A Phantom Study to Evaluate IQ-SPECT as a Potential Replacement for High Resolution Collimators in Myocardial Perfusion SPECT Imaging

Introduction: Traditionally, High Resolution parallel collimators (HiRes) have been utilized for myocardial perfusion imaging at Northwestern Memorial Hospital, injecting 9 mCi of Tc-99m Sestamibi at rest and 29 mCi at stress. The acquisition times of approximately 20 minutes when using HiRes collimators provided adequate counting statistics and image quality when acquired into pixel sizes of 6.6 mm and 8 frame gating at stress. In the attempt to improve diagnostic accuracy a change was made to a higher resolution 4.8 mm pixel and 16 frame gating at both rest and stress. Filtered back projection reconstruction was replaced with an iterative 3D reconstruction algorithm (Flash-3D) to compensate for the increase in image noise. Resting (9 mCi) non-gated image quality remained adequate but 16 frame gated data sets proved to be count poor, noisy and unusable for wall motion and ejection fraction evaluation. Siemens introduced the IQ-SPECT (IQ) technique which utilizes high sensitivity, cardio centric SMARTZOOM collimators in combination with Flash-3D in an effort to reduce imaging time and/or injected dose while maintaining adequate image quality. The objective of this phantom study is to evaluate IQ as an alternative to traditional HiRes parallel hole acquisitions to provide adequate image quality for low dose 16 frame gated imaging.

Methods: An anthropomorphic torso phantom with a cardiac insert was used to simulate a standard patient. The cardiac insert included a simulated mid-inferior wall defect. Clinically relevant amounts of Tc-99m were placed in the cardiac insert, liver and background chambers to simulate a resting 9 mCi dose. The phantom was then positioned on the imaging table and a R-wave generator was used to simulate a 3-lead ECG for gated acquisition. The phantom was first imaged using HiRes collimators and current clinical acquisition parameters – 64 projections, 40 sec/projection totaling an approximate 20-minute acquisition. The phantom was then imaged using the IQ technique using various acquisition parameters with acquisition times ranging from 5 minutes to 19 minutes. The acquisition of 34 projections, 28 sec/projection totaling approximately 16 minutes was utilized for this evaluation as recommended by our Siemens application support specialist.

Results: Both the HiRes non-gated and the non-gated IQ reconstructions allowed good visualization of the mid-inferior wall defect present in the phantom. The IQ data set, however, produced prominent image distortion in the apex, making the apex appear elongated and bulbous. Reconstruction parameters were selected to minimize this distortion. As anticipated, the 16-minute HiRes acquisition resulted in gated images with poor signal/noise ratio, making them questionable for wall motion and ejection fraction determinations. The 16-frame gated IQ showed much improved counting statistics providing what appears to be adequate image quality to allow for wall motion and ejection fraction evaluation.

Conclusions: The result of this phantom study indicated that IQ may be a viable alternative to HiRes cardiac SPECT, providing adequate counting statistics and image quality to allow wall motion and ejection fraction evolution in low dose 16 frame gated-SPECT exams. Further investigation is warranted to evaluate these findings in clinical patient data