Optimal flat-panel detector pixel size for visualization of stents in X-ray fluoroscopy

Visualization of stents is of great interest in X-ray fluoroscopy. For stent visualization, detector pixel size must be small to reduce contrast dilution from partial area effect, but reducing pixel size lowers the number of photons collected, reducing the signal-to-noise ratio (SNR). In this study, we use a computational human observer model to predict human performance for stent detection and stent deployment detection as a function of pixel size. Predictions are made for an ideal and indirect flat-panel (FP) detector at three exposures. For an ideal detector, regardless of the task or exposure, the optimal pixel size was 50 μm. For an indirect detector, optimal pixel size depends on task and exposure. Pixel sizes were 80, 80, and 60 μm for stent detection and 60, 60, and 50 μm for stent deployment, at 5, 10, and 20 μR per frame, respectively.