Performance evaluation of a novel multi-pinhole collimator for dopamine transporter SPECT

Automated identification of uncertain cases in deep learning-based classification of dopamine transporter SPECT to improve clinical utility and acceptance

PURPOSE

Deep convolutional neural networks (CNN) are promising for automatic classification of dopamine transporter (DAT)-SPECT images. Reporting the certainty of CNN-based decisions is highly desired to flag cases that might be misclassified and, therefore, require particularly careful inspection by the user. The aim of the current study was to design and validate a CNN-based system for the identification of uncertain cases.

METHODS

A network ensemble (NE) combining five CNNs was trained for binary classification of [123I]FP-CIT DAT-SPECT images as "normal" or "neurodegeneration-typical reduction" with high accuracy (NE for classification, NEfC). An uncertainty detection module (UDM) was obtained by combining two additional NE, one trained for detection of "reduced" DAT-SPECT with high sensitivity, the other with high specificity. A case was considered "uncertain" if the "high sensitivity" NE and the "high specificity" NE disagreed. An internal "development" dataset of 1740 clinical DAT-SPECT images was used for training (n = 1250) and testing (n = 490). Two independent datasets with different image characteristics were used for testing only (n = 640, 645). Three established approaches for uncertainty detection were used for comparison (sigmoid, dropout, model averaging).

RESULTS

In the test data from the development dataset, the NEfC achieved 98.0% accuracy. 4.3% of all test cases were flagged as "uncertain" by the UDM: 2.5% of the correctly classified cases and 90% of the misclassified cases. NEfC accuracy among "certain" cases was 99.8%. The three comparison methods were less effective in labelling misclassified cases as "uncertain" (40-80%). These findings were confirmed in both additional test datasets.

CONCLUSION

The UDM allows reliable identification of uncertain [123I]FP-CIT SPECT with high risk of misclassification. We recommend that automatic classification of [123I]FP-CIT SPECT images is combined with an UDM to improve clinical utility and acceptance. The proposed UDM method ("high sensitivity versus high specificity") might be useful also for DAT imaging with other ligands and for other binary classification tasks.

Dynamic multi-pinhole collimated brain SPECT of Parkinson's disease by [123I]FP-CIT: a feasibility study of fSPECT

We investigated the feasibility of using a dopamine transporter (DaT) tracer ligand ([123I]FP-CIT) along with novel multi-pinhole brain collimators for dynamic brain single photon emission computed tomography (SPECT) in suspected Parkinson's disease patients. Thirteen patients underwent dynamic tracer acquisitions before standard imaging. Uptake values were corrected for partial volume effects. Specific binding ratio (SBRcalc) was calculated, reflecting binding potential relative to non-displaceable binding (BPND) in the cortex. Additional pharmacokinetic parameters (BPND, R1, k2) were estimated using the simplified reference tissue model, revealing differences between Kahraman low-score (LS) and high-score (HS) groups. Results showed increasing striatal tracer uptake until 100 min post-injection, with consistent values afterward. Uptake and SBRcalc ratios matched visual assessment. LS patients had lower putamen than caudate nucleus tracer uptake, decreased BPND values, while R1 and k2 values were comparable to HS patients. In conclusion, dynamic multi-pinhole SPECT using DaT tracer with the extraction of pharmacokinetic parameters is feasible and could help enable early differentiation of reduced and normal DaT values.

Dopamine Transporter SPECT with 12-Minute Scan Duration Using Multiple-Pinhole Collimators

This study evaluated the potential to reduce the scan duration in dopamine transporter (DAT) SPECT when using a second-generation multiple-pinhole (MPH) collimator designed for brain SPECT with improved count sensitivity and improved spatial resolution compared with parallel-hole and fanbeam collimators. Methods: The retrospective study included 640 consecutive clinical DAT SPECT studies that had been acquired in list mode with a triple-head SPECT system with MPH collimators and a 30-min net scan duration after injection of 181 ± 10 MBq of [123I]FP-CIT. Raw data corresponding to scan durations of 20, 15, 12, 8, 6, and 4 min were obtained by restricting the events to a proportionally reduced time interval of the list-mode data for each projection angle. SPECT images were reconstructed iteratively with the same parameter settings irrespective of scan duration. The resulting 5,120 SPECT images were assessed for a neurodegeneration-typical reduction in striatal signal by visual assessment, conventional specific binding ratio analysis, and a deep convolutional neural network trained on 30-min scans. Results: Regarding visual interpretation, image quality was considered diagnostic for all 640 patients down to a 12-min scan duration. The proportion of discrepant visual interpretations between 30 and 12 min (1.2%) was not larger than the proportion of discrepant visual interpretations between 2 reading sessions of the same reader at a 30-min scan duration (1.5%). Agreement with the putamen specific binding ratio from the 30-min images was better than expected for 5% test-retest variability down to a 10-min scan duration. A relevant change in convolutional neural network-based automatic classification was observed at a 6-min scan duration or less. Conclusion: The triple-head SPECT system with MPH collimators allows reliable DAT SPECT after administration of about 180 MBq of [123I]FP-CIT with a 12-min scan duration.

Multi-pinhole collimator design in different numbers of projections for brain SPECT

Purpose

High resolution and sensitivity brain SPECT is promising for the accurate diagnosis of Alzheimer's disease (AD) and Parkinson's disease (PD). Multi-pinhole (MPH) collimators with a good performance in imaging small field-of-view (FOV) could be better used for brain SPECT. In this study, we aim to evaluate the impact of varying the number of pinholes and the number of projections on the performance of MPH brain SPECT.

Methods

The system design was based on a commercial clinical dual-head SPECT/CT scanner, with target spatial resolutions of 12 mm and 8 mm for AD and PD SPECT, respectively. In total, 1-25 pinholes were modeled for 64, 32, 16, 8, 4, and 2 projections. The 3D NURBS-based HUman Brain (NHUB) phantom was used in the analytical simulation to model 99mTc-HMPAO and 99mTc-TRODAT distributions. The 2D Derenzo hot-rod phantom and star phantom were used in Monte Carlo simulations to evaluate the spatial resolution and angular sampling performance of MPH. The influence of different detector positions was also evaluated for 2, 4, and 6 angular views. The projections were reconstructed using the 3D MPH ML-EM method. Normalized mean square error, coefficient of variation, and image profiles were evaluated.

Results

Along with the decrease in the number of projections, more pinholes are required to achieve the optimum performance. For 32 projections, 9- and 7-pinhole collimators provide the best normalized mean square error (NMSE) to the coefficient of variation (COV) trade-off for 99mTc-HMPAO and 99mTc-TRODAT, respectively. Detector positions substantially affect the image quality for MPH SPECT for 2 and 4 angular views. The smallest rod size for the Derenzo hot-rod phantom, which could be resolved, is 7.9 mm for the MPH general purpose collimator (MPGP) with more than 16 projections and 6.4 mm for MPH high-resolution collimator (MPHR) with more than 8 projections.

Conclusion

The number of pinholes affects the performance of the MPH collimator, especially when the projection views become fewer. More pinholes are required for fewer projections to provide better angular sampling in MPH for complex activity distributions. Detector positions affect the image quality of MPH SPECT for 2 and 4 angular views, where L-mode acquisition is slightly superior to H-mode. MPH collimators exhibited improved spatial resolution and angular sampling compared with both LEHR and single pinhole collimators.

Performance evaluation of a novel multi-pinhole collimator on triple-NaI-detector SPECT/CT for dedicated myocardial imaging

BACKGROUND

In this study we evaluated the imaging capabilities of a novel Multi-pinhole collimator (MPH-Cardiac) specially designed for nuclear cardiology imaging on a Triple-NaI-detector based SPECT/CT system.

METHODS

^99mTc point source measurements covering the field of view (FOV) were used to determine tomographic sensitivity (TS_pointsource) and spatial resolution. Organ-size tomographic sensitivity (TS_organ) was measured with a left ventricle (LV) phantom filled with typical myocardial activity of a patient scan. Reconstructed image uniformity was measured with a 140 mm diameter uniform cylinder phantom. Using the LV phantom once filled with ^99mTc and after with ¹²³I, Contrast-to-noise ratio (CNR) was measured on the reconstructed images by ROI analysis on the myocardium activity and on the LV cavity. Furthermore, a polar map analysis was performed determining Spill-Over-Ratio in water (SOR_water) and image noise. The results were compared with that of a dual-head parallel-hole low energy high resolution (LEHR) collimator system. A patient with suspected coronary artery disease (CAD) was scanned on the LEHR system using local protocol of 16 min total acquisition time, followed by a 4-min MPH-Cardiac scan.

RESULTS

Peak TS_pointsource was found to be 1013 cps/MBq in the axial center of the FOV while it was decreasing toward the radial edges. TS_organ in the CFOV was found to be 134 cps/MBq and 700 cps/MBq for the LEHR and MPH-Cardiac, respectively. Average spatial resolution throughout the FOV was 4.38 mm FWHM for the MPH-Cardiac collimator. Reconstructed image uniformity values were found to be 0.292% versus 0.214% for the LEHR and MPH-Cardiac measurements, respectively. CNR was found to be higher in case of MPH-Cardiac than for LEHR in case of ^99mTc (15.5 vs. 11.7) as well as for ¹²³I (13.5 vs. 8.3). SOR_water values were found to be 28.83% and 21.1% for the ^99mTc measurements, and 31.44% and 24.33% for the ¹²³I measurements for LEHR and MPH-Cardiac, respectively. Pixel noise of the ^99mTc polar maps resulted in values of 0.38% and 0.24% and of the ¹²³I polar maps 0.62% and 0.21% for LEHR and MPH-Cardiac, respectively. Visually interpreting the patient scan images, MPH-Cardiac resulted in better image contrast compared to the LEHR technique with four times shorter scan duration.

CONCLUSIONS

The significant image quality improvement achieved with dedicated MPH-Cardiac collimator on triple head SPECT/CT system paves the way for short acquisition and low-dose cardiovascular SPECT applications.

Molecular Imaging in Parkinsonian Disorders—What’s New and Hot?

Highlights

Abstract

Neurodegenerative parkinsonian disorders are characterized by a great diversity of clinical symptoms and underlying neuropathology, yet differential diagnosis during lifetime remains probabilistic. Molecular imaging is a powerful method to detect pathological changes in vivo on a cellular and molecular level with high specificity. Thereby, molecular imaging enables to investigate functional changes and pathological hallmarks in neurodegenerative disorders, thus allowing to better differentiate between different forms of degenerative parkinsonism, improve the accuracy of the clinical diagnosis and disentangle the pathophysiology of disease-related symptoms. The past decade led to significant progress in the field of molecular imaging, including the development of multiple new and promising radioactive tracers for single photon emission computed tomography (SPECT) and positron emission tomography (PET) as well as novel analytical methods. Here, we review the most recent advances in molecular imaging for the diagnosis, prognosis, and mechanistic understanding of parkinsonian disorders. First, advances in imaging of neurotransmission abnormalities, metabolism, synaptic density, inflammation, and pathological protein aggregation are reviewed, highlighting our renewed understanding regarding the multiplicity of neurodegenerative processes involved in parkinsonian disorders. Consequently, we review the role of molecular imaging in the context of disease-modifying interventions to follow neurodegeneration, ensure stratification, and target engagement in clinical trials.

Multiple-pinhole collimators improve intra- and between-rater agreement and the certainty of the visual interpretation in dopamine transporter SPECT

Background

Multiple-pinhole (MPH) collimators improve the resolution–sensitivity trade-off compared to parallel-hole collimators. This study evaluated the impact of MPH collimators on intra- and between-rater agreement, and on the certainty of visual interpretation in dopamine transporter (DAT)-SPECT.

Methods

The study included 71 patients (62.1 ± 12.7 y). Two SPECT acquisitions were performed in randomized order after a single injection of 182 ± 9 MBq ¹²³I-FP-CIT, one with MPH and one with low-energy–high-resolution–high-sensitivity (LEHRHS) collimators. MPH projections were reconstructed with an iterative 3d Monte Carlo algorithm. LEHRHS projections were reconstructed with filtered backprojection (FBP) or with ordered-subsets expectation–maximization and resolution recovery (OSEM). Images were visually evaluated twice by three independent raters with respect to presence/absence of Parkinson-typical reduction of striatal ¹²³I-FP-CIT uptake using a Likert 6-score (− 3 = clearly normal, …, 3 = clearly reduced). In case of intra-rater discrepancy, an intra-rater consensus was obtained. Intra- and between-rater agreement with respect to the Likert score (6-score and dichotomized score) was characterized by Cohen’s kappa.

Results

Intra-rater kappa of visual scoring of MPH/LEHRHS-OSEM/LEHRHS-FBP images was 0.84 ± 0.12/0.73 ± 0.06/0.73 ± 0.08 (6-score, mean of three raters) and 1.00 ± 0.00/0.96 ± 0.04/0.97 ± 0.03 (dichotomized score). Between-rater kappa of visual scoring (intra-rater consensus) of MPH/LEHRHS-OSEM/LEHRHS-FBP images was 0.70 ± 0.06/0.63 ± 0.08/0.48 ± 0.05 (6-score, mean of three pairs of raters) and 1.00 ± 0.00/0.92 ± 0.04/0.90 ± 0.06 (dichotomized score). There was a decrease of (negative) Likert scores in normal DAT-SPECT by 0.87 ± 0.18 points from the LEHRHS-OSEM to the MPH setting. The (positive) Likert scores of reduced DAT-SPECT did not change on average.

Conclusions

MPH collimators improve intra- and between-rater agreement as well as the certainty of the visual interpretation of DAT-SPECT.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13550-022-00923-w.

Improved Performance of a Multipinhole SPECT for DAT Imaging by Increasing Number of Pinholes at the Expense of Increased Multiplexing

SPECT imaging of dopamine transporters (DAT) in the brain is a widely utilized study to improve the diagnosis of Parkinsonian syndromes, where conventional (parallel-hole and fan-beam) collimators on dual-head scanners are commonly employed. We have designed a multi-pinhole (MPH) collimator to improve the performance of DAT imaging. The MPH collimator focuses on the striatum and hence offers a better trade-off for sensitivity and spatial resolution than the conventional collimators within this clinically most relevant region for DAT imaging. Our original MPH design consisted of 9 pinholes with a background-to-striatal (Bkg/Str) projection multiplexing of 1% only. In this simulation study, we investigated whether further improvements in the performance of MPH imaging could be obtained by increasing the number of pinholes, hence by enhancing the sensitivity and sampling, despite the ambiguity in reconstructing images due to increased multiplexing. We performed analytic simulations of the MPH configurations with 9, 13, and 16 pinholes (aperture diameters: 4-6mm) using a digital phantom modeling DAT imaging. Our quantitative analyses indicated that using 13 (Bkg/Str: 12%) and 16 (Bkg/Str: 22%) pinholes provided better performance than the original 9-pinhole configuration for the acquisition with 2 or 4 angular views, but a similar performance with 8 and 16 views.