Design of a Fast, High-Resolution System for Small Animal Volume MicroCT


Eric C. Frey, Yong Du, and Martin Stumpf



Currently, most commercial x-ray micro-computed tomography (microCT) systems require long acquisition times (ranging from several minutes to tens of minutes) and have relatively poor soft tissue contrast. In addition, due to the relatively long acquisition times, the use of conventional intravenous contrast agents and fast, dynamic acquisitions are impossible. This limits their use for applications such as perfusion imaging. As a result, the utility of microCT systems for small animal imaging has not matched the utility of x-ray CT for humans. One recently introduced commercial system has acquisition speeds as fast as 1 second, but at the cost of image resolution (which is limited to 150 mm), high scanner cost (> 1 million dollars) and large scanner size (it is based on a conventional CT gantry). The ultimate goal of this project is to develop a compact microCT system capable of both fast dynamic and high-resolution in vivo scanning of small animals.


The acquisition time of a microCT system is determined by 3 major factors: exposure, readout and motion times. The exposure time is the time when the animal is exposed to the x-ray beam in order to acquire a projection image with sufficiently low noise to provide useful reconstructions. The readout time is the time required to read the image from the x-ray detector and the motion time is the time required to rotate the detector from one view to the next. In our current microCT system, which is typical of many first-generation commercial systems, these times are 0.3, 1.2, and 0.2 seconds, respectively, per projection. Total acquisition times are on the order of 3 to 10 minutes depending on the number of projection views . The proposed system would use five imaging chains each composed of a detector and an x-ray tube. These chains will be spaced evenly around object imaged. This immediately reduces the effective exposure time, readout time, and motion time by a factor of five. In addition, we propose to further reduce the readout time by using a fast CZT x-ray imaging panel that is capable of acquiring 50 frames per second. In addition, the quantum efficiency of this detector is higher than that of typical phosphor/CCD detectors, allowing further reduction of exposure time for the same noise level. To reduce the motion time we propose to rotate the system continuously rather than using a step-and-shoot system. To minimize motion blur and allow retrospectively gated images we propose to use a gated x-ray source. One limitation of the proposed multi-chain design is the crosstalk between the chains due to scattered photons. To investigate these effects we performed experimental measurements and Monte Carlo simulation studies. To investigate the effect of detector lag and motion blur we performed simulation studies using a simulated resolution phantom.


The resulting system would be capable of obtaining a volume image of a mouse in 0.5 s with a resolution of approximately 70 mm. We found that detector lag and motion blur with the proposed design should not be a major problem. Similarly, the magnitude of scatter from neighboring imaging chains is less than 1% of the incident image intensity and simulations indicate that it will not be a major problem.


The resulting system would allow fast dynamic imaging as well as imaging using conventional contrast agents. The use of multiple sources has the potential to allow for multi-kVp imaging. All of these factors could increase the range of useful applications for small animal microCT imaging.


Internally funded. Hopefully this work will be funded by the Public Health Service .