Retrospective Study of Lung Shunt Fraction Using Data Acquired from 3D SPECT and 2D Planar Imaging
Abstract
Purpose
A modified lung shunt fraction (LSF) based on conventional 2D planar imaging data was proposed. That modified LSF can reduce the 2D overestimation errors and get closer results with 3D SPECT images.
Methods
3D LSF from attenuation corrected SPECT images is treated as gold standard method, a modified LSF based on 2D planar imaging data was derived. The modified LSF includes both attenuation and scatter correction, through introducing a correction factor k, which is a planar imaging system and patient dependent factor. Factor k was in the range of 0.35~0.5 for typical clinical settings.
Results
The test results show that 1) the average difference between 2D LSF and 3D LSF was 3.0% across all patients, while the average difference between the modified LSF and 3D LSF was only 1.2%; 2) For cases where with 2D LSF>5%, the average difference between 2D LSF and 3D LSF was 6.6%, while the average difference between the modified LSF and 3D LSF is only 1.6%；3) For cases where with 2D LSF>10%, the average difference between 2D LSF and 3D LSF was 10.1%, while the average difference between the modified LSF and 3D LSF is only 1.9%. The results indicate that using the modified LSF instead of 2D LSF can reduce the overestimation errors, especially in cases with higher lung shunt fractions.
Conclusion
We proposed a method for determining the corrected LSF based on 2D planar images, which we call the modified LSF. Test results show that using the modified LSF can reduce overestimation errors and yield results closer to those obtained from 3D SPECT/CT images, which we consider the gold standard method, especially in cases with a high lung shunt fraction, where the overestimation error is significantly reduced.