Quantifying ultrasound echogenicity difference for accurate chronic kidney disease diagnosis.

Abstract

Background: This study aims to evaluate the diagnostic ability of quantitative renal echogenicity for chronic kidney disease (CKD) using ultrasounds.

Methods: Between September 2018 and December 2019, patients with kidney disease at our hospital underwent ultrasound imaging and serum creatinine tests. Five radiologists assessed patient ultrasound image echogenic characteristics for CKD diagnosis. Novel quantitative imaging measurements were also collected. Estimated glomerular filtration rate (eGFR) was used to classify patients into CKD or non-CKD groups. The diagnostic accuracy of the radiologists' assessment and quantitative measurements was analyzed using Receiver Operating Characteristic (ROC) curves.

Results: In our study, 339 patients participated. Radiologists' visual analysis showed low accuracy (42.5-46.9%) for CKD diagnosis. Conversely, bilateral mid-renal cortex echogenicity and echogenicity difference value showed a strong correlation with eGFR. We identified two effective CKD indicators: the Echogenicity Difference Value between Right Renal Sinus and Right Renal Cortex, which achieved 0.838 AUC (area under the curve), 77.90% accuracy, 68.45% sensitivity, and 87.13% specificity; and the Echogenicity Difference Value between Left Renal Sinus and Left Renal Cortex, which showed 0.805 AUC, 75.22% accuracy, 70.24% sensitivity, and 80.12% specificity.

Conclusions: The radiologists' visual analysis of renal echogenic characteristics is insufficient for CKD diagnosis. Quantitative echogenicity analysis using the proposed echogenicity difference values is more objective and effective for diagnosing CKD.