Enhanced K-Edge CT Using Dual-Filter Spectral Separation for High-Sensitivity Imaging of Bismuth
Abstract
Purpose
While CT imaging has advanced with improved machine design, further diagnostic information can be achieved by enhancing sensitivity to high-Z therapeutic agents. Current photon-counting CT (PCCT) k-edge imaging sensitivity is limited by spectral overlap between energy bins, reducing contrast-to-noise ratio (CNR) for high-Z targeted pharmaceuticals present at low concentrations. We demonstrate an enhanced k-edge CT (EKCT) method to increase PCCT sensitivity for detecting high-Z contrast agents.
Methods
We developed EKCT to optimize the k-edge signal of bismuth using dual-filter spectral separation on a benchtop CZT photon-counting detector with programmable energy bins. EKCT employs two distinct energy spectra with extensive filtration to isolate relevant photon fluence: a 110 kVp low-energy spectrum (0.89 mm Bi, 1 mm Cu) concentrating 89% fluence below the bismuth k-edge (90.5 keV), and a 110 kVp high-energy spectrum (6 mm Cu) optimized for above-k-edge detection. The technique was optimized analytically using SpekPy spectrum generation and CZT detector response modeling. CNR was compared between conventional single-spectrum k-edge CT (KCT) and EKCT for bismuth sulfate suspensions with standard Iohexol iodine on both at concentrations of 1mg/ml.
Results
Simulations predicted EKCT improves spectral purity to 89-91% of photons in the correct energy bin versus 72-75% for KCT, reducing spectral overlap from 25-28% to less than 11%. On the benchtop PCCT, conventional iodine contrast achieved CNR of 0.71 (30 HU), while bismuth with KCT achieved CNR of 6.5 (86 HU). Bismuth with EKCT achieved CNR of 12.4 (188 HU), representing a 17-fold improvement over iodine and 119% higher signal than KCT. Material differentiation was preserved: bismuth exhibited positive k-edge signal while iodine showed negative signal.
Conclusion
The EKCT technique combined with bismuth contrast achieves 17-fold higher CNR than conventional iodine imaging through physical spectral separation, enabling improved detection of low-concentration high-Z contrast agents.