Super-resolution x-ray luminescence optical tomography imaging

摘要

X-ray luminescence optical tomography (XLOT) is a promising in vivo noninvasive imaging technique. By using x-raysto irradiate nanophosphors (NPs) in imaged region, XLOT can achieve better imaging depth and higher imagingsensitivity than the widely used optical molecular tomographic imaging techniques, e.g., bioluminescence tomography(BLT) or fluorescence molecular tomography (FMT). However, compared with the anatomical imaging techniques, e.g.,x-ray computed tomography (XCT), XLOT has the disadvantage of low spatial resolution limited to millimeters. Toovercome the limitation, recently, many efforts have been dedicated to optimize the data acquisition schemes andimprove reconstruction methods. Nevertheless, challenges remain in XLOT due to the light scattering in biologicaltissues and the severely ill-condition and ill-posed of the inverse problem of XLOT. To improve the spatial resolution ofXLOT, inspired by super-resolution localization optical microscopy, in this work, we propose a novel imaging method,termed as SR-XLOT, which is achieved by locating the position of NPs in each frame by using the single emitterlocalization methods (e.g., Gaussian fitting method). After reconstructing the positions of NPs in each frame, asuper-resolution XLOT image can be generated by superimposing the identified positions of NPs from all frames intoone image. To evaluate the performance of the proposed SR-XLOT method, a series of numerical simulationexperiments were performed. The experimental results indicate that when using SR-XLOT method, the spatial resolutionof XLOT can be significantly improved, compared with the conventional reconstruction methods. As a result, themethod makes it possible to implement a super-resolution XLOT imaging, which is attractive for medical diagnostic anddrug research.