Abstract:

Objective: The goal of this research is to evaluate the efficiency of computed tomography texture analysis in differentiating renal cell carcinoma from a high-attenuation renal cyst on non-contrast computed tomography. Methods: Forty-nine non-contrast abdominal computed tomography examinations, 27 patients with high-attenuation renal cyst and 22 patients with renal cell carcinoma were evaluated retrospectively. Region of interest was drawn to cover the entire lesion in the sections. Gray-level intensity (Hounsfield Unit value), entropy, standard deviation, uniformity, kurtosis, skewness, size% lower, size % mean, size% upper, values were obtained by texture analysis. The findings of both groups were compared statistically. Results: Mean and median gray-level intensity values and entropy values were significantly higher in renal cell carcinoma than in high-attenuation renal cyst (p<0.001). There was no significant difference in other parameters. When receiver operator characteristics analysis was performed for the mean value, the area under the curve value was found to be 0,754. When the threshold value was selected as 34.5708, 72.7% sensitivity and 66.7% specificity was found. Conclusion: Texture analysis may be useful in differentiating renal cell carcinoma from high-attenuation renal cysts on non-contrast computed tomography.

Keywords: