Diagnostic Performance of Gray-scale Sonographic Findings for the Detection of Acute Pyelonephritis

ABSTRACT

PURPOSE: To evaluate the diagnostic performance of gray-scale renal sonographic findings for the diagnosis of acute pyelonephritis (APN) by using computed tomography as a reference standard. MATERIALS AND METHODS: We retrospectively reviewed gray-scale renal sonographic findings of 48 patients for the detection of APN. All patients had clinical symptoms such as fever, flank pain, or dysuria and were confirmed as APN by contrast-enhanced CT. The presence of sonographic findings such as renal swelling, alteration of the parenchymal echogenicity, wall thickening of the renal pelvis, loss of the renal sinus fat echogenicity, and loss of the corticomedullary differentiation were evaluated. We also categorized all patients into mild APN or severe APN groups according to the volume of the morbid renal parenchyma on contrast-enhanced CT, and evaluated the aforementioned sonographic findings between the two groups. RESULTS: Overall diagnostic sensitivity, specificity, and accuracy of gray-scale renal ultrasonography (US) for the detection of APN were 32.5%, 72.0%, and 58.5%, respectively. The sensitivity and specificity of each sonographic finding were measured for each group. Renal swelling sensitivity and specificity were 33.8% and 70.8% for the mild APN group, but 45.8% and 66.7% for the severe APN group. Sensitivity and specificity for alteration of the parenchymal echogenicity were 41.7% and 79.2% for the mild APN group, but 58.3% and 66.7% for the severe APN group. The sensitivity and specificity for wall thickening of the renal pelvis was 37.5% and 95.8% for the mild APN group, but 50.0% and 95.8% for the severe APN group. The sensitivity and specificity of loss of the renal sinus fat echogenicity were 12.5% and 83.3% for the mild APN group, but 12.5% and 91.7% for the severe APN group. The sensitivity and specificity of the loss of the corticomedullary differentiation were 12.5% and 95.8% for the mild APN group, but 20.8% and 75.0% for the severe APN group. There was no significant difference of gray-scale renal US diagnostic accuracy for the detection of APN between the mild and severe APN groups (56.3% 58.3%; p > 0.05). CONCLUSION: Although overall gray-scale renal US has poor sensitivity for the detection of APN, wall thickening of the renal pelvis is the most specific sonographic finding in the both mild and severe APN groups.