D-SPECT (Spectrum Dynamics, Israel) is a novel SPECT system for cardiac perfusion studies. Based on CZT detectors, region-centric scanning, high-sensitivity collimators and resolution recovery, it offers potential advantages over conventional systems. A series of measurements were made on a beta-version D-SPECT system in order to evaluate its performance in terms of energy resolution, scatter fraction, sensitivity, count rate capability and resolution. Corresponding measurements were also done on a conventional SPECT system (CS) for comparison. The energy resolution of the D-SPECT system at 140 keV was 5.5% (CS: 9.25%), the scatter fraction 30% (CS: 34%), the planar sensitivity 398 s(-1) MBq(-1) per head ((99m)Tc, 10 cm) (CS: 72 s(-1) MBq(-1)), and the tomographic sensitivity in the heart region was in the range 647-1107 s(-1) MBq(-1) (CS: 141 s(-1) MBq(-1)). The count rate increased linearly with increasing activity up to 1.44 M s(-1). The intrinsic resolution was equal to the pixel size, 2.46 mm (CS: 3.8 mm). The average reconstructed resolution using the standard clinical filter was 12.5 mm (CS: 13.7 mm). The D-SPECT has superior sensitivity to that of a conventional system with similar spatial resolution. It also has excellent energy resolution and count rate characteristics, which should prove useful in dynamic and dual radionuclide studies.

Fast and high-quality simultaneous DR MPI is feasible with D-SPECT in a single imaging session with comparable diagnostic performance and image quality to conventional SPECT and to a separate rest (201)Tl D-SPECT acquisition.

No abstract

Background: Compared to filtered backprojection (FBP) and iterative reconstruction with ordered subset expectation maximization (OSEM), wide beam reconstruction (WBR), which incorporates resolution recovery and models noise during reconstruction without applying a post-processing filter, has been reported to allow half-time gated myocardial perfusion single-photon emission computed tomography (SPECT) acquisition with preserved diagnostic quality. We postulated that with further noise modeling even shorter acquisition times would be possible. Methods: The half-time WBR algorithm was modified for "quarter-time" acquisitions based upon anthropomorphic cardiac phantom data and a pilot group of 48 patients (pts). Pilot pts underwent 180-degree, 64-stop, full-time single-day rest (R) (25 second-per-stop [sps]) and stress (S) (20sps), and then "quarter-time" either R (6sps) (nϭ27 pts) or S (4 sps) (n ϭ 21pts) 9 mCi/32 mCi R/S 99m Tc-sestamibi SPECT. A 90°-angled dual-headed camera with high resolution parallel-hole collimators was used. Subsequently, using the same protocol, 134 consecutive pts (61 men, 73 women, mean weight ϭ 182 lbs., mean chest circumference ϭ 41 in.) were imaged both at R and S with full-time FBP and OSEM and also quarter-time WBR using the modified algorithm. Anticipating reconstruction artifacts in low count density R 6sps scans, a R 10sps acquisition was simulated by randomly dropping counts from each stop of the full time R acquisition while maintaining Poisson statistics, and the WBR algorithm was separately optimized for R 10sps SPECT. Blinded observers graded perfusion scans for quality (1ϭpoor to 5ϭexcellent) based on myocardial uniformity, endocardial/epicardial edge definition, and background noise. Perfusion defects were scored using a 17-segment model. Results: For the 134 prospective pts mean image quality for R full-time OSEM and quarter-time WBR was equivalent (3.5) and superior to FBP (3.1) (p Ͻ 0.0001). For S, quarter-time WBR quality (4.2) was superior to both full-time OSEM (3.8) and FBP (3.4) (p's Ͻ 0.0001). Reconstruction artifacts (myocardial "streaks" or clustered hot pixels) were more frequent with quarter-time WBR than with full-time OSEM (14 R, 5 S vs. 1 R, 0 S), but did not confound interpretation. For R WBR, 10sps acquisitions were superior to 6sps (quality 3.7 vs. 3.5, p ϭ .003) and artifacts were less frequent (8 vs.14). In pts with chest circumferences Ն 44 in. (nϭ15), R image quality was better for 10sps than for 6sps (3.6 vs. 3.2, p ϭ 0.03). Of the 19 patients with abnormal scans (summed stress scores [SSSs] Ͼ 2 by OSEM), mean SSSs, summed rest scores, and summed difference scores were not significantly different with quarter-time WBR vs. full-time OSEM (8.6 vs 9.3), (6.9 vs. 8.0), (1.9 vs. 1.3) (p's NS). Only 1 patient with normal full-time OSEM had abnormal quarter-time WBR (SSS ϭ 3). Conclusions: For perfusion SPECT quarter-time WBR affords image quality and defect characterization equivalent to full-time OSEM. Lengthening WBR R acquisitions to 10sps may be advant...