Is Acoustic Radiation Force Impulse (ARFI) elastography a useful method for predicting kidney pathology?

Purpose: Was to establish if ARFI is a useful method for predicting the kidney pathology. Methods: Our study included 110 subjects: 78 “normal” subjects (70.9%) (including healthy volunteers without known kidney pathology with normal renal ultrasonography, without history of renal disease, without supplementary investigations and hospitalized patients for various diseases with normal renal ultrasonography, normal serum creatinine and blood nitrogen urea and without proteinuria or hematuria) and 32 (29.1%) patients with various etiologies of kidney diseases. The mean age of subjects was 48.5 ± 18.4 years, 47.9% being women and 52.1% men. Five valid ARFI measurements were performed in the parenchyma of the each kidney with the subjects in left or right lateral decubitus, median values were calculated and expressed in meters/second (m/s). Results: Valid ARFI measurements were obtained in all subjects in the right kidney and in 98.1% of cases in the left kidney. The mean ARFI values obtained in the right and left kidney were similar in “normal” kidneys and also in kidneys with pathology: 2.37 ± 0.74 m/s vs. 2.38 ± 0.75 m/s, p = 0.92 and 2.02 ± 0.74 m/s vs. 2.05 ± 0.88 m/s, p = 0.91. Because the mean value of KS was similar in the right and left kidney in both “normal” subjects and those with kidney pathology, we chose to study the influence of kidney diseases on KS values obtained in the right kidney. The mean ARFI values obtained in the right kidney were significantly higher in “normal kidneys” as compared with kidneys with pathology: 2.37 ± 0.74 m/s vs. 2.02 ± 0.74 m/s, p = 0.02. For a KS cut-off value≤2.59 m/s, ARFI had 81.2% Se, 41% Sp, 36.1% PPV and 84.2% NPV (AUROC = 0.621, p = 0.03) for predicting the presence of kidney pathology. Conclusion: KS values obtained by ARFI are significantly lower in kidney with pathology as compared with “normal” kidneys. For a cut-off value≤2.59 m/s ARFI had an enough good Se and NPV for predicting the presence of kidney pathology.