Development of a new toolbox for mouse PET-CT brain image analysis fully based on CT images and validation in a PD mouse model.

Automatic analysis toolboxes are popular in brain image analysis, both in clinical and in preclinical practices. In this regard, we proposed a new toolbox for mouse PET-CT brain image analysis including a new Statistical Parametric Mapping-based template and a pipeline for image registration of PET-CT images based on CT images. The new templates is compatible with the common coordinate framework (CCFv3) of the Allen Reference Atlas (ARA) while the CT based registration step allows to facilitate the analysis of mouse PET-CT brain images. From the ARA template, we identified 27 volumes of interest that are relevant for in vivo imaging studies and provided binary atlas to describe them. We acquired 20 C57BL/6 mice with [18F]FDG PET-CT, and 12 of them underwent 3D T2-weighted high-resolution MR scans. All images were elastically registered to the ARA atlas and then averaged. High-resolution MR images were used to validate a CT-based registration pipeline. The resulting method was applied to a mouse model of Parkinson's disease subjected to a test-retest study (n = 6) with the TSPO-specific radioligand [18F]VC701. The identification of regions of microglia/macrophage activation was performed in comparison to the Ma and Mirrione template. The new toolbox identified 11 (6 after false discovery rate adjustment, FDR) brain sub-areas of significant [18F]VC701 uptake increase versus the 4 (3 after FDR) macro-regions identified by the Ma and Mirrione template. Moreover, these 11 areas are functionally connected as found by applying the Mouse Connectivity tool of ARA. In conclusion, we developed a mouse brain atlas tool optimized for PET-CT imaging analysis that does not require MR. This tool conforms to the CCFv3 of ARA and could be applied to the analysis of mouse brain disease models.

68Ga-DOTATOC PET/MR imaging and radiomic parameters in predicting histopathological prognostic factors in patients with pancreatic neuroendocrine well-differentiated tumours.

PURPOSE: To explore the role of fully hybrid 68Ga-DOTATOC PET/MR imaging and radiomic parameters in predicting histopathological prognostic factors in patients with pancreatic neuroendocrine tumours (PanNETs) undergoing surgery. METHODS: One hundred eighty-seven consecutive 68Ga-DOTATOC PET/MRI scans (March 2018-June 2020) performed for gastroenteropancreatic neuroendocrine tumour were retrospectively evaluated; 16/187 patients met the eligibility criteria (68Ga-DOTATOC PET/MRI for preoperative staging of PanNET and availability of histological data). PET/MR scans were qualitatively and quantitatively interpreted, and the following imaging parameters were derived: PET-derived SUVmax, SUVmean, somatostatin receptor density (SRD), total lesion somatostatin receptor density (TLSRD), and MRI-derived apparent diffusion coefficient (ADC), arterial and late enhancement, necrosis, cystic degeneration, and maximum diameter. Additionally, first-, second-, and higher-order radiomic parameters were extracted from both PET and MRI scans. Correlations with several PanNETs' histopathological prognostic factors were evaluated using Spearman's coefficient, while the area under the curve (AUC) of the receiver operating characteristic (ROC) curve was used to evaluate parameters' predictive performance. RESULTS: Primary tumour was detected in all 16 patients (15/16 by 68Ga-DOTATOC PET and 16/16 by MRI). SUVmax and SUVmean resulted good predictors of lymphnodal (LN) involvement (AUC of 0.850 and 0.783, respectively). Second-order radiomic parameters GrayLevelVariance and HighGrayLevelZoneEmphasis extracted from T2 MRI demonstrated significant correlations with LN involvement (adjusted p = 0.009), also showing good predictive performance (AUC = 0.992). CONCLUSION: This study demonstrates the role of the fully hybrid PET/MRI tool for the synergic function of imaging parameters extracted by the two modalities and highlights the potentiality of imaging and radiomic parameters in assessing histopathological features of PanNET aggressiveness.

Radiomics in pancreatic neuroendocrine tumors: methodological issues and clinical significance.

PURPOSE: To present the state-of-art of radiomics in the context of pancreatic neuroendocrine tumors (PanNETs), with a focus on the methodological and technical approaches used, to support the search of guidelines for optimal applications. Furthermore, an up-to-date overview of the current clinical applications of radiomics in the field of PanNETs is provided. METHODS: Original articles were searched on PubMed and Science Direct with specific keywords. Evaluations of the selected studies have been focused mainly on (i) the general radiomic workflow and the assessment of radiomic features robustness/reproducibility, as well as on the major clinical applications and investigations accomplished so far with radiomics in the field of PanNETs: (ii) grade prediction, (iii) differential diagnosis from other neoplasms, (iv) assessment of tumor behavior and aggressiveness, and (v) treatment response prediction. RESULTS: Thirty-one articles involving PanNETs radiomic-related objectives were selected. In regard to the grade differentiation task, yielded AUCs are currently in the range of 0.7-0.9. For differential diagnosis, the majority of studies are still focused on the preliminary identification of discriminative radiomic features. Limited information is known on the prediction of tumors aggressiveness and of treatment response. CONCLUSIONS: Radiomics is recently expanding in the setting of PanNETs. From the analysis of the published data, it is emerging how, prior to clinical application, further validations are necessary and methodological implementations require optimization. Nevertheless, this new discipline might have the potential in assisting the current urgent need of improving the management strategies in PanNETs patients.

PET textural features stability and pattern discrimination power for radiomics analysis: An "ad-hoc" phantoms study.

PURPOSE: The analysis of PET images by textural features, also known as radiomics, shows promising results in tumor characterization. However, radiomic metrics (RMs) analysis is currently not standardized and the impact of the whole processing chain still needs deep investigation. We characterized the impact on RM values of: i) two discretization methods, ii) acquisition statistics, and iii) reconstruction algorithm. The influence of tumor volume and standardized-uptake-value (SUV) on RM was also investigated. METHODS: The Chang-Gung-Image-Texture-Analysis (CGITA) software was used to calculate 39 RMs using phantom data. Thirty noise realizations were acquired to measure statistical effect size indicators for each RM. The parameter η2 (fraction of variance explained by the nuisance factor) was used to assess the effect of categorical variables, considering η2 < 20% and 20% < η2 < 40% as representative of a "negligible" and a "small" dependence respectively. The Cohen's d was used as discriminatory power to quantify the separation of two distributions. RESULTS: We found the discretization method based on fixed-bin-number (FBN) to outperform the one based on fixed-bin-size in units of SUV (FBS), as the latter shows a higher SUV dependence, with 30 RMs showing η2 > 20%. FBN was also less influenced by the acquisition and reconstruction setup:with FBN 37 RMs had η2 < 40%, only 20 with FBS. Most RMs showed a good discriminatory power among heterogeneous PET signals (for FBN: 29 out of 39 RMs with d > 3). CONCLUSIONS: For RMs analysis, FBN should be preferred. A group of 21 RMs was suggested for PET radiomics analysis.

Simultaneous reconstruction of attenuation and activity in cardiac PET can remove CT misalignment artifacts.

BACKGROUND: Misalignment between positron emission tomography (PET) and computed tomography (CT) data is known to generate artifactual defects in cardiac PET images due to imprecise attenuation correction (AC). In this work, the use of a maximum likelihood attenuation and activity (MLAA) algorithm is proposed to avoid such artifacts in time-of-flight (TOF) PET. METHODS: MLAA was implemented and tested using a thorax/heart phantom and retrospectively on fourteen (13)N-ammonia PET/CT perfusion studies. Global and local misalignments between PET and CT data were generated by shifting matched CT images or using CT data representative of the end-inspiration phase. PET images were reconstructed with MLAA and a 3D-ordered-subsets-expectation-maximization (OSEM)-TOF algorithm. Images obtained with 3D-OSEM-TOF and matched CT were used as references. These images were compared (qualitatively and semi-quantitatively) with those reconstructed with 3D-OSEM-TOF and MLAA for which a misaligned CT was used, respectively, for AC and initialization. RESULTS: Phantom experiment proved the capability of MLAA to converge toward the correct emission and attenuation distributions using, as input, only PET emission data, but convergence was very slow. Initializing MLAA with phantom CT images markedly improved convergence speed. In patient studies, when shifted or end-inspiration CT images were used for AC, 3D-OSEM-TOF reconstructions showed artifacts of increasing severity, size, and frequency with increasing mismatch. Such artifacts were absent in the corresponding MLAA images. CONCLUSION: The proposed implementation of the MLAA algorithm is a feasible and robust technique to avoid AC mismatch artifacts in cardiac PET studies provided that a CT of the source is available, even if poorly aligned.

Evaluation of image reconstruction algorithms encompassing Time-Of-Flight and Point Spread Function modelling for quantitative cardiac PET: phantom studies.

BACKGROUND: To perform kinetic modelling quantification, PET dynamic data must be acquired in short frames, where different critical conditions are met. The accuracy of reconstructed images influences quantification. The added value of Time-Of-Flight (TOF) and Point Spread Function (PSF) in cardiac image reconstruction was assessed. METHODS: A static phantom was used to simulate two extreme conditions: (i) the bolus passage and (ii) the steady uptake. Various count statistics and independent noise realisations were considered. A moving phantom filled with two different radionuclides was used to simulate: (i) a great range of contrasts and (ii) the cardio/respiratory motion. Analytical and iterative reconstruction (IR) algorithms also encompassing TOF and PSF modelling were evaluated. RESULTS: Both analytic and IR algorithms provided good results in all the evaluated conditions. The amount of bias introduced by IR was found to be limited. TOF allowed faster convergence and lower noise levels. PSF achieved near full myocardial activity recovery in static conditions. Motion degraded performances, but the addition of both TOF and PSF maintained the best overall behaviour. CONCLUSIONS: IR accounting for TOF and PSF can be recommended for the quantification of dynamic cardiac PET studies as they improve the results compared to analytic and standard IR.

Motion-Tracking Hardware and Advanced Applications in PET and PET/CT.

Respiratory and cardiac motions represent important sources of image degradation in both PET and computed tomography (CT) studies that need to be taken into account and compensated to improve image quality and quantitative accuracy. This review describes the hardware needed to perform respiratory and cardiac gating with PET and PET/CT systems. In particular, most of the proposed motion-tracking devices for the management of respiratory, cardiac, and multidimensional movements are described and compared. Some advanced applications in PET and PET/CT made possible by the gating technology are considered and analyzed.

Physical performance of the new hybrid PET∕CT Discovery-690.

PURPOSE: The aim of this work was the assessment of the physical performance of the new hybrid PET∕CT system: Discovery-690. METHODS: The Discovery-690 combines a lutetium-yttrium-orthosilicate (LYSO) block detector designed PET tomograph with a 64-slice CT scanner. The system is further characterized by a dedicated powerful computing platform implementing fully 3D-PET iterative reconstruction algorithms. These algorithms can account for time of flight (TOF) information and∕or a 3D model of the PET point spread function (PSF). PET physical performance was measured following NEMA NU-2-2007 procedures. Furthermore, specific tests were used: (i) to measure the energy and timing resolution of the PET system and (ii) to evaluate image quality, by using phantoms representing different clinical conditions (e.g., brain and whole body). Data processing and reconstructions were performed as required by standard procedures. Further reconstructions were carried out to evaluate the performance of the new reconstruction algorithms. In particular, four algorithms were considered for the reconstruction of the PET data: (i) HD = standard configuration, without TOF and PSF, (ii) TOF = HD + TOF, (iii) PSF = HD + PSF, and (iv) TOFPSF = HD + TOF + PSF. RESULTS: The transverse (axial) spatial resolution values were 4.70 (4.74) mm and 5.06 (5.55) mm at 1 cm and 10 cm off axis, respectively. Sensitivity (average between 0 and 10 cm) was 7.5 cps∕kBq. The noise equivalent count rate (NECR) peak was 139.1 kcps at 29.0 kBq∕ml. The scatter fraction at the NECR peak was 37%. The correction accuracy for the dead time losses and random event counts had a maximum absolute error below the NECR peak of 2.09%. The average energy and timing resolution were 12.4% and 544.3 ps, respectively. PET image quality was evaluated with the NEMA IEC Body phantom by using four reconstruction algorithms (HD, TOF, PSF, and TOFPSF), as previously described. The hot contrast (after 3 iterations and for a lesion∕background activity ratio of 4:1) for the spheres of 10, 13, 17, and 22 mm was (HD) 29.8, 45.4, 55.4, and 68.1%; (TOF) 39.9, 53.5, 62.7, and 72.2%; (PSF) 28.3, 47.3, 60.4, and 71.8%; (TOFPSF) 43.8, 62.9, 70.6, and 76.4%. The cold contrast for the spheres of 28 and 37 mm was (HD) 62.4 and 65.2%; (TOF) 77.1 and 81.4%; (PSF) 62.0 and 65.2%; (TOFPSF) 77.3 and 81.6%. Similar hot and cold contrast trends were found during the analyses of other phantoms representing different clinical conditions (brain and whole body). Nevertheless, the authors observed a predominant role of either TOF or PSF, depending on the specific characteristics and dimensions of the phantoms. CONCLUSIONS: Discovery-690 shows very good PET physical performance for all the standard NEMA NU-2-2007 measurements. Furthermore, the new reconstruction algorithms available for PET data (TOF and PSF) allow further improvements of the D-690 image quality performance both qualitatively and quantitatively.