Diagnostic Utility of Dysmorphic Erythrocyte Identification Using the Atellica 1500 Automated Urinalysis System in the Evaluation of Glomerular Hematuria

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Context.— Dysmorphic red blood cells (dRBCs) in urine are useful noninvasive markers for distinguishing glomerular hematuria. Objective.— To assess the diagnostic utility of dRBC identification, using image data from the Atellica 1500 Automated Urinalysis System (Siemens Healthineers). Design.— We retrospectively evaluated 162 patients from Severance Hospital (n = 46; August 2019–February 2024) and Yongin Severance Hospital (n = 116; between March 2020 and February 2024) who underwent renal, urinary, and/or urogenital biopsies during periods when the same Atellica 1500 Automated Urinalysis System was in use. dRBCs were evaluated by trained medical technologists following recent literature criteria. Inclusion requires hematuria (≥2+ on dipstick) and dRBC testing within 3 months before biopsy. From biopsy, imaging, and clinical findings, patients were classified as having glomerular hematuria (n = 90) or nonglomerular diseases (n = 72). Results.— The area under the receiver operating characteristic curve (AUROC) for %dRBC in diagnosing glomerular disease was 0.766. At cutoffs of 17% and 28% dRBCs, sensitivity was 32.2% and 21.1% with specificity of 95.8% and 97.2%, respectively. %G1-dRBC yielded an AUROC of 0.636, with sensitivities of 31.1% and 26.7% at 1% and 2% cutoffs, respectively, for corresponding specificities. In immunoglobulin A nephropathy (n = 67), %G1-dRBC was higher in cases with segmental glomerulosclerosis (n = 53). Multivariate analysis identified that younger age, higher %dRBC, lower dipstick blood/leukocyte esterase, and higher dipstick protein grades were independent predictors of glomerular diseases. Conclusions.— dRBC identification using digital urine sediment analyzer provides high specificity and positive predictive value for glomerular hematuria, with advantages in efficiency, reproducibility, and image archiving.