Magnetic resonance (MR) imaging was evaluated for its possible diagnostic application in determining the survival of fetal central nervous system tissue grafts in the injured spinal cord. Hemisection cavities were made at the T11-L1 level of eight adult female cats. Immediately thereafter, several pieces of tissue, either obtained from the fetal cat brain stem on embryonic Day 37 (E-37), from the fetal neocortex on E-37, or from the fetal spinal cord on E-23, were implanted into the cavities made in seven cats. The eighth cat served as a control for the effect of the lesion only. In another group of four animals, a static-load compression injury was made at the L-2 level. Seven weeks later, the lesion was resected in three cases and fragments of either fetal brain-stem or spinal cord tissue were introduced. A small cyst was observed in a fourth cat in the compression injury group and a suspension of dissociated E-23 brain-stem cells was injected into this region of cavitation without disturbing the surrounding leptomeninges. Five months to 2 years posttransplantation, MR imaging was performed with a 2.0-tesla VIS imaging spectrometer by acquiring multislice spin-echo images (TR 1000 msec, TE 30 msec) in both the transverse and sagittal planes. Collectively, these intermediate-weighted images revealed homogeneous, slightly hyperintense signals at the graft site relative to the neighboring host tissue in seven of the 11 graft recipients. Two of the remaining four cats exhibited signals from the graft site that were approximately isointense with the adjacent host spinal cord, and the final two cats and the lesion-only control presented with very hypointense transplant/resection regions. The hyperintense and isointense images were tentatively interpreted as representing viable graft tissue, whereas the hypointense transplant/resection sites were considered to be indicative of a lack of transplant survival or the absence of tissue in the lesion-only control animal. Postmortem gross inspection of fixed specimens and light microscopy verified the MR findings in the control animal in 10 of the 11 graft recipients by showing either transplants and/or cysts corresponding to the MR images obtained. In one cat in the hemisection group, histological analysis revealed a very small piece of graft tissue that was not detected on the MR images. Therefore, it is suggested that within certain spatial- and contrast-resolution limits, MR imaging can reliably detect the presence of transplanted neural tissue in both the hemisected and compression-injured spinal cord of living animals.(ABSTRACT TRUNCATED AT 400 WORDS)