression were applied to develop and train the model on a total of 113 patients from 2 clinical trials. The model was validated on an independent external cohort of 69 patients. In addition, we validated 4 different PET segmentation methods to calculate radiomics features. We identified a subset of patients at high risk for progression with significant inferior 3-year PFS outcomes of 38.1% vs 88.4% for patients in the low-risk group in the training cohort (P < .001) and 38.5% vs 75.0% in the validation cohort (P = .015), respectively. The overall survival was also significantly better in the low-risk group (P = .022 and P < .001). We provide a formula to calculate a risk score for individual patients based on the model. In conclusion, we developed a prognostic model for PFS combining radiomics and clinical features in a large cohort of patients with R/R cHL. This model calculates a PET-based risk profile and can be applied to develop risk-stratified treatment strategies for patients with R/R cHL. These trials were registered at www.clinicaltrials.gov as #NCT02280993, #NCT00255723, and #NCT01508312.

Automatic lesion detection and segmentation in 18F-flutemetamol positron emission tomography images using deep learning

BackgroundBeta amyloid in the brain, which was originally confirmed by post-mortem examinations, can now be confirmed in living patients using amyloid positron emission tomography (PET) tracers, and the accuracy of diagnosis can be improved by beta amyloid plaque confirmation in patients. Amyloid deposition in the brain is often associated with the expression of dementia. Hence, it is important to identify the anatomically and functionally meaningful areas of the human brain cortex surface using PET to diagnose the possibility of developing dementia. In this study, we demonstrated the validity of automated 18F-flutemetamol PET lesion detection and segmentation based on a complete 2D U-Net convolutional neural network via masking treatment strategies.MethodsPET data were first normalized by volume and divided into five amyloid accumulation zones through axial, coronary, and thalamic slices. A single U-Net was trained using a divided dataset for one of these zones. Ground truth segmentations were obtained by manual delineation and thresholding (1.5 × background).ResultsThe following intersection over union values were obtained for the various slices in the verification dataset: frontal lobe axial/sagittal: 0.733/0.804; posterior cingulate cortex and precuneus coronal/sagittal: 0.661/0.726; lateral temporal lobe axial/coronal: 0.864/0.892; parietal lobe axial/coronal: 0.542/0.759; and striatum axial/sagittal: 0.679/0.752. The U-Net convolutional neural network architecture allowed fully automated 2D division of the 18F-flutemetamol PET brain images of Alzheimer's patients.ConclusionsAs dementia should be tested and evaluated in various ways, there is a need for artificial intelligence programs. This study can serve as a reference for future studies using auxiliary roles and research in Alzheimer's diagnosis.

The Value of FET PET/CT in Recurrent Glioma with a Different IDH Mutation Status: The Relationship between Imaging and Molecular Biomarkers

The evaluation of treatment response remains a challenge in glioma cases because the neuro oncological therapy can lead to the development of treatment-related changes (TRC) that mimic true progression (TP). Positron emission tomography (PET) using O-(2-[18F] fluoroethyl-)-L-tyrosine (18F-FET) has been shown to be a useful tool for detecting TRC and TP. We assessed the diagnostic performance of different 18F-FET PET segmentation approaches and different imaging biomarkers for differentiation between late TRC and TP in glioma patients. Isocitrate dehydrogenase (IDH) status was evaluated as a predictor of disease outcome. In our study, the proportion of TRC in IDH wild type (IDHwt) and IDH mutant (IDHm) subgroups was without significant difference. We found that the diagnostic value of static and dynamic biomarkers of 18F-FET PET for discrimination between TRC and TP depends on the IDH mutation status of the tumor. Dynamic 18F-FET PET acquisition proved helpful in the IDH wild type (IDHwt) subgroup, as opposed to the IDH mutant (IDHm) subgroup, providing an early indication to discontinue dynamic imaging in the IDHm subgroup.

A Multi-Atlas-Based [18F]9-Fluoropropyl-(+)-Dihydrotetrabenazine Positron Emission Tomography Image Segmentation Method for Parkinson's Disease Quantification.

Objectives[18F]9-fluoropropyl-(+)-dihydrotetrabenazine ([18F]-FP-DTBZ) positron emission tomography (PET) provides reliable information for the diagnosis of Parkinson’s disease (PD). In this study, we proposed a multi-atlas-based [18F]-FP-DTBZ PET image segmentation method for PD quantification assessment.MethodsA total of 99 subjects from Xuanwu Hospital of Capital Medical University were included in this study, and both brain PET and magnetic resonance (MR) scans were conducted. Data from 20 subjects were used to generate atlases, based on which a multi-atlas-based [18F]-FP-DTBZ PET segmentation method was developed especially for striatum and its subregions. The proposed method was compared with the template-based method through striatal subregion parcellation performance and the standard uptake value ratio (SUVR) quantification accuracy. Discriminant analysis between healthy controls (HCs) and PD patients was further performed.ResultsSegmentation results of the multi-atlas-based method showed better consistency than the template-based method with the ground truth, yielding a dice coefficient of 0.81 over 0.73 on the full striatum. The SUVRs calculated by the multi-atlas-based method had an average interclass correlation coefficient (ICC) of 0.953 with the standardized result, whereas the template-based method only reached 0.815. The SUVRs of HCs were generally higher than that of patients with PD and showed significant differences in all of the striatal subregions (all p < 0.001). The median and posterior putamen performed best in discriminating patients with PD from HCs.ConclusionThe proposed multi-atlas-based [18F]-FP-DTBZ PET image segmentation method achieved better performance than the template-based method, indicating great potential in improving accuracy and efficiency for PD diagnosis in clinical routine.

Computer-Aided Classification Framework of Parkinsonian Disorders Using 11C-CFT PET Imaging.

PurposeTo investigate the usefulness of a novel computer-aided classification framework for the differential diagnosis of parkinsonian disorders (PDs) based on 11C-methyl-N-2β-carbomethoxy-3β-(4-fluorophenyl)-tropanel (11C-CFT) positron emission tomography (PET) imaging.MethodsPatients with different forms of PDs—including Parkinson's disease (PD), multiple system atrophy (MSA) and progressive supranuclear palsy (PSP)—underwent dopamine transporter (DAT) imaging with 11C-CFT PET. A novel multistep computer-aided classification framework—consisting of magnetic resonance imaging (MRI)-assisted PET segmentation, feature extraction and prediction, and automatic subject classification—was developed. A random forest method was used to assess the diagnostic relevance of different regions to the classification process. Finally, the performance of the computer-aided classification system was tested using various training strategies involving patients with early and advanced disease stages.ResultsAccuracy values for identifying PD, MSA, and PSP were 85.0, 82.2, and 89.7%, respectively—with an overall accuracy of 80.4%. The caudate and putamen provided the highest diagnostic relevance to the proposed classification framework, whereas the contribution of midbrain was negligible. With the exception of sensitivity for diagnosing PSP, the strategy comprising both early and advanced disease stages performed better in terms of sensitivity, specificity, positive predictive value, and negative predictive value within each PDs subtype.ConclusionsThe proposed computer-aided classification framework based on 11C-CFT PET imaging holds promise for improving the differential diagnosis of PDs.

Experimental phantom evaluation to identify robust positron emission tomography (PET) radiomic features

BackgroundRadiomics analysis usually involves, especially in multicenter and large hospital studies, different imaging protocols for acquisition, reconstruction, and processing of data. Differences in protocols can lead to differences in the quantification of the biomarker distribution, leading to radiomic feature variability. The aim of our study was to identify those radiomic features robust to the different degrading factors in positron emission tomography (PET) studies. We proposed the use of the standardized measurements of the European Association Research Ltd. (EARL) accreditation to retrospectively identify the radiomic features having low variability to the different systems and reconstruction protocols. In addition, we presented a reproducible procedure to identify PET radiomic features robust to PET/CT imaging metal artifacts. In 27 heterogeneous homemade phantoms for which ground truth was accurately defined by CT segmentation, we evaluated the segmentation accuracy and radiomic feature reliability given by the contrast-oriented algorithm (COA) and the 40% threshold PET segmentation. In the comparison of two data sets, robustness was defined by Wilcoxon rank tests, bias was quantified by Bland–Altman (BA) plot analysis, and strong correlations were identified by Spearman correlation test (r > 0.8 and p satisfied multiple test Bonferroni correction).ResultsForty-eight radiomic features were robust to system, 22 to resolution, 102 to metal artifacts, and 42 to different PET segmentation tools. Overall, only 4 radiomic features were simultaneously robust to all degrading factors. Although both segmentation approaches significantly underestimated the volume with respect to the ground truth, with relative deviations of −62 ± 36% for COA and −50 ± 44% for 40%, radiomic features derived from the ground truth were strongly correlated and/or robust to 98 radiomic features derived from COA and to 102 from 40%.ConclusionIn multicenter studies, we recommend the analysis of EARL accreditation measurements in order to retrospectively identify the robust PET radiomic features. Furthermore, 4 radiomic features (area under the curve of the cumulative SUV volume histogram, skewness, kurtosis, and gray-level variance derived from GLRLM after application of an equal probability quantization algorithm on the voxels within lesion) were robust to all degrading factors. In addition, the feasibility of 40% and COA segmentations for their use in radiomics analysis has been demonstrated.

Clinically feasible semi-automatic workflows for measuring metabolically active tumour volume in metastatic melanoma

PurposeMetabolically active tumour volume (MATV) is a potential quantitative positron emission tomography (PET) imaging biomarker in melanoma. Accumulating data indicate that low MATV may predict increased chance of response to immunotherapy and overall survival. However, metastatic melanoma can present with numerous (small) tumour lesions, making manual tumour segmentation time-consuming. The aim of this study was to evaluate multiple semi-automatic segmentation workflows to determine reliability and reproducibility of MATV measurements in patients with metastatic melanoma.MethodsAn existing cohort of 64 adult patients with histologically proven metastatic melanoma was used in this study. 18F-FDG PET/CT diagnostic baseline images were acquired using a European Association of Nuclear Medicine (EANM) Research Limited–accredited Siemens Biograph mCT PET/CT system (Siemens Healthineers, Knoxville, USA). PET data were analysed using manual, gradient-based segmentation and five different semi-automatic methods: three direct PET image–derived delineations (41MAX, A50P and SUV40) and two based on a majority-vote approach (MV2 and MV3), without and with (suffix ‘+’) manual lesion addition. Correlation between the different segmentation methods and their respective associations with overall survival was assessed.ResultsCorrelation between the MATVs derived by the manual segmentation and semi-automated tumour segmentations ranged from R2 = 0.41 for A50P to R2 = 0.85 for SUV40+ and MV2+, respectively. Manual MATV segmentation did not differ significantly from the semi-automatic methods SUV40 (∆MATV mean ± SD 0.08 ± 0.60 mL, P = 0.303), SUV40+ (∆MATV − 0.10 ± 0.51 mL, P = 0.126), MV2+ (∆MATV − 0.09 ± 0.62 mL, P = 0.252) and MV3+ (∆MATV − 0.03 ± 0.55 mL, P = 0.615). Log-rank tests showed statistically significant overall survival differences between above and below median MATV patients for all segmentation methods with areas under the ROC curves of 0.806 for manual segmentation and between 0.756 [41MAX] and 0.807 [MV3+] for semi-automatic segmentations.ConclusionsSimple and fast semi-automated FDG PET segmentation workflows yield accurate and reproducible MATV measurements that correlate well with manual segmentation in metastatic melanoma. The most readily applicable and user-friendly SUV40 method allows feasible MATV measurement in prospective multicentre studies required for validation of this potential PET imaging biomarker for clinical use.

Regional SUV quantification in hybrid PET/MR, a comparison of two atlas-based automatic brain segmentation methods

BackgroundQuantitative analysis of brain positron-emission tomography (PET) depends on structural segmentation, which can be time-consuming and operator-dependent when performed manually. Previous automatic segmentation usually registered subjects’ images onto an atlas template (defined as RSIAT here) for group analysis, which changed the individuals’ images and probably affected regional PET segmentation. In contrast, we could register atlas template to subjects’ images (RATSI), which created an individual atlas template and may be more accurate for PET segmentation. We segmented two representative brain areas in twenty Parkinson disease (PD) and eight multiple system atrophy (MSA) patients performed in hybrid positron-emission tomography/magnetic resonance imaging (PET/MR). The segmentation accuracy was evaluated using the Dice coefficient (DC) and Hausdorff distance (HD), and the standardized uptake value (SUV) measurements of these two automatic segmentation methods were compared, using manual segmentation as a reference.ResultsThe DC of RATSI increased, and the HD decreased significantly (P < 0.05) compared with the RSIAT in PD, while the results of one-way analysis of variance (ANOVA) found no significant differences in the SUVmean and SUVmax among the two automatic and the manual segmentation methods. Further, RATSI was used to compare regional differences in cerebral metabolism pattern between PD and MSA patients. The SUVmean in the segmented cerebellar gray matter for the MSA group was significantly lower compared with the PD group (P < 0.05), which is consistent with previous reports.ConclusionThe RATSI was more accurate for the caudate nucleus and putamen automatic segmentation and can be used for regional PET analysis in hybrid PET/MR.

Pancreatic Tumor Segmentation Based on Optimized K-Means Clustering and Saliency Map Model

Segmentation of positron emission tomography (PET) plays a major role in research and clinical applications. The segmentation of pancreatic tumors using PET / CT is challenging due to a significant amount of noise that may result in serious segmentation inaccuracies. The evaluation of the results of segmentation in medical imaging is due to the presence of a gold standard. Therefore, the performance evaluation of these methods would be necessary. This paper suggested a new object segmentation method that is based on K-means clustering with Saliency Maps. The K-means clustering approach restricts every pixel of the image that belongs to a single cluster. One drawback with using the K-means algorithm to segment objects is that segments are not connected and can be widely scattered. It is known that using saliency region, the approximate location of the desired object in the map can easily be identified. In this proposed method, the saliency map is used to distinguish the desired object cluster from the image from the background cluster, and then, to map the object clusters together Experimental results shows that the proposed algorithm outperforms dramatically in terms of visual plausibility and computational cost compared to state-of - the-art methods and achieves excellent performance for object segmentation.

A smart and operator independent system to delineate tumours in Positron Emission Tomography scans

Positron Emission Tomography (PET) imaging has an enormous potential to improve radiation therapy treatment planning offering complementary functional information with respect to other anatomical imaging approaches.The aim of this study is to develop an operator independent, reliable, and clinically feasible system for biological tumour volume delineation from PET images. Under this design hypothesis, we combine several known approaches in an original way to deploy a system with a high level of automation.The proposed system automatically identifies the optimal region of interest around the tumour and performs a slice-by-slice marching local active contour segmentation. It automatically stops when a “cancer-free” slice is identified. User intervention is limited at drawing an initial rough contour around the cancer region. By design, the algorithm performs the segmentation minimizing any dependence from the initial input, so that the final result is extremely repeatable.To assess the performances under different conditions, our system is evaluated on a dataset comprising five synthetic experiments and fifty oncological lesions located in different anatomical regions (i.e. lung, head and neck, and brain) using PET studies with 18F-fluoro-2-deoxy-d-glucose and 11C-labeled Methionine radio-tracers.Results on synthetic lesions demonstrate enhanced performances when compared against the most common PET segmentation methods. In clinical cases, the proposed system produces accurate segmentations (average dice similarity coefficient: 85.36 ± 2.94%, 85.98 ± 3.40%, 88.02 ± 2.75% in the lung, head and neck, and brain district, respectively) with high agreement with the gold standard (determination coefficient R2 = 0.98). We believe that the proposed system could be efficiently used in the everyday clinical routine as a medical decision tool, and to provide the clinicians with additional information, derived from PET, which can be of use in radiation therapy, treatment, and planning.

Automatic lesion detection and segmentation of 18F-FET PET in gliomas: A full 3D U-Net convolutional neural network study.

IntroductionAmino-acids positron emission tomography (PET) is increasingly used in the diagnostic workup of patients with gliomas, including differential diagnosis, evaluation of tumor extension, treatment planning and follow-up. Recently, progresses of computer vision and machine learning have been translated for medical imaging. Aim was to demonstrate the feasibility of an automated 18F-fluoro-ethyl-tyrosine (18F-FET) PET lesion detection and segmentation relying on a full 3D U-Net Convolutional Neural Network (CNN).MethodsAll dynamic 18F-FET PET brain image volumes were temporally realigned to the first dynamic acquisition, coregistered and spatially normalized onto the Montreal Neurological Institute template. Ground truth segmentations were obtained using manual delineation and thresholding (1.3 x background). The volumetric CNN was implemented based on a modified Keras implementation of a U-Net library with 3 layers for the encoding and decoding paths. Dice similarity coefficient (DSC) was used as an accuracy measure of segmentation.ResultsThirty-seven patients were included (26 [70%] in the training set and 11 [30%] in the validation set). All 11 lesions were accurately detected with no false positive, resulting in a sensitivity and a specificity for the detection at the tumor level of 100%. After 150 epochs, DSC reached 0.7924 in the training set and 0.7911 in the validation set. After morphological dilatation and fixed thresholding of the predicted U-Net mask a substantial improvement of the DSC to 0.8231 (+ 4.1%) was noted. At the voxel level, this segmentation led to a 0.88 sensitivity [95% CI, 87.1 to, 88.2%] a 0.99 specificity [99.9 to 99.9%], a 0.78 positive predictive value: [76.9 to 78.3%], and a 0.99 negative predictive value [99.9 to 99.9%].ConclusionsWith relatively high performance, it was proposed the first full 3D automated procedure for segmentation of 18F-FET PET brain images of patients with different gliomas using a U-Net CNN architecture.

Validating PET segmentation of thoracic lesions—is 4D PET necessary?

Respiratory-induced motions are prone to degrade the positron emission tomography (PET) signal with the consequent loss of image information and unreliable segmentations. This phantom study aims to assess the discrepancies relative to stationary PET segmentations, of widely used semi-automatic PET segmentation methods on heterogeneous target lesions influenced by motion during image acquisition. Three target lesions included dual F-18 Fluoro-deoxy-glucose (FDG) tracer concentrations as high- and low tracer activities relative to the background. Four different tracer concentration arrangements were segmented using three SUV threshold methods (Max40%, SUV40% and 2.5SUV) and a gradient based method (GradientSeg). Segmentations in static 3D-PET scans (PETsta) specified the reference conditions for the individual segmentation methods, target lesions and tracer concentrations. The motion included PET images followed a 4D-PET (PET4D) and a 3D-PET (PETmot) scan protocol. Moreover, motion-corrected PET images (PETdeb) were derived from the PETmot images. Segmentations in PET4D, PETmot and PETdeb were compared to the PETsta segmentations according to volume changes (ΔVol) and an error estimate (lowUptakeerror) for the lesion part covering the low tracer concentration. In PET4D images, all segmentation methods provided lowUptakeerror estimates equivalent to PETsta segmentations and, except for the Max40% segmentations, a slight volume expansion. In the PETmot images, the GradientSeg method results in an average 0.43 increased volume and an overestimation of 0.33 for the lowUptakeerror. The most accurate segmentations in PETmot, relative to PETsta, were accomplished by the 2.5SUV and SUV40% methods. In the PETdeb images, the GradientSeg method solitary provided segmentations equivalent to segmentation in PETsta images. The use of FDG with various tracer concentrations revealed, according to PETsta images, that the most constant segmentations for motion-corrected PET images (PET4D or PETdeb) were achieved using the GradientSeg method. In the absence of PET4D or PETdeb images, the 2.5SUV and SUV40% methods are most consistent to PETsta segmentations.

Regional tau deposition and subregion atrophy of medial temporal structures in early Alzheimer's disease: A combined positron emission tomography/magnetic resonance imaging study

IntroductionMolecular imaging and selective hippocampal subfield atrophy are a focus of recent Alzheimer's disease (AD) research. Here, we investigated correlations between molecular imaging and hippocampal subfields in early AD. MethodsWe investigated 18 patients with early AD and 18 healthy control subjects using 11C-Pittsburgh compound-B (PIB) positron emission tomography (PET) and 18F-THK5351 PET and automatic segmentation of hippocampal subfields with high-resolution T2-weighted magnetic resonance imaging. The PET images were normalized and underwent voxelwise regression analysis with each subregion volumes using SPM12. ResultsAs for 18F-THK5351 PET, the bilateral perirhinal cortex volumes were significantly associated with the ipsilateral or bilateral temporal lobar uptakes, whereas hippocampal subfields showed no correlations. 11C-PIB PET showed relatively broad negative correlation with the right cornu ammonis 3 volumes. DiscussionRegional tau deposition was correlated with extrahippocampal subregional atrophy and not with hippocampal subfields, possibly reflecting different underlying mechanisms of atrophy in early AD. Amyloid might be associated with right cornu ammonis 3 atrophy.

Generic and robust method for automatic segmentation of PET images using an active contour model.

Although positron emission tomography (PET) images have shown potential to improve the accuracy of targeting in radiation therapy planning and assessment of response to treatment, the boundaries of tumors are not easily distinguishable from surrounding normal tissue owing to the low spatial resolution and inherent noisy characteristics of PET images. The objective of this study is to develop a generic and robust method for automatic delineation of tumor volumes using an active contour model and to evaluate its performance using phantom and clinical studies. MASAC, a method for automatic segmentation using an active contour model, incorporates the histogram fuzzy C-means clustering, and localized and textural information to constrain the active contour to detect boundaries in an accurate and robust manner. Moreover, the lattice Boltzmann method is used as an alternative approach for solving the level set equation to make it faster and suitable for parallel programming. Twenty simulated phantom studies and 16 clinical studies, including six cases of pharyngolaryngeal squamous cell carcinoma and ten cases of nonsmall cell lung cancer, were included to evaluate its performance. Besides, the proposed method was also compared with the contourlet-based active contour algorithm (CAC) and Schaefer's thresholding method (ST). The relative volume error (RE), Dice similarity coefficient (DSC), and classification error (CE) metrics were used to analyze the results quantitatively. For the simulated phantom studies (PSs), MASAC and CAC provide similar segmentations of the different lesions, while ST fails to achieve reliable results. For the clinical datasets (2 cases with connected high-uptake regions excluded) (CSs), CAC provides for the lowest mean RE (-8.38% ± 27.49%), while MASAC achieves the best mean DSC (0.71 ± 0.09) and mean CE (53.92% ± 12.65%), respectively. MASAC could reliably quantify different types of lesions assessed in this work with good accuracy, resulting in a mean RE of -13.35% ± 11.87% and -11.15% ± 23.66%, a mean DSC of 0.89 ± 0.05 and 0.71 ± 0.09, and a mean CE of 19.19% ± 7.89% and 53.92% ± 12.65%, for PSs and CSs, respectively. The authors' results demonstrate that the developed novel PET segmentation algorithm is applicable to various types of lesions in the authors' study and is capable of producing accurate and consistent target volume delineations, potentially resulting in reduced intraobserver and interobserver variabilities observed when using manual delineation and improved accuracy in treatment planning and outcome evaluation.

A fully automatic, threshold-based segmentation method for the estimation of the Metabolic Tumor Volume from PET images: validation on 3D printed anthropomorphic oncological lesions

18F-Fluorodeoxyglucose (18F-FDG) Positron Emission Tomography (PET) is a standard functional diagnostic technique to in vivo image cancer. Different quantitative paramters can be extracted from PET images and used as in vivo cancer biomarkers. Between PET biomarkers Metabolic Tumor Volume (MTV) has gained an important role in particular considering the development of patient-personalized radiotherapy treatment for non-homogeneous dose delivery. Different imaging processing methods have been developed to define MTV. The different proposed PET segmentation strategies were validated in ideal condition (e.g. in spherical objects with uniform radioactivity concentration), while the majority of cancer lesions doesn't fulfill these requirements. In this context, this work has a twofold objective: 1) to implement and optimize a fully automatic, threshold-based segmentation method for the estimation of MTV, feasible in clinical practice 2) to develop a strategy to obtain anthropomorphic phantoms, including non-spherical and non-uniform objects, miming realistic oncological patient conditions. The developed PET segmentation algorithm combines an automatic threshold-based algorithm for the definition of MTV and a k-means clustering algorithm for the estimation of the background. The method is based on parameters always available in clinical studies and was calibrated using NEMA IQ Phantom. Validation of the method was performed both in ideal (e.g. in spherical objects with uniform radioactivity concentration) and non-ideal (e.g. in non-spherical objects with a non-uniform radioactivity concentration) conditions. The strategy to obtain a phantom with synthetic realistic lesions (e.g. with irregular shape and a non-homogeneous uptake) consisted into the combined use of standard anthropomorphic phantoms commercially and irregular molds generated using 3D printer technology and filled with a radioactive chromatic alginate. The proposed segmentation algorithm was feasible in a clinical context and showed a good accuracy both in ideal and in realistic conditions.

A novel phantom technique for evaluating the performance of PET auto-segmentation methods in delineating heterogeneous and irregular lesions.

BackgroundPositron Emission Tomography (PET)-based automatic segmentation (PET-AS) methods can improve tumour delineation for radiotherapy treatment planning, particularly for Head and Neck (H&N) cancer. Thorough validation of PET-AS on relevant data is currently needed. Printed subresolution sandwich (SS) phantoms allow modelling heterogeneous and irregular tracer uptake, while providing reference uptake data. This work aimed to demonstrate the usefulness of the printed SS phantom technique in recreating complex realistic H&N radiotracer uptake for evaluating several PET-AS methods.MethodsTen SS phantoms were built from printouts representing 2mm-spaced slices of modelled H&N uptake, printed using black ink mixed with 18F-fluorodeoxyglucose, and stacked between 2mm thick plastic sheets. Spherical lesions were modelled for two contrasted uptake levels, and irregular and spheroidal tumours were modelled for homogeneous, and heterogeneous uptake including necrotic patterns. The PET scans acquired were segmented with ten custom PET-AS methods: adaptive iterative thresholding (AT), region growing, clustering applied to 2 to 8 clusters, and watershed transform-based segmentation. The difference between the resulting contours and the ground truth from the image template was evaluated using the Dice Similarity Coefficient (DSC), Sensitivity and Positive Predictive value.ResultsRealistic H&N images were obtained within 90 min of preparation. The sensitivity of binary PET-AS and clustering using small numbers of clusters dropped for highly heterogeneous spheres. The accuracy of PET-AS methods dropped between 4% and 68% for irregular lesions compared to spheres of the same volume. For each geometry and uptake modelled with the SS phantoms, we report the number of clusters resulting in optimal segmentation. Radioisotope distributions representing necrotic uptakes proved most challenging for most methods. Two PET-AS methods did not include the necrotic region in the segmented volume.ConclusionsPrinted SS phantoms allowed identifying advantages and drawbacks of the different methods, determining the most robust PET-AS for the segmentation of heterogeneities and complex geometries, and quantifying differences across methods in the delineation of necrotic lesions. The printed SS phantom technique provides key advantages in the development and evaluation of PET segmentation methods and has a future in the field of radioisotope imaging.

PET image segmentation using a Gaussian mixture model and Markov random fields.

BackgroundClassification algorithms for positron emission tomography (PET) images support computational treatment planning in radiotherapy. Common clinical practice is based on manual delineation and fixed or iterative threshold methods, the latter of which requires regression curves dependent on many parameters.MethodsAn improved statistical approach using a Gaussian mixture model (GMM) is proposed to obtain initial estimates of a target volume, followed by a correction step based on a Markov random field (MRF) and a Gibbs distribution to account for dependencies among neighboring voxels. In order to evaluate the proposed algorithm, phantom measurements of spherical and non-spherical objects with the smallest diameter being 8 mm were performed at signal-to-background ratios (SBRs) between 2.06 and 9.39. Additionally 68Ga-PET data from patients with lesions in the liver and lymph nodes were evaluated.ResultsThe proposed algorithm produces stable results for different reconstruction algorithms and different lesion shapes. Furthermore, it outperforms all threshold methods regarding detection rate, determines the spheres’ volumes more accurately than fixed threshold methods, and produces similar values as iterative thresholding. In a comparison with other statistical approaches, the algorithm performs equally well for larger volumes and even shows improvements for small volumes and SBRs. The comparison with experts’ manual delineations on the clinical data shows the same qualitative behavior as for the phantom measurements.ConclusionsIn conclusion, a generic probabilistic approach that does not require data measured beforehand is presented whose performance, robustness, and swiftness make it a feasible choice for PET segmentation.

Simultaneous Reconstruction and Segmentation of Dynamic PET via Low-Rank and Sparse Matrix Decomposition.

Although of great clinical value, accurate and robust reconstruction and segmentation of dynamic positron emission tomography (PET) images are great challenges due to low spatial resolution and high noise. In this paper, we propose a unified framework that exploits temporal correlations and variations within image sequences based on low-rank and sparse matrix decomposition. Thus, the two separate inverse problems, PET image reconstruction and segmentation, are accomplished in a simultaneous fashion. Considering low signal to noise ratio and piece-wise constant assumption of PET images, we also propose to regularize low-rank and sparse matrices with vectorial total variation norm. The resulting optimization problem is solved by augmented Lagrangian multiplier method with variable splitting. The effectiveness of proposed approach is validated on realistic Monte Carlo simulation datasets and the real patient data.

Pathology-validated PET image data sets and their role in PET segmentation

Positron emission tomography (PET)/computed tomography has recently been finding broader application for the diagnosis, treatment and therapy assessment of malignant disease. Accurate definition of the tumor border is extremely important for the success of localized tumor therapies. PET promises to provide the metabolically active tumor volume and, at present, it is used for target definition in a variety of tumors. This process is, however, subject to uncertainties of different origin. Resolving these uncertainties is challenging, since validating PET images and segmentation contours against tumor pathology is experimentally difficult. In addition to accurate lesion contouring, this challenges validation of PET tracers and investigations of tumor functional heterogeneity. In this paper, we briefly review the present studies providing PET image data sets with pathology validation. We focus on the specimen handling techniques aimed at achieving higher geometrical accuracy of the pathology-derived “ground truth”. We also summarize the main findings obtained for the PET segmentation techniques which have been tested with the help of these data sets. Finally, we provide a critical summary of the current state of the art in pathological validation of PET images and briefly discuss future possibilities in this direction.

Evaluation of advanced automatic PET segmentation methods using nonspherical thin‐wall inserts

The use of positron emission tomography (PET) within radiotherapy treatment planning requires the availability of reliable and accurate segmentation tools. PET automatic segmentation (PET-AS) methods have been recommended for the delineation of tumors, but there is still a lack of thorough validation and cross-comparison of such methods using clinically relevant data. In particular, studies validating PET segmentation tools mainly use phantoms with thick plastic walls inserts of simple spherical geometry and have not specifically investigated the effect of the target object geometry on the delineation accuracy. Our work therefore aimed at generating clinically realistic data using nonspherical thin-wall plastic inserts, for the evaluation and comparison of a set of eight promising PET-AS approaches. Sixteen nonspherical inserts were manufactured with a plastic wall of 0.18 mm and scanned within a custom plastic phantom. These included ellipsoids and toroids derived with different volumes, as well as tubes, pear- and drop-shaped inserts with different aspect ratios. A set of six spheres of volumes ranging from 0.5 to 102 ml was used for a baseline study. A selection of eight PET-AS methods, written in house, was applied to the images obtained. The methods represented promising segmentation approaches such as adaptive iterative thresholding, region-growing, clustering and gradient-based schemes. The delineation accuracy was measured in terms of overlap with the computed tomography reference contour, using the dice similarity coefficient (DSC), and error in dimensions. The delineation accuracy was lower for nonspherical inserts than for spheres of the same volume in 88% cases. Slice-by-slice gradient-based methods, showed particularly lower DSC for tori (DSC < 0.5), caused by a failure to recover the object geometry. The region-growing method reached high levels of accuracy for most inserts (DSC > 0.76 except for tori) but showed the largest errors in the recovery of pears and drops dimensions (higher than 10% and 30% of the true length, respectively). Large errors were visible for one of the gradient-based contouring methods when delineating drop-shaped inserts. Low DSC due to systematic underestimation of the volumes was observed for our fuzzy clustering method when using nonspherical inserts. The adaptive iterative thresholding method produced the highest DSC score on our nonspherical dataset (DSC > 0.83, except for tori) and showed robustness to the insert geometry. This study investigated the accuracy of eight PET-AS methods for the delineation of objects with a range of nonspherical geometries. The authors' results confirmed the robustness of some segmentation approaches, but also showed the weaknesses of some of the other methods implemented, which were not observed with spherical inserts. This work therefore highlights the importance of using a variety of thin-wall inserts with complex geometries for the validation of PET-AS methods and provided useful information for further development of the methods tested.