RATIONALE AND OBJECTIVES. The use of rapid dynamic magnetic resonance (MR) imaging after bolus intravenous contrast injection, to improve detection and delineation of parenchymal disease, was evaluated in an experimental model of splenic metastasis. METHODS. An experimental model for splenic metastasis was first developed in the New Zealand White rabbit, Magnetic resonance studies were then obtained at 1.5 tesla in six animals, A 25 ga needle was used to penetrate the spleen and inject 0.1 mt of minced, screened VX2 adenocarcinoma (obtained from a carrier rabbit), The injections were performed by two techniques, percutaneously using ultrasound guidance (n = 3) and at the time of abdominal laparotomy (n = 3), The animals were imaged at 1.5 tesla on day 10 after implantation, Breath-hold T2-weighted and T1-weighted scans were acquired prior to contrast injection, A dose of 0.3 mmol/kg gadoteridol (ProHance(R)) was then administered intravenously using an MR-compatible power injector, with both dynamic and delayed postcontrast scans obtained, The lesion was confirmed in each animal by gross pathologic and microscopic exam. RESULTS. On region of interest analysis of T2-weighted scans, the lesions could not be differentiated by signal intensity (with any statistical significance) from normal surrounding splenic parenchyma, Lesion conspicuity, assessed by signal difference/noise ratio on dynamic turbo-FLASH scans, increased from 6 +/- 5 precontrast to a peak of 16 +/- 5 at 31 seconds postcontrast, with P < 0.003 (n = 6), Lesion conspicuity steadily diminished from this time to 5 minutes postinjection, although it was still improved at 5 minutes over precontrast scans. CONCLUSIONS. On early dynamic contrast enhanced breath-hold MR, marked improvement in the differentiation of splenic metastases from surrounding normal parenchyma is achieved, The capability of dynamic MR in this regard is analogous to that demonstrated for helical computed tomography.