Joint Temporal and Multi-energy Processing for Photon-Counting Fluoroscopy

Spectral imaging is being increasingly explored for interventional applications, but noise amplification resulting from material decomposition has traditionally impeded clinical adoption. To address this problem, we propose a novel joint processing strategy that utilizes both the pre- and post-contrast acquisitions in angiography applications. By acquiring spectral measurements in both images and jointly process them in a model-based material decomposition approach, we may simultaneously leverage temporal and spectral information to improve imaging performance beyond what is achievable through conventional approaches either information by itself - e.g., temporal subtraction, or spectral imaging using just the post-contrast measurements. Furthermore, the proposed strategy enables three material decomposition to be performed using just two spectral channels, allowing the separation of soft tissue, calcium, and exogenous contrast agent that is not possible using, for example, conventional digital subtraction angiography. We demonstrate the theoretical noise advantage of the joint processing strategy in simulation studies, and show initial experimental results of three material decomposition on a benchtop system with a photon counting detector.