Automated body contour detection in SPECT: effects on quantitative studies

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Image Reconstruction

Background:

• Accurate in vivo quantitation in single photon emission computed tomography (SPECT) is crucial for diagnostic accuracy.
• Tissue attenuation correction is a significant challenge in SPECT imaging.
• Existing methods for body contour detection can impact image quantitation.

Purpose of the Study:

• To develop and evaluate a novel method for detecting body contours to improve tissue attenuation correction in SPECT.
• To assess the impact of contour detection accuracy on image reconstruction using various attenuation correction techniques.

Main Methods:

• Developed a new method for body contour detection using patient axial length and Fourier filtering of contours defined by a bit-plane algorithm.
• Evaluated the method's effect on image reconstruction using precorrection (Sorenson), postcorrection (Chang), weighted backprojection, and radial postcorrection (Tanaka) methods.
• Analyzed counts in regions-of-interest in phantom images reconstructed with different contour accuracy and correction techniques.

Main Results:

• The novel contour detection method rapidly derives best-fit contours.
• Inaccurately detected contours had a greater impact on image quantitation with radial postcorrection, weighted backprojection, and postcorrection methods compared to precorrection.
• Precorrection methods demonstrated greater robustness to contour detection inaccuracies.

Conclusions:

• Accurate body contour detection is vital for reliable SPECT quantitation.
• The developed contour detection method, particularly when used with precorrection techniques, enhances the accuracy of in vivo quantitation in SPECT.
• This method offers a significant improvement for SPECT imaging requiring precise tissue attenuation correction.