Small field-of-view dedicated cardiac SPECT systems: impact of projection truncation

Abstract

Purpose

Small field-of-view (FOV) dedicated cardiac SPECT systems suffer from truncated projection data. This results in (1) neglect of liver activity that otherwise could be used to estimate (and subsequently correct) the amount of scatter in the myocardium by model-based scatter correction, and (2) distorted attenuation maps. In this study, we investigated to what extent truncation impacts attenuation correction and model-based scatter correction in the cases of ^99mTc, ²⁰¹Tl, and simultaneous ^99mTc/ ²⁰¹Tl studies. In addition, we evaluated a simple correction method to mitigate the effects of truncation.

Methods

Digital thorax phantoms of different sizes were used to simulate the full FOV SPECT projections for ^99mTc, ²⁰¹Tl, and simultaneous ^99mTc/ ²⁰¹Tl studies. Small FOV projections were obtained by artificially truncating the full FOV projections. Deviations from ideal heart positioning were simulated by axially shifting projections resulting in more severe liver truncation. Effects of truncation on SPECT images were tested for ordered subset (OS) expectation maximization reconstruction with (1) attenuation correction and detector response modelling (OS-AD), and (2) with additional Monte-Carlo-based scatter correction (OS-ADS). To correct truncation-induced artefacts, we axially extended truncated projections on both sides by duplicating pixel values on the projection edge.

Results

For both ^99mTc and ²⁰¹Tl, differences in the reconstructed myocardium between full FOV and small FOV projections were negligible. In the nine myocardial segments, the maximum deviations of the average pixel values were 1.3% for OS-AD and 3.5% for OS-ADS. For the simultaneous ^99mTc/ ²⁰¹Tl studies, reconstructed ²⁰¹Tl SPECT images from full FOV and small FOV projections showed clearly different image profiles due to truncation. The maximum deviation in defected segments was found to be 49% in the worst-case scenario. However, artificially extending projections reduced deviations in defected segments to a few percent.

Conclusion

Our results indicate that, for single isotope studies, using small FOV systems has little impact on attenuation correction and model-based scatter correction. For simultaneous ^99mTc/ ²⁰¹Tl studies, artificial projection extension almost fully eliminates the adverse effects of projection truncation.