The Feasibility of Differentiating Lewy Body Dementia and Alzheimer's Disease by Deep Learning Using ECD SPECT Images.

The correct differential diagnosis of dementia has an important impact on patient treatment and follow-up care strategies. Tc-99m-ECD SPECT imaging, which is low cost and accessible in general clinics, is used to identify the two common types of dementia, Alzheimer's disease (AD) and Lewy body dementia (LBD). Two-stage transfer learning technology and reducing model complexity based on the ResNet-50 model were performed using the ImageNet data set and ADNI database. To improve training accuracy, the three-dimensional image was reorganized into three sets of two-dimensional images for data augmentation and ensemble learning, then the performance of various deep learning models for Tc-99m-ECD SPECT images to distinguish AD/normal cognition (NC), LBD/NC, and AD/LBD were investigated. In the AD/NC, LBD/NC, and AD/LBD tasks, the AUC values were around 0.94, 0.95, and 0.74, regardless of training models, with an accuracy of 90%, 87%, and 71%, and F1 scores of 89%, 86%, and 76% in the best cases. The use of transfer learning and a modified model resulted in better prediction results, increasing the accuracy by 32% for AD/NC. The proposed method is practical and could rapidly utilize a deep learning model to automatically extract image features based on a small number of SPECT brain perfusion images in general clinics to objectively distinguish AD and LBD.

Detection of Alzheimer's disease using ECD SPECT images by transfer learning from FDG PET.

OBJECTIVE: To develop a practical method to rapidly utilize a deep learning model to automatically extract image features based on a small number of SPECT brain perfusion images in general clinics to objectively evaluate Alzheimer's disease (AD). METHODS: For the properties of low cost and convenient access in general clinics, Tc-99-ECD SPECT imaging data in brain perfusion detection was used in this study for AD detection. Two-stage transfer learning based on the Inception v3 network model was performed using the ImageNet dataset and ADNI database. To improve training accuracy, the three-dimensional image was reorganized into three sets of two-dimensional images for data augmentation and ensemble learning. The effect of pre-training parameters for Tc-99m-ECD SPECT image to distinguish AD from normal cognition (NC) was investigated, as well as the effect of the sample size of F-18-FDG PET images used in pre-training. The same model was also fine-tuned for the prediction of the MMSE score from the Tc-99m-ECD SPECT image. RESULTS: The AUC values of w/wo pre-training parameters for Tc-99m-ECD SPECT image to distinguish AD from NC were 0.86 and 0.90. The sensitivity, specificity, precision, accuracy, and F1 score were 100%, 75%, 76%, 86%, and 86%, respectively for the training model with 1000 cases of F-18-FDG PET image for pre-training. The AUC values for various sample sizes of the training dataset (100, 200, 400, 800, 1000 cases) for pre-training were 0.86, 0.91, 0.95, 0.97, and 0.97. Regardless of the pre-training condition ECD dataset used, the AUC value was greater than 0.85. Finally, predicting cognitive scores and MMSE scores correlated (R2 = 0.7072). CONCLUSIONS: With the ADNI pre-trained model, the sensitivity and accuracy of the proposed deep learning model using SPECT ECD perfusion images to differentiate AD from NC were increased by approximately 30% and 10%, respectively. Our study indicated that the model trained on PET FDG metabolic imaging for the same disease could be transferred to a small sample of SPECT cerebral perfusion images. This model will contribute to the practicality of SPECT cerebral perfusion images using deep learning technology to objectively recognize AD.

Time of flight dual photon emission computed tomography.

Time-of-flight dual photon emission computed tomography (TOF-DuPECT) is an imaging system that can obtain radionuclide distributions using time information recorded from two cascade-decay photons. The potential decay locations in the image space, a hyperbolic response curve, can be determined via time-difference-of-arrival (TDOA) estimations from two instantaneous coincidence photons. In this feasibility study, Monte Carlo simulations were performed to generate list-mode coincidence data. A full-ring positron emission tomography-like detection system geometry was built in the simulation environment. A contrast phantom and a Jaszczak-like phantom filled with Selenium-75 (Se-75) were used to evaluate the image quality. A TOF-DuPECT system with varying coincidence time resolution (CTR) was then evaluated. We used the stochastic origin ensemble (SOE) algorithm to reconstruct images from the recorded list-mode data. The results indicate that the SOE method can be successfully employed for the TOF-DuPECT system and can achieve acceptable image quality when the CTR is less than 100 ps. Therefore, the TOF-DuPECT imaging system is feasible. With the improvement of the detector with time, future implementations and applications of TOF-DuPECT are promising. Further quantitative imaging techniques such as attenuation and scatter corrections for the TOF-DuPECT system will be developed in future.

Intratumor Heterogeneity Assessed by 18F-FDG PET/CT Predicts Treatment Response and Survival Outcomes in Patients with Hodgkin Lymphoma.

RATIONALE AND OBJECTIVES: Radiomic analysis of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) images enables the extraction of quantitative information of intratumour heterogeneity. This study investigated whether the baseline 18F-FDG PET/CT radiomics can predict treatment response and survival outcomes in patients with Hodgkin lymphoma (HL). MATERIALS AND METHODS: Thirty-five patients diagnosed with HL who underwent 18F-FDG PET/CT scans before and during chemotherapy were retrospectively enrolled in this investigation. For each patient, we extracted 709 radiomic features from pretreatment PET/CT images. Clinical variables (age, gender, B symptoms, bulky tumor, and disease stage) and radiomic signatures (intensity, texture, and wavelet) were analyzed according to response to therapy, progression-free survival (PFS), and overall survival (OS). Receiver operating characteristic curve, logistic regression, and Cox proportional hazards model were used to examine potential predictive and prognostic factors. RESULTS: High-intensity run emphasis (HIR) of PET and run-length nonuniformity (RLNU) of CT extracted from gray-level run-length matrix (GLRM) in high-frequency wavelets were independent predictive factors for the treatment response (odds ratio [OR] = 36.4, p = 0.014; OR = 30.4, p = 0.020). Intensity nonuniformity (INU) of PET and wavelet short run emphasis (SRE) of CT from GLRM and Ann Arbor stage were independently related to PFS (hazard ratio [HR] = 9.29, p = 0.023; HR = 18.40, p = 0.012; HR = 7.46, p = 0.049). Zone-size nonuniformity (ZSNU) of PET from gray-level size zone matrix (GLSZM) was independently associated with OS (HR = 41.02, p = 0.001). Based on these factors, a prognostic stratification model was devised for the risk stratification of patients. The proposed model allowed the identification of four risk groups for PFS and OS (p < 0.001 and p < 0.001). CONCLUSION: HIR_GLRMPET and RLNU_GLRMCT in high-frequency wavelets serve as independent predictive factors for treatment response. ZSNU_GLSZMPET, INU_GLRMPET, and wavelet SRE_GLRMCT serve as independent prognostic factors for survival outcomes. The present study proposes a prognostic stratification model that may be clinically beneficial in guiding risk-adapted treatment strategies for patients with HL.

Prognostic Value of Pretreatment Radiomic Features of 18F-FDG PET in Patients With Hodgkin Lymphoma.

PURPOSE: This study investigated whether a radiomic analysis of pretreatment F-FDG PET can predict prognosis in patients with Hodgkin lymphoma (HL). METHODS: Forty-two patients who were diagnosed as having HL and underwent pretreatment F-FDG PET scans were retrospectively enrolled. For each patient, we extracted 450 radiomic features from PET images. The prognostic significance of the clinical and radiomic features was assessed in relation to progression-free survival (PFS) and overall survival (OS). Receiver operating characteristic curve, Cox proportional hazards regression, and Kaplan-Meier analyses were performed to examine the potential independent predictors and to evaluate the predictive value. RESULTS: Intensity nonuniformity extracted from a gray-level run-length matrix and the Ann Arbor stage were independently associated with PFS (hazard ratio [HR] = 22.8, P < 0.001; HR = 7.6, P = 0.024) and OS (HR = 14.5, P = 0.012; HR = 8.5, P = 0.048), respectively. In addition, SUV kurtosis was an independent prognosticator for PFS (HR = 6.6, P = 0.026). We devised a prognostic scoring system based on these 3 risk predictors. The proposed scoring system further improved the risk stratification of the current staging classification (P < 0.001). CONCLUSIONS: The radiomic feature intensity nonuniformity is an independent prognostic predictor of PFS and OS in patients with HL. We devised a prognostic scoring system, which may be more beneficial for patient risk stratification in guiding therapy compared with the current Ann Arbor staging system.

A noise smoothing origin ensemble algorithm based on local filtering.

An origin ensemble (OE) image reconstruction algorithm can be used for the fast reconstruction of unconventional geometrical images, e.g. in a Compton camera (CC) system. Due to the low-count rate in the emission data, the reconstructed image is often noisy and inhomogeneous in density. In this study, we propose a way to smooth out the noise in the OE algorithm. During the OE reconstruction, the algorithm stochastically modifies the current location to a random new voxel along the probable corresponding curve of each event depending on the relative event density of the new and old locations. In the original OE technique, the event density is simply the number of events in the voxel. In the proposed method, the event density is estimated from the filtering of a kernel window centered on the voxel. Incorporating the regional filtering is similar to performing an OE algorithm on a smoothed image at each iteration and enables the reconstruction of a smoother image. A Flangeless Esser PET phantom and a multi-activity phantom are used to study the property of the new reconstruction algorithm. The results indicate that the proposed method performs better than a conventional OE algorithm in terms of normalized mean square error (NMSE) and structural similarity (SSIM). Both contrast noise ratio (CNR) and reconstruction accuracy of the new method are better than the conventional OE algorithm and their performances improve with the increase of object size. The median-OE possesses the highest overall image quality and recovery rate among the three filter-OE algorithms and is the method of choice for image reconstruction. Comparing to conventional post-smoothing OEs, the NMSE of median-OE improves 57.6% (46.9%) and the SSIM increased by 73.2% (51.1%) for the Esser (multi-activity) phantom. The proposed OE algorithm is simple and efficient for noise smoothing without complex calculations and highly suited for low-count cases.

A comparison of dose distributions in gross tumor volume between boron neutron capture therapy alone and combined boron neutron capture therapy plus intensity modulation radiation therapy for head and neck cancer.

Nine patients with recurrent head and neck (H&N) cancer received boron neutron capture therapy (BNCT) in one fraction at the Tsing-Hua Open pool reactor (THOR) utilizing the THORplan treatment planning system (TPS). The aims of the present study were to evaluate the use of intensity modulated radiation therapy (IMRT) of 45 Gy in 20 fractions to compensate for the dose heterogeneity in gross tumor volume observed with single-fraction BNCT with mean prescription dose 19 Gy (w), and to evaluate planning quality indices of simulated BNCT+IMRT versus single-fraction BNCT alone. All IMRT plans were generated using the Eclipse TPS which employs the analytical anisotropic algorithm. The conformity index for the gross tumor volume (GTV) was better for the BNCT+IMRT plan than for the BNCT-alone plan (p = 0.003). In addition, the BNCT+IMRT plan provided significantly better homogeneity in the GTV (p = 0.03). The cold spots in inhomogeneous dose distribution in the BNCT plan may be a key factor for H&N cancer recurrence. Our results suggest that single-fraction BNCT combined with compensated multi-fraction IMRT improves treatment homogeneity and conformity than single-fraction BNCT alone, especially for tumor volumes >100 cm3, and possibly increases local tumor control.

The Dosimetric Impact of Shifts in Patient Positioning during Boron Neutron Capture Therapy for Brain Tumors.

Unlike conventional photon radiotherapy, sophisticated patient positioning tools are not available for boron neutron capture therapy (BNCT). Thus, BNCT remains vulnerable to setup errors and intra-fractional patient motion. The aim of this study was to estimate the impact of deviations in positioning on the dose administered by BNCT for brain tumors at the Tsing Hua open-pool reactor (THOR). For these studies, a simulated head model was generated based on computed tomography (CT) images of a patient with a brain tumor. A cylindrical brain tumor 3 cm in diameter and 5 cm in length was modeled at distances of 6.5 cm and 2.5 cm from the posterior scalp of this head model (T6.5 cm and T2.5 cm, respectively). Radiation doses associated with positioning errors were evaluated for each distance, including left and right shifts, superior and inferior shifts, shifts from the central axis of the beam aperture, and outward shifts from the surface of the beam aperture. Rotational and tilting effects were also evaluated. The dose prescription was 20 Gray-equivalent (Gy-Eq) to 80 % of the tumor. The treatment planning system, NCTPlan, was used to perform dose calculations. The average decreases in mean tumor dose for T6.5 cm for the 1 cm, 2 cm, and 3 cm lateral shifts composed by left, right, superior, and inferior sides, were approximately 1 %, 6 %, and 11 %, respectively, compared to the dose administered to the initial tumor position. The decreases in mean tumor dose for T6.5 cm were approximately 5 %, 11 %, and 15 % for the 1 cm, 2 cm, and 3 cm outward shifts, respectively. For a superficial tumor at T2.5cm, no significant decrease in average mean tumor dose was observed following lateral shifts of 1 cm. Rotational and tilting up to 15° did not result in significant difference to the tumor dose. Dose differences to the normal tissues as a result of the shifts in positioning were also minimal. Taken together, these data demonstrate that the mean dose administered to tumors at greater depths is potentially more vulnerable to deviations in positioning, and greater shift distances resulted in reduced mean tumor doses at the THOR. Moreover, these data provide an estimation of dose differences that are caused by setup error or intra-fractional motion during BNCT, and these may facilitate more accurate predictions of actual patient dose in future treatments.

Optical and thermal simulations for the design of optodes for minimally invasive optogenetics stimulation or photomodulation of deep and large cortical areas in non-human primate brain.

The use of optogenetics or photobiomodulation in non-human primate (NHP) requires the ability to noninvasively stimulate large and deep cortical brain tissues volumes. In this context, the optical and geometrical parameters of optodes are critical. Methods and general guidelines to optimize these parameters have to be defined. OBJECTIVE: We propose the design of an optode for safe and efficient optical stimulation of a large volume of NHP cortex, down to 3-5 mm depths without inserting fibers into the cortex. APPROACH: Monte Carlo simulations of optical and thermal transport have been carried out using the Geant4 application for tomographic emission (GATE) platform. Parameters such as the fiber diameter, numerical aperture, number of fibers and their geometrical arrangement have been studied. Optimal hardware parameters are proposed to obtain homogeneous fluence above the fluence threshold for opsin activation without detrimental thermal effects. MAIN RESULTS: The simulations show that a large fiber diameter and a large numerical aperture are preferable since they allow limiting power concentration and hence the resulting thermal increases at the brain surface. To obtain a volume of 200-500 mm3 of brain tissues receiving a fluence above the opsin activation threshold for optogenetics or below a phototocixity threshold for photobiomodulation, a 4 fibers configuration is proposed. The optimal distance between the fibers was found to be 4 mm. A practical implementation of the optode has been performed and the corresponding fluence and thermal maps have been simulated. SIGNIFICANCE: The present study defines a method to optimize the design of optode and the choice of stimulation parameters for optogenetics and more generally light delivery to deep and large volumes of tissues in NHP brain with a controlled irradiance dosimetry. The general guidelines are the use of silica fibers with a large numerical aperture and a large diameter. The combination of several fibers is required if large volumes need to be stimulated while avoiding thermal effects.

Adapted estimate of neural activity based on blood-oxygen-level dependent signal by a model-free spatio-temporal clustering analysis.

In this study, we detected brain activity by comparing the overall temporal response of the blood oxygen level referring to hemodynamic response with a modeled hemodynamic response (MHR). However, in a conventional analysis by statistical parametric mapping (SPM) method, the MHR is assumed to be a fixed-response function, which may bias the conclusions about brain activation, such as the shapes of the response curve or the different response delays to stimuli. Therefore, to improve detection efficacy, we applied a spatio-temporal clustering analysis (sTCA) to determine the MHR, which is calculated from the prospective voxels with no a priori information about the experiment design. With the sTCA method, these prospective voxels are detected by the feature with the largest temporal clustering within which these voxels react simultaneously, irrespective of where the variant hemodynamic response occurs. This estimated MHR (eMHR) is then applied to search for brain activation. Preliminary results show that the eMHR signal response closely resembles the real signal response of the target area. Moreover, the activation detection using eMHR method is more sensitive for the human visual and motor tasks than that with the canonical hemodynamic response embedded in the SPM analysis as the default MHR (dMHR). The more precise location of brain activation made possible by the improved sensitivity should provide helpful information about the stimulation of neuron activity.

Preliminary dosimetric study on feasibility of multi-beam boron neutron capture therapy in patients with diffuse intrinsic pontine glioma without craniotomy.

Diffuse intrinsic pontine glioma is a very frustrating disease. Since the tumor infiltrates the brain stem, surgical removal is often impossible. For conventional radiotherapy, the dose constraint of the brain stem impedes attempts at further dose escalation. Boron neutron capture therapy (BNCT), a targeted radiotherapy, carries the potential to selectively irradiate tumors with an adequate dose while sparing adjacent normal tissue. In this study, 12 consecutive patients treated with conventional radiotherapy in our institute were reviewed to evaluate the feasibility of BNCT. NCTPlan Ver. 1.1.44 was used for dose calculations. Compared with two and three fields, the average maximal dose to the normal brain may be lowered to 7.35 ± 0.72 Gy-Eq by four-field irradiation. The mean ratio of minimal dose to clinical target volume and maximal dose to normal tissue was 2.41 ± 0.26 by four-field irradiation. A therapeutic benefit may be expected with multi-field boron neutron capture therapy to treat diffuse intrinsic pontine glioma without craniotomy, while the maximal dose to the normal brain would be minimized by using the four-field setting.

A Novel Two-Compartment Model for Calculating Bone Volume Fractions and Bone Mineral Densities From Computed Tomography Images.

Osteoporosis is a disease characterized by a degradation of bone structures. Various methods have been developed to diagnose osteoporosis by measuring bone mineral density (BMD) of patients. However, BMDs from these methods were not equivalent and were incomparable. In addition, partial volume effect introduces errors in estimating bone volume from computed tomography (CT) images using image segmentation. In this study, a two-compartment model (TCM) was proposed to calculate bone volume fraction (BV/TV) and BMD from CT images. The TCM considers bones to be composed of two sub-materials. Various equivalent BV/TV and BMD can be calculated by applying corresponding sub-material pairs in the TCM. In contrast to image segmentation, the TCM prevented the influence of the partial volume effect by calculating the volume percentage of sub-material in each image voxel. Validations of the TCM were performed using bone-equivalent uniform phantoms, a 3D-printed trabecular-structural phantom, a temporal bone flap, and abdominal CT images. By using the TCM, the calculated BV/TVs of the uniform phantoms were within percent errors of ±2%; the percent errors of the structural volumes with various CT slice thickness were below 9%; the volume of the temporal bone flap was close to that from micro-CT images with a percent error of 4.1%. No significant difference (p >0.01) was found between the areal BMD of lumbar vertebrae calculated using the TCM and measured using dual-energy X-ray absorptiometry. In conclusion, the proposed TCM could be applied to diagnose osteoporosis, while providing a basis for comparing various measurement methods.

Noninvasive measurement of radiopharmaceutical time-activity data using external thermoluminescent dosimeters (TLDs).

In this study, we present a new method for estimating the time-activity data using serial timely measurements of thermoluminescent dosimeters (TLDs). The approach is based on the combination of the measurement of surface dose using TLD and Monte Carlo (MC) simulation to estimate the radiopharmaceutical time-activity data. It involves four steps: (1) identify the source organs and outline their contours in computed tomography images; (2) compute the S values on the body surface for each source organ using a MC code; (3) obtain a serial measurement of the dose with numerous TLDs placed on the body surface; (4) solve the dose-activity equation to generate organ cumulative activity for each period of measurement. The activity of each organ at the time of measurement is simply the cumulative activity divided by the timespan between measurements. The usefulness of this method was studied using a MC simulation based on an Oak Ridge National Laboratory mathematical phantom with 18F-FDG filled in six source organs. Numerous TLDs were placed on different locations of the surface and were repeatedly read and replaced. The time-activity curves (TACs) of all organs were successfully reconstructed. Experiments on a physical phantom were also performed. Preliminary results indicate that it is an effective, robust, and simple method for assessing the TAC. The proposed method holds great potential for a range of applications in areas such as targeted radionuclide therapy, pharmaceutical research, and patient-specific dose estimation.

Optimal technique of linear accelerator-based stereotactic radiosurgery for tumors adjacent to brainstem.

Stereotactic radiosurgery (SRS) is a well-established technique that is replacing whole-brain irradiation in the treatment of intracranial lesions, which leads to better preservation of brain functions, and therefore a better quality of life for the patient. There are several available forms of linear accelerator (LINAC)-based SRS, and the goal of the present study is to identify which of these techniques is best (as evaluated by dosimetric outcomes statistically) when the target is located adjacent to brainstem. We collected the records of 17 patients with lesions close to the brainstem who had previously been treated with single-fraction radiosurgery. In all, 5 different lesion catalogs were collected, and the patients were divided into 2 distance groups-1 consisting of 7 patients with a target-to-brainstem distance of less than 0.5cm, and the other of 10 patients with a target-to-brainstem distance of ≥ 0.5 and < 1cm. Comparison was then made among the following 3 types of LINAC-based radiosurgery: dynamic conformal arcs (DCA), intensity-modulated radiosurgery (IMRS), and volumetric modulated arc radiotherapy (VMAT). All techniques included multiple noncoplanar beams or arcs with or without intensity-modulated delivery. The volume of gross tumor volume (GTV) ranged from 0.2cm(3) to 21.9cm(3). Regarding the dose homogeneity index (HIICRU) and conformity index (CIICRU) were without significant difference between techniques statistically. However, the average CIICRU = 1.09 ± 0.56 achieved by VMAT was the best of the 3 techniques. Moreover, notable improvement in gradient index (GI) was observed when VMAT was used (0.74 ± 0.13), and this result was significantly better than those achieved by the 2 other techniques (p < 0.05). For V4Gy of brainstem, both VMAT (2.5%) and IMRS (2.7%) were significantly lower than DCA (4.9%), both at the p < 0.05 level. Regarding V2Gy of normal brain, VMAT plans had attained 6.4 ± 5%; this was significantly better (p < 0.05) than either DCA or IMRS plans, at 9.2 ± 7% and 8.2 ± 6%, respectively. Owing to the multiple arc or beam planning designs of IMRS and VMAT, both of these techniques required higher MU delivery than DCA, with the averages being twice as high (p < 0.05). If linear accelerator is only 1 modality can to establish for SRS treatment. Based on statistical evidence retrospectively, we recommend VMAT as the optimal technique for delivering treatment to tumors adjacent to brainstem.

Fuzzy C-means clustering of magnetic resonance imaging on apparent diffusion coefficient maps for predicting nodal metastasis in head and neck cancer.

OBJECTIVE: The present study evaluated and analyzed apparent diffusion coefficients (ADCs) from partitions through a fuzzy C-means (FCM) technique for distinguishing nodal metastasis in head and neck cancer. METHODS: MRI studies of 169 lymph node lesions, dissected from 22 patients with a histopathologically confirmed lymph node status, were analyzed using in-house software developed using MATLAB(®) (The MathWorks(®) Inc., Natick, MA). A radiologist manually contoured the lesions, and ADCs for each lesion were divided into two (low and high) and three (low, intermediate and high) partitions by using the FCM clustering algorithm. RESULTS: The results showed that the low-value ADC clusters were more sensitive (95.7%) in distinguishing malignant from benign lesions than the whole-lesion mean ADC values (78.3%), while retaining a high specificity (approximately 90%). Moreover, receiver-operating characteristic curves demonstrated that the low-value ADC clusters used as a predictor of malignancy for lymph nodes could achieve a higher area under the curve (0.949 and 0.944 for two and three partitions, respectively). CONCLUSION: The segmentation by ADC values of lesions through the FCM technique enables the efficient characterization of the lymph node pathology and can help distinguish malignant from benign lymph nodes. ADVANCES IN KNOWLEDGE: Tumour heterogeneity may degrade the prediction of metastatic lymph nodes that involves using mean region-of-interest ADC values. The clustering of ADC values in lesions by using FCM can improve the diagnostic accuracy of nodal metastasis and reduce interreader variance.

Recovering the triple coincidence of non-pure positron emitters in preclinical PET.

Non-pure positron emitters, with their long half-lives, allow for the tracing of slow biochemical processes which cannot be adequately examined by the commonly used short-lived positron emitters. Most of these isotopes emit high-energy cascade gamma rays in addition to positron decay that can be detected and create a triple coincidence with annihilation photons. Triple coincidence is discarded in most scanners, however, the majority of the triple coincidence contains true photon pairs that can be recovered. In this study, we propose a strategy for recovering triple coincidence events to raise the sensitivity of PET imaging for non-pure positron emitters. To identify the true line of response (LOR) from a triple coincidence, a framework utilizing geometrical, energy and temporal information is proposed. The geometrical criterion is based on the assumption that the LOR with the largest radial offset among the three sub pairs of triple coincidences is least likely to be a true LOR. Then, a confidence time window is used to test the valid LOR among those within triple coincidence. Finally, a likelihood ratio discriminant rule based on the energy probability density distribution of cascade and annihilation gammas is established to identify the true LOR. An Inveon preclinical PET scanner was modeled with GATE (GEANT4 application for tomographic emission) Monte Carlo software. We evaluated the performance of the proposed method in terms of identification fraction, noise equivalent count rates (NECR), and image quality on various phantoms. With the inclusion of triple coincidence events using the proposed method, the NECR was found to increase from 11% to 26% and 19% to 29% for I-124 and Br-76, respectively, when 7.4-185 MBq of activity was used. Compared to the reconstructed images using double coincidence, this technique increased the SNR by 5.1-7.3% for I-124 and 9.3-10.3% for Br-76 within the activity range of 9.25-74 MBq, without compromising the spatial resolution or contrast. We conclude that the proposed method can improve the counting statistics of PET imaging for non-pure positron emitters and is ready to be implemented on current PET systems.

Image reconstruction of optical computed tomography by using the algebraic reconstruction technique for dose readouts of polymer gel dosimeters.

Optical computed tomography (optical CT) has been proven to be a useful tool for dose readouts of polymer gel dosimeters. In this study, the algebraic reconstruction technique (ART) for image reconstruction of gel dosimeters was used to improve the image quality of optical CT. Cylindrical phantoms filled with N-isopropyl-acrylamide polymer gels were irradiated using a medical linear accelerator. A circular dose distribution and a hexagonal dose distribution were produced by applying the VMAT technique and the six-field dose delivery, respectively. The phantoms were scanned using optical CT, and the images were reconstructed using the filtered back-projection (FBP) algorithm and the ART. For the circular dose distribution, the ART successfully reduced the ring artifacts and noise in the reconstructed image. For the hexagonal dose distribution, the ART reduced the hot spots at the entrances of the beams and increased the dose uniformity in the central region. Within 50% isodose line, the gamma pass rates for the 2 mm/3% criteria for the ART and FBP were 99.2% and 88.1%, respectively. The ART could be used for the reconstruction of optical CT images to improve image quality and provide accurate dose conversion for polymer gel dosimeters.

Efficient simulation of voxelized phantom in GATE with embedded SimSET multiple photon history generator.

GEANT4 Application for Tomographic Emission (GATE) is a powerful Monte Carlo simulator that combines the advantages of the general-purpose GEANT4 simulation code and the specific software tool implementations dedicated to emission tomography. However, the detailed physical modelling of GEANT4 is highly computationally demanding, especially when tracking particles through voxelized phantoms. To circumvent the relatively slow simulation of voxelized phantoms in GATE, another efficient Monte Carlo code can be used to simulate photon interactions and transport inside a voxelized phantom. The simulation system for emission tomography (SimSET), a dedicated Monte Carlo code for PET/SPECT systems, is well-known for its efficiency in simulation of voxel-based objects. An efficient Monte Carlo workflow integrating GATE and SimSET for simulating pinhole SPECT has been proposed to improve voxelized phantom simulation. Although the workflow achieves a desirable increase in speed, it sacrifices the ability to simulate decaying radioactive sources such as non-pure positron emitters or multiple emission isotopes with complex decay schemes and lacks the modelling of time-dependent processes due to the inherent limitations of the SimSET photon history generator (PHG). Moreover, a large volume of disk storage is needed to store the huge temporal photon history file produced by SimSET that must be transported to GATE. In this work, we developed a multiple photon emission history generator (MPHG) based on SimSET/PHG to support a majority of the medically important positron emitters. We incorporated the new generator codes inside GATE to improve the simulation efficiency of voxelized phantoms in GATE, while eliminating the need for the temporal photon history file. The validation of this new code based on a MicroPET R4 system was conducted for (124)I and (18)F with mouse-like and rat-like phantoms. Comparison of GATE/MPHG with GATE/GEANT4 indicated there is a slight difference in energy spectra for energy below 50 keV due to the lack of x-ray simulation from (124)I decay in the new code. The spatial resolution, scatter fraction and count rate performance are in good agreement between the two codes. For the case studies of (18)F-NaF ((124)I-IAZG) using MOBY phantom with 1  ×  1 × 1 mm(3) voxel sizes, the results show that GATE/MPHG can achieve acceleration factors of approximately 3.1 × (4.5 ×), 6.5 × (10.7 ×) and 9.5 × (31.0 ×) compared with GATE using the regular navigation method, the compressed voxel method and the parameterized tracking technique, respectively. In conclusion, the implementation of MPHG in GATE allows for improved efficiency of voxelized phantom simulations and is suitable for studying clinical and preclinical imaging.

A novel adaptive discrete cosine transform-domain filter for gap-inpainting of high resolution PET scanners.

PURPOSE: Several positron emission tomography (PET) scanners with special detector block arrangements have been developed in recent years to improve the resolution of PET images. However, the discontinuous detector blocks cause gaps in the sinogram. This study proposes an adaptive discrete cosine transform-based (aDCT) filter for gap-inpainting. METHODS: The gap-corrupted sinogram was morphologically closed and subsequently converted to the DCT domain. A certain number of the largest coefficients in the DCT spectrum were identified to determine the low-frequency preservation region. The weighting factors for the remaining coefficients were determined by an exponential weighting function. The aDCT filter was constructed and applied to two digital phantoms and a simulated phantom introduced with various levels of noise. RESULTS: For the Shepp-Logan head phantom, the aDCT filter filled the gaps effectively. For the Jaszczak phantom, no secondary artifacts were induced after aDCT filtering. The percent mean square error and mean structure similarity of the aDCT filter were superior to those of the DCT2 filter at all noise levels. For the simulated striatal dopamine innervation study, the aDCT filter recovered the shape of the striatum and restored the striatum to reference activity ratios to the ideal value. CONCLUSIONS: The proposed aDCT filter can recover the missing gap data in the sinogram and improve the image quality and quantitative accuracy of PET images.

A simple algorithm for subregional striatal uptake analysis with partial volume correction in dopaminergic PET imaging.

OBJECTIVE: In positron emission tomography (PET) of the dopaminergic system, quantitative measurements of nigrostriatal dopamine function are useful for differential diagnosis. A subregional analysis of striatal uptake enables the diagnostic performance to be more powerful. However, the partial volume effect (PVE) induces an underestimation of the true radioactivity concentration in small structures. This work proposes a simple algorithm for subregional analysis of striatal uptake with partial volume correction (PVC) in dopaminergic PET imaging. METHODS: The PVC algorithm analyzes the separate striatal subregions and takes into account the PVE based on the recovery coefficient (RC). The RC is defined as the ratio of the PVE-uncorrected to PVE-corrected radioactivity concentration, and is derived from a combination of the traditional volume of interest (VOI) analysis and the large VOI technique. The clinical studies, comprising 11 patients with Parkinson's disease (PD) and 6 healthy subjects, were used to assess the impact of PVC on the quantitative measurements. Simulations on a numerical phantom that mimicked realistic healthy and neurodegenerative situations were used to evaluate the performance of the proposed PVC algorithm. In both the clinical and the simulation studies, the striatal-to-occipital ratio (SOR) values for the entire striatum and its subregions were calculated with and without PVC. RESULTS: In the clinical studies, the SOR values in each structure (caudate, anterior putamen, posterior putamen, putamen, and striatum) were significantly higher by using PVC in contrast to those without. Among the PD patients, the SOR values in each structure and quantitative disease severity ratings were shown to be significantly related only when PVC was used. For the simulation studies, the average absolute percentage error of the SOR estimates before and after PVC were 22.74% and 1.54% in the healthy situation, respectively; those in the neurodegenerative situation were 20.69% and 2.51%, respectively. CONCLUSIONS: We successfully implemented a simple algorithm for subregional analysis of striatal uptake with PVC in dopaminergic PET imaging. The PVC algorithm provides an accurate measure of the SOR in the entire striatum and its subregions, and improves the correlation between the SOR values and the clinical disease severity of PD patients.