Feasibility of dual-energy CBCT material decomposition in the human torso with 2D anti-scatter grids and grid-based scatter sampling

TL;DR

This study demonstrates that dual-energy CBCT with 2D anti-scatter grids and residual scatter correction can accurately quantify materials like iodine in the human torso, overcoming scatter-related limitations.

Contribution

It introduces a combined approach of 2D anti-scatter grids and measurement-based scatter correction for improved material decomposition in DE CBCT of the human torso.

Findings

Iodine quantification errors were reduced below 1.5 mg/ml with scatter correction.

A 2D anti-scatter grid enables differentiation of iodine and water in torso-sized phantoms.

Robust scatter suppression improves the accuracy of DE material decomposition.

Abstract

Background: Dual-energy (DE) imaging techniques in cone-beam computed tomography (CBCT) have potential clinical applications, including material quantification and improved tissue visualization. However, the performance of DE CBCT is limited by the effects of scattered radiation, which restricts its use to small object imaging. Purpose: This study investigates the feasibility of DE CBCT material decomposition by reducing scatter with a 2D anti-scatter grid and a measurement-based scatter correction method. Methods: A 2D anti-scatter grid prototype was utilized with a residual scatter correction method in a linac-mounted CBCT system to investigate the effects of robust scatter suppression in DE CBCT. Scans were acquired at 90 and 140 kVp using phantoms that mimic head, thorax, and abdomen/pelvis anatomies. The effect of a 2D anti-scatter grid with and without residual scatter correction on iodine concentration quantification and contrast visualization in VME images was evaluated. Results: In CBCT images, a 2D grid with or without scatter correction can differentiate iodine and water after DE processing in human torso-sized phantom images. However, iodine quantification errors were up to 10 mg/ml in pelvis phantoms when only the 2D grid was used. Adding scatter correction to 2D-grid CBCT reduced iodine quantification errors below 1.5 mg/ml in pelvis phantoms, comparable to iodine quantification errors in multidetector CT. Conclusions: This study indicates that accurate DE decomposition is potentially feasible in DE CBCT of the human torso if robust scatter suppression is achieved with 2D anti-scatter grids and residual scatter correction. This approach can potentially enable better contrast visualization and tissue and contrast agent quantification in various CBCT applications.