MRI in large animals in clinical research typically refers to any animal larger than a ferret or rabbit. In veterinary medicine, dogs and cats are considered small animals whereas pigs, horses, sheep, goats, and cows fall under the category of large animals. Unlike imaging in mice and rats, which are typically performed on high-field, small bore MRI scanners, MRI on large animals is routinely performed at clinical field strengths. In veterinary medicine, the bulk of diagnostic MR imaging is neurological scans followed by musculoskeletal imaging. Because of the acoustic noise (1) and the need to prevent motion, MRI scanning in large animals is usually performed under general anesthesia. Inhalational anesthesia, such as isoflurane or sevoflurane, is frequently performed using a ventilator, which has the added advantage of providing a means to perform suspended respiration or breath-holds by pausing the ventilator. However, one must recognize that suspending the ventilator also suspends the anesthetic delivery. Thus, care must be taken not to perform breath holds too frequently less the animal wakes up. Many anesthetics can have profound effects on imaging studies. For instance, several groups have shown that fMRI results can be vastly different in awake dogs versus anesthetized dogs.(2,3) In a similar manner, inhalational anesthesia can cause vasodilation, which can affect hemodynamic studies in the heart. Physiological monitoring to assess patient anesthetic depth becomes more important in anesthetized animals. Fortunately, most MR-compatible stand-alone physiological monitors as well as vendor systems that are used in human patients can be used in large animals without major modifications. Pediatric settings, if available, will usually give more reliable waveforms than adult settings in most large animal patients.Depending on the animal size, standard receiver coils can often be used in large animals. Brain scans can usually be performed using the vendor head or knee coil. Because of the different conformation of the head in most large animals, scan planes in the head will need to be adapted. For instance, images acquired in the axial plane may yield coronal brain scans. Many groups will scan large animals other than non-human primates in the prone position, but our group has also had good success using the supine position for all species. On the other hand, typical knee coils are not very useful for imaging the knee or stifle in most quadrupeds. Flexible receive only coils that are wrapped around the joint of interest can provide better coverage. For extremity imaging, positioning the animal with the joint of interest down and the contralateral leg positioned away from the coil will minimize motion of the joint from gradient switching and also decrease the likelihood of wrap artifact from the non-imaged leg.Species differences from man must also be considered when planning experimental models and imaging. For example, dogs usually have thirteen thoracic vertebrae and seven lumbar vertebrate, but pigs usually have fifteen thoracic vertebrae and six to seven lumbar vertebrae. (People usually have twelve thoracic and five lumbar vertebrae.) In addition, the spinal cord extends further in most quadrupeds than humans. Similarly, the number of liver lobes in different species is distinct from humans. Moreover, the orientation of the heart orientation in the deeper chest of most animals other than non-human primates (NHPs) will require different orientations of the double oblique planes relative to humans to obtain long and short-axis images and can make planning navigators in the diaphragm challenging to avoid saturation bands in the heart.Significant variations may occur between species that must be considered in the choice of animal model. The most classic example is the degree of vascular collateralization in the heart with dogs having the highest degree of collateralization relative to rabbits and pigs, who seldom have any degree of collateralization. As a results, reperfused, myocardial infarction models in dogs will demonstrate drastically different contrast kinetics than in pigs on first, pass contrast-enhanced imaging as well as late gadolinium enhancement. Because most large animals other than NHPs have a complete Circle of Willis, complete ligation of the common carotid artery will seldom cause a stroke. The newest wave in preclinical MRI studies for clinical translation is the use of spontaneously occurring disease in pets as more relevant testing for new pulse sequences, devices, and drug therapies. Due to the high degree of inbreeding in domestic dogs and cats—especially in pure-bred lines, many diseases that are prevalent in patients occur spontaneously in pets. Unlike transgenic mice where specific genes are knocked in or knocked out to cause disease, the genetic mutations in dogs and cats that leads to naturally occurring diseases is often highly variable or not yet elucidated as in human diseases. Presently, over 450 diseases are recognized in domesticated dogs with over 360 of these diseases having analogous disease in humans, including many types of cancer, heart disease and metabolic diseases related to obesity.(4,5) Moreover pets with these diseases may receive many concurrent medications akin to what occurs in our patient population. These models may prove to be better than our current models in which disease conditions are artificially induced in laboratory animal species. In addition, dogs and cats are sufficiently large enough to enable imaging with clinical MRI scanners to enable rapid clinical translation.