Fundamental image quality limits for microcomputed tomography in small animals.

Small-animal imaging has become increasingly more important as transgenic and knockout mice are produced to model human diseases. One imaging technique that has emerged is microcomputed tomography (micro-CT). For live-animal imaging, the precision in the images will be determined by the x-ray dose given to the animal. As a result, we propose a simple method to predict the noise performance of an x-ray micro-CT system as a function of dose and image resolution. An ideal, quantum-noise limited micro-CT scanner, assumed to have perfect resolution and ideal efficiency, was modeled. Using a simplified model, the coefficient of variation (COV) of the linear attenuation coefficient was calculated for a range of entrance doses and isotropic voxel sizes. COV calculations were performed for the ideal case and with simulated imperfections in efficiency and resolution. Our model was validated in phantom studies and mouse images were acquired with a specimen scanner to illustrate the results. A simplified model of noise propagation in the case of isotropic resolution indicates that the COV in the linear attenuation coefficient is proportional to (dose)(-1/2) and to the (isotropic voxel size)(-2) in the reconstructed volume. Therefore an improvement in the precision can be achieved only by increasing the isotropic voxel size (thereby decreasing the resolution of the image) or by increasing the x-ray dose. For the ideal scanner, a COV of 1% in the linear attenuation coefficient for an image of a mouse exposed to 0.25 Gy is obtained with a minimum isotropic voxel size of 135 microm. However, the same COV is achieved at a dose of 5.0 Gy with a 65 microm isotropic voxel size. Conversely, for a 68 mm diameter rat, a COV of 1% obtained from an image at 5.0 Gy would require an isotropic voxel size of 100 microm. These results indicate that short-term, potentially lethal, effects of ionizing radiation will limit high-resolution live animal imaging. As improvements in detector technology allow the resolution to improve, by decreasing the detector element size to tens of microns or less, high quality images will be limited by the x-ray dose administered. For the highest quality images, these doses will approach the lethal dose or LD50 for the animals. Approaching the lethal dose will affect the way experiments are planned, and may reduce opportunities for experiments involving imaging the same animal over time. Dose considerations will become much more important for live small-animal imaging as the limits of resolution are tested.

[1]  Robertson Js,et al.  Vertebrate Radiobiology (Lethal Actions and Associated Effects) , 1957 .

[2]  F Sato,et al.  Late effects of whole or partial body x-irradiation on mice: life shortening. , 1981, International journal of radiation biology and related studies in physics, chemistry, and medicine.

[3]  D. Wagner,et al.  Prominent role of P-selectin in the development of advanced atherosclerosis in ApoE-deficient mice. , 2000, Circulation.

[4]  J. Ordóñez‐Llanos,et al.  Expression of human apolipoprotein A-II in apolipoprotein E-deficient mice induces features of familial combined hyperlipidemia. , 2000, Journal of lipid research.

[5]  David C. Lee,et al.  Transgenic Mice Reveal Roles for TGFα and EGF Receptor in Mammary Gland Development and Neoplasia , 1997, Journal of Mammary Gland Biology and Neoplasia.

[6]  D L Parker,et al.  Design constraints in computed tomography: a theoretical review. , 1982, Medical physics.

[7]  G. Barnes,et al.  Semiempirical model for generating tungsten target x-ray spectra. , 1991, Medical physics.

[8]  R. Brooks,et al.  Statistical limitations in x-ray reconstructive tomography. , 1976, Medical physics.

[9]  J C Gore,et al.  STATISTICAL LIMITATIONS IN COMPUTED TOMOGRAPHY , 1978 .

[10]  R. Mole,et al.  Quantitative observations on recovery from whole body irradiation in mice. II. Recovery during and after daily irradiation. , 1957, The British journal of radiology.

[11]  D. Holdsworth,et al.  Micro-CT in small animal and specimen imaging , 2002 .

[12]  L. Feldkamp,et al.  Practical cone-beam algorithm , 1984 .

[13]  A Fenster,et al.  A high-resolution XRII-based quantitative volume CT scanner. , 1993, Medical physics.

[14]  W. Muller,et al.  Mammary gland neoplasia: insights from transgenic mouse models , 2000, BioEssays : news and reviews in molecular, cellular and developmental biology.

[15]  S J Riederer,et al.  Noise Due to Photon Counting Statistics in Computed X‐Ray Tomography , 1977, Journal of computer assisted tomography.

[16]  D Heinegård,et al.  Up-regulation of cartilage oligomeric matrix protein at the onset of articular cartilage degeneration in a transgenic mouse model of osteoarthritis. , 2000, Arthritis and rheumatism.

[17]  W Huda,et al.  Radiation exposure and image quality in chest CT examinations. , 2001, AJR. American journal of roentgenology.

[18]  R F KALLMAN,et al.  The influence of strain on acute x-ray lethality in the mouse. I. LD50 and death rate studies. , 1956, Radiation research.

[19]  K Faulkner,et al.  Noise and contrast detection in computed tomography images. , 1984, Physics in medicine and biology.