Abstract
28
Objectives: Traditional protocols for myocardial perfusion imaging (MPI) with SPECT use a fixed administered patient dose. A constant dose leads to lower image quality in heavier patients due to increased attenuation. Some clinics adjust the administered dose for patients with body mass greater than 100 kg, and the American (ASNC) guidelines call for a linear adjustment of 11.5 MBq/kg for patients heavier than 70 kg. A study using visual assessment of image quality showed that administered patient dose is related to patient weight by a power function but that this relationship can be approximated as linear. We propose to use an observer-independent quantification of noise to determine the weight-based adjustment of dose that is required to minimize noise variation between patients. We have previously developed a technique to estimate image noise from a single patient acquisition and the corresponding attenuation map. Our objective was to employ this technique to determine a weight-based dose function which would provide the adjusted dose needed to produce a constant level of noise.
Methods: Voxel-by-voxel image noise was calculated from the square root of the sum of detected counts using an attenuated forward projection from each voxel of a reconstructed image. The mean noise estimate for voxels in the myocardium was calculated for rest studies of 27 patients (17 male) acquired on a dedicated cardiac SPECT camera (350 MBq Tc-99m tetrofosmin, 5 min acquisition, patient weights were 45-102 kg). Approximating a linear relationship, dose/patient weight [MBq/kg] was plotted as a function of the estimated noise. This was repeated for 2×, ½, and ¼ dose simulations.
Results: The fit to the data was found to be D/W = 395 · σ-1.922 (R2 = 0.97) where D was the administered dose [MBq], W was the patient weight [kg], and σ was the % standard deviation (noise) in a reconstructed image voxel. Male and female patients showed equivalent results when evaluated separately. Among the 27 patients, image noise varied with patient weight from 6.5-13.5% for a 350 MBq fixed administered dose; image noise was 9.7% for a 70 kg patient. Maintenance of a constant noise level of (9.9 ± 1.1)% for all patients, required a dose per weight of 4.8192 MBq/kg (225-510 MBq for the range of weights in our patient set). For a half-dose protocol, an administered dose of 175 MBq in a 70 kg patient gives a noise level of 13.9%. Maintaining this noise level required a dose per weight of 2.4759 MBq/kg. For our patient set, this equated to 113-255 MBq and produced a noise level of (14.0 ± 1.5)%.
Conclusion: We developed a method for calculating the weight-based administered dose required to achieve a consistent image noise level over a range of patient body weights. We evaluated the method for a dedicated cardiac pinhole SPECT camera, but the technique could also be applied to other SPECT cameras. Research Support: This work is supported by the M. Hildred Blewett Fellowship of the American Physical Society, www.aps.org, the Queen Elizabeth II Graduate Scholarship in Science and Technology, the Kiwanis Club of Ottawa Medical Foundation and Dr. Marwah, and NSERC grants: RGPIN 261765-2011 and RGPIN 2016‑05658.