Novel reconstruction of low-dose DCP-CT images using a regularized least-squares method based on voxel-level TAC correction (RLS-VC)

Abstract Dynamic cerebral perfusion computed tomography (DCP-CT) is an advanced imaging technique that helps in the clinical diagnosis of cerebrovascular diseases (CVDs). However, radiation dose deposition during repeated CT scans seriously limits its clinical application. In this study, we propose a regularized least-squares method with high interpretability based on voxel-level time-attenuation curve (TAC) correction (RLS-VC) for DCP-CT image reconstruction with a dual low-dose imaging protocol that involves both sparse sampling and low-mAs X-ray emission. The theory of third-order Hermite interpolation (THI) is applied to voxel-level TAC correction during dynamic image reconstruction. The peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) are used to quantitatively evaluate the proposed method in terms of imaging accuracy and noise reduction, while hemodynamic maps, including cerebral blood flow (CBF) and cerebral blood volume (CBV), are calculated to validate its ability to restore hemodynamic parameters. It is proven that the proposed RLS-VC method for low-dose DCP-CT imaging has better performance than several state-of-the-art dynamic CT imaging methods, including PICCS, ndiNLM and PIDT, as well as the commercial FBP method. It can be expected that the RLS-VC method can play an important role in the promotion of the clinical application of DCP-CT for the diagnosis of CVD.