Addressing current limitations in dual-energy computed tomography

Abstract

Dual-energy CT (DECT) provides better material differentiation compared to conventional CT. However, DECT is fundamentally limited by noise amplification and the need for dual dataset acquisition. Noise amplification during signal decomposition significantly limits utility of DECT basis material images. Conventional noise suppression algorithms limit signal variation between neighboring pixels which inevitably sacrifices spatial resolution. For noise suppression in DECT, we propose an Image-domain Decomposition method through Entropy Minimization (IDEM) that avoids degradation to spatial resolution by exploiting strong signal correlations between images. As supported by phantom and patient studies, our algorithm has the unique capability of reducing noise standard deviation on DECT decomposed images by roughly one order of magnitude while preserving spatial resolution and image noise power spectra. The need for projection data with two different effective x-ray spectra restricts DECT applications to specialized scanners. We propose a hardware-based acquisition method known as PM-DECT, which utilizes an attenuation sheet with a spatially-varying pattern, known as a primary beam modulator, to enable single-scan DECT on a conventional CT scanner. Phantom studies demonstrate that PM-DECT retains a high level of spatial resolution compared to conventional CT scans and can achieve electron density values with approximately 1% error. Granting the opportunity for high-quality single-scan DECT on conventional CT scanners via limited hardware modification, PM-DECT has the potential to liberate DECT from specialized scanners, extending clinical availability and implementation.