Purpose: To determine the accuracy of breath-hold magnetic resonance (MR) imaging for preoperative staging of patients with organ-confined (stage I) renal cell carcinoma.
Materials and methods: Preoperative MR examinations of 43 patients (50 lesions) who underwent nephrectomy were reviewed. The MR examination consisted entirely of breath-hold sequences, and images were retrospectively evaluated by 2 blinded radiologists. Reviewers independently evaluated each case for findings that could affect the radiologic staging, particularly those that distinguish between organ-confined (stage I) and non-organ-confined (>stage II) disease. Each reviewer assigned a stage, and results were correlated with findings at surgery and pathologic examination.
Results: The difference between both reviewers and pathologic findings in evaluating an intact renal capsule (stage I) was statistically significant (P < 0.05) and resulted in a statistically significant difference between radiologic and pathologic staging (Wilcoxon test, P < 0.05). The kappa test demonstrated moderate agreement between radiologic and pathologic staging (82% and 80% for reviewers 1 and 2, kappa = 0.54 and 0.80, respectively) and substantial agreement (90%, kappa = 0.80) between the 2 reviewers in assigning a radiologic stage.
Conclusion: Breath-hold MR imaging has an accuracy ranging between 80% and 82% in staging patients with organ-confined renal cell carcinoma, with substantial (90%) agreement between readers.