TY - JOUR T1 - Image quality vs. sampling strategy for offset flat-panel cone-beam CT JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 2007 LP - 2007 VL - 52 IS - supplement 1 AU - Eberhard Hansis AU - David Sowards-Emmerd AU - Joerg Bredno AU - Jiong Wang AU - Lingxiong Shao Y1 - 2011/05/01 UR - http://jnm.snmjournals.org/content/52/supplement_1/2007.abstract N2 - 2007 Objectives Flat-panel cone-beam computed tomography using a laterally offset flat-panel X-ray detector is being used for localization and attenuation correction in SPECT/CT, as well as for target localization in image-guided radiotherapy. In this contribution we investigate the image quality implications of different acquisition sampling strategies, namely the influence of the angular sampling (number of projections per spin), vertical detector binning, and pulsed vs. continuous X-ray modes. Modified sampling strategies could lead to accelerated acquisitions and data processing, and could potentially enable dose reduction. Methods Different sampling strategies were compared quantitatively on simulated software phantom data, as well as on scans of two anthropomorphic hardware phantoms and a quality assessment phantom, performed on a Philips BrightView XCT system. Reconstructions were generated using un-regularized iterative maximum-likelihood reconstruction with 1 mm isotropic voxels. Results A marked decrease in image quality is found when reducing the projection number below 360, from the current value of 720. Slight sub-sampling artifacts are often hidden by noise that is present in the acquisitions. Continuous and pulsed X-ray modes perform differently: the former conserves total dose and suffers mainly from angular averaging effects; the latter suffers more from noise and streak artifacts with reduced total dose at reduced projection number. The vertical detector pixel size after binning can be increased to 1.17 mm from the current default of 0.39 or 0.78 mm without loss in axial resolution. Sample phantom and clinical images with different simulated acquisition strategies are shown. Conclusions Sub-sampled acquisitions could be acceptable for many practical applications. Pulsed and continuous X-ray modes perform substantially differently ER -