Reduction of CT dose for CT-based PET attenuation correction

Abstract

One important goal of quantitative imaging using positron emission tomography (PET) combined with X-ray computed tomography (CT) is to accurately measure a tumor's characteristics both before and during therapy to determine as early as possible the efficacy of the treatment. The transmission CT scans in the dual modality PET/CT are used for attenuation correction of the PET emission data. Two significant challenges for quantitative PET/CT imaging come from respiratory motion in lung cancer imaging and estimation of the attenuation coefficients for high atomic number materials in bone imaging. Longer duration respiratory-gated CT has been proposed for attenuation correction of phase-matched respiratory-gated PET and motion estimation. Dual energy CT (DECT) has been proposed for accurate CT based PET attenuation correction (CTAC) for bone imaging. However, for both methods, the radiation dose from the CT scan is unacceptably high with the current CT techniques. This directly limits the clinical application of the quantitative PET imaging. Ultra-low dose CT for PET attenuation correction is studied for lung cancer imaging. Selected combinations of dose reduced acquisition and noise suppression methods are investigated by taking advantage of the reduced requirement of CT for PET CTAC. The impact of these methods on PET quantitation is evaluated through simulations on different digital phantoms. When CT is not used for diagnostic and anatomical localization purposes, it is shown that ultra-low dose CT for PET/CT is feasible. The noise and bias propagation from DECT acquisitions to PET or SPECT are studied for bone imaging, and related dose minimization are investigated. It is shown that through appropriate selection of CT techniques, DECT could deliver the same radiation dose as that of a single spectra CT and provide accurate attenuation correction for PET imaging containing high-Z materials. Finally, phantom-based measured experiments are performed to characterize the simulation and provide spectra validation.