Real-Time Elastography for Distinguishing Angiomyolipoma From Renal Cell Carcinoma: Preliminary Observations

OBJECTIVE. The objective of this study was to determine the diagnostic performance of sonoelastography for differentiating angiomyolipomas from renal cell carcinomas. SUBJECTS AND METHODS. Twenty-eight angiomyolipomas and 19 renal cell carcinomas were prospectively examined with real-time elastography. Lesions were classified according to four elastographic patterns on the basis of the distribution of the blue area (representing no strain and hardest tissue component). The elasticity patterns and the strain ratios of the angiomyolipomas and renal cell carcinomas were evaluated independently by two observers. Diagnostic performance and interobserver agreement were analyzed. RESULTS. All angiomyolipomas were classified as having a high-strain elastographic pattern (blue areas in < 50% of lesion, considered type 1 or type 2) by both radiologists, whereas 18 of 19 renal cell carcinomas were classified as having a low-strain elastographic pattern (blue areas in >= 50% of lesion, considered type 3 or 4) by both radiologists. The respective mean strain ratios measured by two radiologists were 0.15 +/- 0.06 and 0.18 +/- 0.09 for the angiomyolipomas and 0.64 +/- 0.15 and 0.63 +/- 0.19 for the renal cell carcinomas. There were significant differences between the elasticity patterns and strain ratios for angiomyolipomas and renal cell carcinomas (p < 0.001). Interobserver agreement was excellent for elasticity patterns and strain ratios, with a weighted kappa coefficient of 0.96 and an intraclass correlation coefficient score of 0.95. CONCLUSION. Our results show that real-time elastography may be useful in differentiating angiomyolipomas from renal cell carcinomas, by use of both elasticity patterns and strain ratios.