Feasibility of Material Decomposition-based Photon-counting CT Thermometry

TL;DR

This paper demonstrates the feasibility of using material decomposition-based photon-counting CT for real-time thermometry during thermal ablation, achieving accurate temperature measurements in simulations.

Contribution

It introduces a novel one-step algorithm for simultaneous material decomposition and temperature estimation in photon-counting CT, improving robustness and accuracy.

Findings

The one-step algorithm provides temperature predictions with centigrade accuracy.

Photon-counting CT thermometry is robust to tissue heterogeneity and noise.

Numerical simulations validate the method's potential for clinical use.

Abstract

Over the past decades, thermal ablation procedures such as high intensity focused ultrasound (HIFU) have been developed vaporize cancerous tissues in a focal area. Thermal ablation has the potential to non-invasively eliminate tumors and treat other medical conditions with high efficacy and better patient outcomes. However, it is necessary for treatment to be coupled with real time temperature monitoring in order to deliver sufficient thermal dosage to the target while sparing surrounding, healthy tissues. To this end, computed tomography (CT) has been explored to offer fine spatial resolution at a rapid speed. However, current CT thermometry techniques are sensitive to heterogeneous tissue properties, imaging noise, and artifacts. To address this challenge, this paper utilizes the emerging photon-counting CT technology, performs tomographic thermometry based on material decomposition, and obtains excellent results in realistic numerical simulation. Specifically, three algorithms were designed to decouple material composition and thermal expansion from spectral CT reconstruction and compared in a comparative study. The best algorithm, referred to as the one-step algorithm, uniquely finds both material decomposition and the associated temperature at the same time in a closed form solution, which is robust to changes in tissue composition and generates temperature predictions with centigrade accuracy under realistic CT image noise levels.