Reconstructing Accurate Synthetic Hounsfield Units with Spectral CT Data

TL;DR

This paper presents a method to reconstruct spectral CT data into accurate Hounsfield Units that mimic conventional CT numbers, reducing artifacts and enabling easier transition from traditional to spectral CT imaging.

Contribution

The paper introduces a novel reconstruction method that produces accurate, artifact-free CT numbers from spectral CT data, aligning with conventional CT standards.

Findings

Reconstructed CT numbers are accurate across various tissues.

Images are free of beam hardening artifacts.

Method facilitates transition from conventional to spectral CT.

Abstract

Purpose: To evaluate a proposed method to reconstruct CT numbers that accurately mimic conventional CT numbers from spectral CT data, as would have been produced by a conventional system without effects of beam hardening. Approach: We implement the proposed method for simulated ideal and non-ideal photon counting multi-bin systems, the latter based on a photon counting Silicon detector, and compare them with a simulated ideal conventional energy integrating system with a cupping correction algorithm. We compare the systems using a mathematical phantom of a size and composition that produces severe cupping and beam hardening artefacts when imaged with a conventional system with no cupping correction. Results: The resulting images show CT numbers that are consistently accurate for a varying range of tissues and are free of beam hardening artefacts. Conclusions: This method could facilitate use of established rules-of-thumb regarding absolute CT numbers for various organs and conditions during the transition from conventional CT to spectral CT.