Abstract
PURPOSE: Implantation of carmustine wafers (Gliadel) in vivo is accompanied by characteristic serial changes on MRI and CT, such as transient hyperintensity of the wafers on T(1)-weighted images (T(1)WIs) and considerable gas accumulation in surgical resection cavities. The purpose of this study was to evaluate intrinsic imaging changes to carmustine wafers in vitro. METHODS: Three phantoms simulating a surgical resection cavity were constructed. Each contained either a carmustine wafer fixed with oxidized regenerated cellulose and fibrin sealant, an unfixed carmustine wafer, or a fixed polyethylene control disk, immersed in phosphate-buffered saline. Image acquisition of the phantoms was performed on MRI and CT until 182 days after construction. The radiological appearances of the object in each phantom were assessed by visual evaluation and quantification of the region of interest. The volume of gas around the objects at 24 h after constructing the phantoms was also measured. RESULTS: The carmustine wafers showed low signal intensities on T(1)WIs and T(2)-weighted images (T(2)WIs), and high densities on CT images at 24 h. The signal intensities and CT densities gradually approximated those of saline over a period of months. However, the carmustine wafers never showed hyperintensity on T(1)WIs in vitro. The fixed carmustine wafer showed slower radiological changes, as compared to the unfixed wafer. The gas volume around the fixed carmustine wafer was greater than that around the fixed control disk. CONCLUSION: Changes to the carmustine wafers probably reflected penetration of fluid inside and degradation of the hydrophobic matrix. Reported transient hyperintensity of wafers on T(1)WIs in vivo is regarded as the result of biological reactions, whereas the initial production of gas is considered as an intrinsic characteristic of wafers.