2021 Volume 70 Issue 3 Pages 456-464
We evaluated the performance of AUTION EYE AI-4510 (ARKRAY, 2019), an automatic urine sediment analyzer that is based on the unstained image analysis method. Our results showed that its basic performance, such as parallel accuracy and indoor reproducibility, was satisfactory. In comparison with microscopic examination, the sensitivity and specificity for red blood cells, white blood cells, squamous epithelial cells, hyaline casts, and bacteria were 69.0% and 95.1%, 60.9% and 99.4%, 34.1% and 99.6%, 52.6% and 77.4%, and 48.5% and 93.8%, respectively. The reason for such differences was considered to be the difficulty in distinguishing similar components owing to insufficient shape recognition. However, sequential version-up improved the specificity. Improvement of working efficiency, shortening of examination time, and accurate detection of urine sediments are required for this automatic urine sediment analyzer. This device has basic performance comparable to that of microscopic examination; thus, it is considered that this device could be introduced into routine laboratory use.
