Dual energy CT for attenuation correction with PET/CT.

Area of Science:

• Medical Imaging Physics
• Radiological Sciences
• Computed Tomography

Background:

• Quantitative imaging with PET/CT and SPECT/CT is increasing.
• Monoenergetic images synthesized from dual-energy CT (DECT) are valuable for estimating attenuation coefficients at PET and SPECT energies.
• Variations in noise and bias of synthesized monoenergetic images with energy require evaluation.

Purpose of the Study:

• To evaluate energy-dependent noise and bias properties of monoenergetic images synthesized from DECT.
• To assess the suitability of these images for PET and SPECT attenuation correction.
• To compare DECT-based attenuation correction with single-kVp CT-based correction for PET quantitation.

Main Methods:

• Used analytic approximations and simulations to estimate noise and bias of synthesized monoenergetic images (40-520 keV).
• Employed dual-kVp spectra (80 and 140 kVp) from a GE Lightspeed VCT scanner with 0.5 mm Cu filtration.
• Evaluated noise suppression (sinogram smoothing) and dose minimization strategies, comparing DECT with single-kVp CT for PET attenuation correction.

Main Results:

• Minimum noise observed at an energy between the mean energies of the two spectra.
• Normalized coefficient of variation in attenuation maps plateaued around 160 keV and remained constant up to 511 keV.
• Bias was minimal (within 2.4% of true values); sinogram smoothing reduced noise significantly without increasing bias.
• DECT with optimized techniques achieved similar radiation dose to single CT, reducing PET bias in contrast regions by ≈2.6x and slightly reducing overall RMSE.

Conclusions:

• DECT for higher-energy attenuation correction exhibits energy-dependent noise amplification.
• Sinogram smoothing effectively reduces noise in DECT-derived attenuation maps without introducing bias.
• Optimized DECT techniques offer comparable radiation dose to single CT while enhancing PET image quantitative accuracy.