A study of the performance characteristics of an X-ray detector for region-of-interest angiography

Minimally invasive image-guided neuro-vascular interventions require very high image resolution and quality, specifically over regions-of-interest (ROI) crucial to the procedure. ROI images allow limited patient integral radiation dose while permitting rapid frame transfer of high-resolution images.; The design and performance of a CCD based x-ray detector was assessed for neuro-vascular procedures. The detector consists of a 250 μm thick CsI(Tl) phosphor fiber-optically coupled through a 2:1 taper to a CCD chip, with an effective image pixel size of 50 μm and a frame rate of 5 fps in the 2:1 pixel-binned mode. The characteristics of the camera including the MTF, NEQ, DQE and observer performance were evaluated experimentally and through theoretical modeling. The MTF was non-zero (>1%), at the Nyquist frequency of 10 cycles/mm while the DQE(0) had a value of ∼54%. All values were measured using head equivalent attenuating material in the beam at 80 kVp. Observer studies performed using the 2 Alternative Forced Choice (2AFC) method, revealed that 100 μm diameter iodinated vessels with 2 cm length could be seen with greater than 75% confidence level. The theoretical modeling included parallel cascade calculations for the signal and noise characteristics of the detector chain. The observer studies included an ideal observer performance calculation based on the integral signal to noise ratio in the image. Probabilities of visualization of various objects of interest in a neuro-intervention were assessed. There was good agreement between experiment and theory and the results suggest ways for further performance improvement for the ROI CCD camera. Additionally, possible direct flat panel ROI detectors designs were also considered for the future.