Integration of advanced 3D SPECT modeling into the open-source STIR framework

Deep learning approximation of attenuation maps for myocardial perfusion SPECT with an IQ[Formula: see text]SPECT collimator

BACKGROUND

The use of CT images for attenuation correction of myocardial perfusion imaging with single photon emission computer tomography (SPECT) increases diagnostic confidence. However, acquiring a CT image registered to a SPECT image is often not possible because most scanners are SPECT-only. It is possible to approximate attenuation maps using deep learning, but this has not yet been shown to work for a SPECT scanner with an IQ[Formula: see text]SPECT collimator. This study investigates whether it is possible to approximate attenuation maps from non-attenuation-corrected (nAC) reconstructions acquired with a SPECT scanner equipped with an IQ[Formula: see text]SPECT collimator.

METHODS

Attenuation maps and reconstructions were acquired retrospectively for 150 studies. A U-Net was trained to predict attenuation maps from nAC reconstructions using the conditional generative adversarial network framework. Predicted attenuation maps are compared to real attenuation maps using the normalized mean absolute error (NMAE). Attenuation-corrected reconstructions were computed, and the resulting polar maps were compared by pixel and by average perfusion per segment using the absolute percent error (APE). The training and evaluation code is available at https://gitlab.ub.uni-bielefeld.de/thuxohl/mu-map .

RESULTS

Predicted attenuation maps are similar to real attenuation maps, achieving an NMAE of 0.020±0.007. The same is true for polar maps generated from reconstructions with predicted attenuation maps compared to CT-based attenuation maps. Their pixel-wise absolute distance is 3.095±3.199, and the segment-wise APE is 1.155±0.769.

CONCLUSIONS

It is feasible to approximate attenuation maps from nAC reconstructions acquired by a scanner with an IQ[Formula: see text]SPECT collimator using deep learning.

Simulation studies of a full-ring, CZT SPECT system for whole-body imaging of 99m Tc and 177 Lu

BACKGROUND

Single photon emission computed tomography (SPECT) is an imaging modality that has demonstrated its utility in a number of clinical indications. Despite this progress, a high sensitivity, high spatial resolution, multi-tracer SPECT with a large field of view suitable for whole-body imaging of a broad range of radiotracers for theranostics is not available.

PURPOSE

With the goal of filling this technological gap, we have designed a cadmium zinc telluride (CZT) full-ring SPECT scanner instrumented with a broad-energy tungsten collimator. The final purpose is to provide a multi-tracer solution for brain and whole-body imaging. Our static SPECT does not rely on the dual- and the triple-head rotational SPECT standard paradigm, enabling a larger effective area in each scan to increase the sensitivity. We provide a demonstration of the performance of our design using a realistic model of our detector with simulated body-sized phantoms filled with 99m Tc and 177 Lu.

METHODS

We create a realistic model of our detector by using a combination of a Geant4 Application for Tomographic Emission (GATE) Monte Carlo simulation and a finite element model for the CZT response, accounting for low-energy tail effects in CZT that affects the sensitivity and the scatter correction. We implement a modified dual-energy-window scatter correction adapted for CZT. Other corrections for attenuation, detector and collimator response, and detector gaps and edges are also included. The images are reconstructed using the maximum-likelihood expectation-maximization. Detector and reconstruction performance are characterized with point sources, Derenzo phantoms, and a body-sized National Electrical Manufacturers Association (NEMA) Image Quality (IQ) phantom for both 99m Tc and 177 Lu.

RESULTS

Our SPECT design can resolve 7.9 mm rods for 99m Tc (140 keV) and 9.5 mm for 177 Lu (208 keV) in a hot-rod Derenzo phantom with a 3-min exposure and reach an image contrast of 78% for 99m Tc and 57% for 177 Lu using the NEMA IQ phantom with a 6-min exposure. Our modified scatter correction shows an improved contrast-recovery ratio compared to a standard correction.

CONCLUSIONS

In this paper, we demonstrate the good performance of our design for whole-body imaging purposes. This adds to our previous demonstration of improved qualitative and quantitative 99m Tc imaging of brain perfusion and 123 I imaging of dopamine transport with respect to state-of-the-art NaI dual-head cameras. We show that our design provides similar IQ and contrast to the commercial full-ring SPECT VERITON for 99m Tc. Regarding 177 Lu imaging of the 208 keV emissions, our design provides similar contrast to that of other state-of-the-art SPECTs with a significant reduction in exposure. The high sensitivity and extended energy range up to 250 keV makes our SPECT design a promising alternative for clinical imaging and theranostics of emerging radionuclides.

NEMA NU 1-2018 performance characterization and Monte Carlo model validation of the Cubresa Spark SiPM-based preclinical SPECT scanner

BACKGROUND

The Cubresa Spark is a novel benchtop silicon-photomultiplier (SiPM)-based preclinical SPECT system. SiPMs in SPECT significantly improve resolution and reduce detector size compared to preclinical cameras with photomultiplier tubes requiring highly magnifying collimators. The NEMA NU 1 Standard for Performance Measurements of Gamma Cameras provides methods that can be readily applied or extended to characterize preclinical cameras with minor modifications. The primary objective of this study is to characterize the Spark according to the NEMA NU 1-2018 standard to gain insight into its nuclear medicine imaging capabilities. The secondary objective is to validate a GATE Monte Carlo simulation model of the Spark for use in preclinical SPECT studies.

METHODS

NEMA NU 1-2018 guidelines were applied to characterize the Spark's intrinsic, system, and tomographic performance with single- and multi-pinhole collimators. Phantoms were fabricated according to NEMA specifications with deviations involving high-resolution modifications. GATE was utilized to model the detector head with the single-pinhole collimator, and NEMA measurements were employed to tune and validate the model. Single-pinhole and multi-pinhole SPECT data were reconstructed with the Software for Tomographic Image Reconstruction and HiSPECT, respectively.

RESULTS

The limiting intrinsic resolution was measured as 0.85 mm owing to a high-resolution SiPM array combined with a 3 mm-thick scintillation crystal. The average limiting tomographic resolution was 1.37 mm and 1.19 mm for the single- and multi-pinhole collimators, respectively, which have magnification factors near unity at the center of rotation. The maximum observed count rate was 15,400 cps, and planar sensitivities of 34 cps/MBq and 150 cps/MBq were measured at the center of rotation for the single- and multi-pinhole collimators, respectively. All simulated tests agreed well with measurement, where the most considerable deviations were below 7%.

CONCLUSIONS

NEMA NU 1-2018 standards determined that a SiPM detector mitigates the need for highly magnifying pinhole collimators while preserving detailed information in projection images. Measured and simulated NEMA results were highly comparable with differences on the order of a few percent, confirming simulation accuracy and validating the GATE model. Of the collimators initially provided with the Spark, the multi-pinhole collimator offers high resolution and sensitivity for organ-specific imaging of small animals, and the single-pinhole collimator enables high-resolution whole-body imaging of small animals.

Towards accurate partial volume correction in 99mTc oncology SPECT: perturbation for case-specific resolution estimation

BACKGROUND

Currently, there is no consensus on the optimal partial volume correction (PVC) algorithm for oncology imaging. Several existing PVC methods require knowledge of the reconstructed resolution, usually as the point spread function (PSF)-often assumed to be spatially invariant. However, this is not the case for SPECT imaging. This work aimed to assess the accuracy of SPECT quantification when PVC is applied using a case-specific PSF.

METHODS

Simulations of SPECT [Formula: see text]Tc imaging were performed for a range of activity distributions, including those replicating typical clinical oncology studies. Gaussian PSFs in reconstructed images were estimated using perturbation with a small point source. Estimates of the PSF were made in situations which could be encountered in a patient study, including; different positions in the field of view, different lesion shapes, sizes and contrasts, noise-free and noisy data. Ground truth images were convolved with the perturbation-estimated PSF, and with a PSF reflecting the resolution at the centre of the field of view. Both were compared with reconstructed images and the root-mean-square error calculated to assess the accuracy of the estimated PSF. PVC was applied using Single Target Correction, incorporating the perturbation-estimated PSF. Corrected regional mean values were assessed for quantitative accuracy.

RESULTS

Perturbation-estimated PSF values demonstrated dependence on the position in the Field of View and the number of OSEM iterations. A lower root mean squared error was observed when convolution of the ground truth image was performed with the perturbation-estimated PSF, compared with convolution using a different PSF. Regional mean values following PVC using the perturbation-estimated PSF were more accurate than uncorrected data, or data corrected with PVC using an unsuitable PSF. This was the case for both simple and anthropomorphic phantoms. For the simple phantom, regional mean values were within 0.7% of the ground truth values. Accuracy improved after 5 or more OSEM iterations (10 subsets). For the anthropomorphic phantoms, post-correction regional mean values were within 1.6% of the ground truth values for noise-free uniform lesions.

CONCLUSION

Perturbation using a simulated point source could potentially improve quantitative SPECT accuracy via the application of PVC, provided that sufficient reconstruction iterations are used.

Evaluation of a variable-aperture full-ring SPECT system using large-area pixelated CZT modules: A simulation study for brain SPECT applications

PURPOSE

Single photon emission computed tomography (SPECT) scanners using cadmium zinc telluride (CZT) offer compact, lightweight, and improved imaging capability over conventional NaI(Tl)-based SPECT scanners. The main purpose in this study is to propose a full-ring SPECT system design with eight large-area CZT detectors that can be used for a broad spectrum of SPECT radiopharmaceuticals and demonstrate the performance of our system in comparison to the reference conventional NaI(Tl)-based two-head Anger cameras.

METHODS

A newly designed full-ring SPECT system is composed of eight large-area CZT cameras (128 mm × 179.2 mm effective area) that can be independently swiveled around their own axes of rotation independently and can have radial motion for varying aperture sizes that can be adapted to different sizes of imaging volume. Extended projection data were generated by conjoining projections of two adjacent detectors to overcome the limited field-of-view (FOV) by each CZT camera. Using Monte Carlo simulations, we evaluated this new system design with digital phantoms including a Derenzo hot rod phantom and a Zubal brain phantom. Comparison of performance metrics such as spatial resolution, sensitivity, contrast-to-noise ratio (CNR), and contrast-recovery ratio was made between our design and conventional SPECT scanners having different pixel sizes and radii of rotation (one clinically well-known type and two arbitrary types matched to our proposed CZT-SPECT geometries).

RESULTS

The proposed scanner could result in up to about three times faster in acquisition time over conventional scan time at same acquisition time per step. The spatial resolution improvement, or deterioration, of our proposed scanner compared to the clinical-type scanner was dependent upon the location of the point source. However, there were overall performance improvements over the three different setups of the conventional scanner particularly in volume sensitivity (approximately up to 1.7 times). Overall, we successfully reconstructed the phantom image for both ^99m Tc-based perfusion and ¹²³ I-based dopamine transporter (DaT) brain studies simulated for our new design. In particular, the striatal/background contrast-recovery ratio in 3-to-1 reference ratio was over 0.8 for the ¹²³ I-based DaT study.

CONCLUSIONS

We proposed a variable-aperture full-ring SPECT system using combined pixelated CZT and energy-optimized parallel-hole collimator modules and evaluated the performance of this scanner using relevant digital phantoms and MC simulations. Our studies demonstrated the potential of our new full-ring CZT-SPECT design, showing reduced acquisition time and improved sensitivity with acceptable CNR and spatial resolution.

Optimization of the reconstruction parameters in [123I]FP-CIT SPECT

The aim of this work was to obtain a set of parameters to be applied in [¹²³I]FP-CIT SPECT reconstruction in order to minimize the error between standardized and true values of the specific uptake ratio (SUR) in dopaminergic neurotransmission SPECT studies. To this end, Monte Carlo simulation was used to generate a database of 1380 projection data-sets from 23 subjects, including normal cases and a variety of pathologies. Studies were reconstructed using filtered back projection (FBP) with attenuation correction and ordered subset expectation maximization (OSEM) with correction for different degradations (attenuation, scatter and PSF). Reconstruction parameters to be optimized were the cut-off frequency of a 2D Butterworth pre-filter in FBP, and the number of iterations and the full width at Half maximum of a 3D Gaussian post-filter in OSEM. Reconstructed images were quantified using regions of interest (ROIs) derived from Magnetic Resonance scans and from the Automated Anatomical Labeling map. Results were standardized by applying a simple linear regression line obtained from the entire patient dataset. Our findings show that we can obtain a set of optimal parameters for each reconstruction strategy. The accuracy of the standardized SUR increases when the reconstruction method includes more corrections. The use of generic ROIs instead of subject-specific ROIs adds significant inaccuracies. Thus, after reconstruction with OSEM and correction for all degradations, subject-specific ROIs led to errors between standardized and true SUR values in the range [-0.5, +0.5] in 87% and 92% of the cases for caudate and putamen, respectively. These percentages dropped to 75% and 88% when the generic ROIs were used.