Pixelated scintillators for ultra-high resolution cone-beam CT

PROJECT SUMMARY / ABSTRACT Indirect-detection x-ray Flat Panel Detectors (FPDs) use a scintillator to convert x-rays to light, which is then detected by photodiodes coupled to readout electronics. During the past two decades, the latter evolved from noisy amorphous Silicon arrays with coarse pixels (>130 µm) to low-noise CMOS sensors with very fine pixels (<100 µm). There has been comparatively little innovation in the scintillator, which is now the principal factor limiting FPD spatial resolution. The scintillator bur is caused by the lateral spread of light between the x-ray interaction site and the photodiodes. To overcome this challenge, we propose to use laser ablation to pixelate the scintillator film – here, a ~700 µm thick CsI:Tl commonly used in FPDs - at a pitch matching the readout array. We will develop Atomic Layer Deposition (ALD) techniques to coat the high-aspect ratio pixelation grooves with an optimized combination of absorptive and reflective layers to ensure that there is no inter-pixel cross-talk and that the majority of x-ray interaction light is directed towards the photodiodes. To mitigate signal losses due to the loss of film volume in pixelation grooves, we will use a novel crystalline form of micro-columnar CsI:Tl (CMS CsI:Tl) which enhances the sensor signal-to-noise ratio (SNR) by exploiting the higher density and increased transparency of the crystalline material. The proposed scintillator has the potential to substantially improve the performance of modern FPDs by enabling ultra- high spatial resolution imaging without sacrificing x-ray attenuation provided by using a relatively thick CsI:Tl. We have performed initial experimental studies of this approach and found ~20% better limiting spatial resolution than conventional detectors. This proposal will build on this early work by (i) refining the pixelation technique to obtain even thinner pixel grooves for better detection efficiency, and (ii) optimizing the CMS deposition process and the inter-pixel coatings for improved light output. Our technology will benefit applications where visualization of ~100 µm details is desired, but currently challenged by image noise due to body size and/or patient dose: 2D and 3D angiography, pulmonology, breast, otolaryngology imaging, and orthopedics. For an initial demonstration of potential clinical utility, we target quantitative in vivo assessment of bone microarchitecture in osteoporosis (OP) and osteoarthritis (OA). We will pursue the following specific aims: Aim 1: Optimize the pixelation process to maximize spatial resolution, detection efficiency, and brightness of the pixelated CsI:Tl films. Achieve light yield approaching that of a conventional scintillator and improved SNR for ~100 µm features. Aim 2: Validation in quantitative high-resolution Cone Beam CT of trabecular bone. Characterize of the pixelated detector in terms of fundamental metrics of CBCT imaging performance and in 3D trabecular measurements (bone volume, trabecular thickness and spacing) in a range of body sites pertinent to OA and OP. The results will inform future development of other possible applications of FPDs based on pixelated scintillators with high detective efficiency, e.g. in interventional radiology and pulmonary imaging.