MRI of Wilms' tumor: promise as the primary imaging method.

The magnetic resonance appearance of Wilms' tumor in 14 patients is described and its clinical utility is evaluated. In all cases, magnetic resonance was correlated with surgical and pathologic findings to assess accuracy. Magnetic resonance accurately identified the primary tumor and its renal origin in all cases, and tumor margins and local extension were accurately demonstrated. Tumor margins were smooth and well defined in nine of 12 cases. Local extension and size were accurately assessed, but because capsular invasion could not be predicted, four surgically proven instances of capsular invasion were missed. Metastatic spread into the liver and inferior vena cava was well documented in four cases and excluded in 10. Magnetic resonance was sensitive for identifying lymph-node enlargement in five of 14 cases, but could not predict the etiology of the enlargement. All Wilms' tumors had signal intensities consistent with prolonged T1 and T2 relaxation times. Signal intensity was highly variable, mainly because of necrosis and hemorrhage within the tumor. Magnetic resonance based on signal intensity could not distinguish Wilms' tumor from other solid renal tumors. Magnetic resonance has the potential for providing the same information as computed tomography, sonography, liver spleen radionuclide scanning, and excretory urography. Although expensive, magnetic resonance will be cost-effective if it can replace all the above techniques. This limited study indicates that magnetic resonance has promise as the primary imaging technique for Wilms' tumors.