Translator Disclaimer
23 March 2007 Advanced imaging technique for automated classification of casts and crystals in urine
Author Affiliations +
We present the development and demonstration of a novel technique for microscopic analysis of urine particles. Casts and crystals in urine are indicative of clinically important abnormalities. Current urinalysis techniques using flow cytometry and image analysis are limited by their inability to detect, identify and classify crystals and casts. Casts, crystals and yeast cells reported by current automated urine particle analyzers must be confirmed by a second microscopic review involving a human operator to prevent false positives. Human examination of suspect urine samples is resource intensive and time consuming. We introduce a new imaging method to add functionality for recognition of casts and crystals in urine. Our approach uses a polarization microscopy technique to aid classification of crystals, casts and other urine particles. Crystals and casts in urine exhibit unique interference patterns when imaged using a fixed polarizer and analyzer in a crossed configuration. These interference patterns are a measure of birefringence (retardation angle) of the cast or crystal being imaged. Preliminary experiments indicate that uric acid shows a polarization color, and larger crystals exhibit a series of concentric black lines. We show that these unique 'signatures' when used in conjunction with a hierarchical pattern recognition technique can reliably classify the analytes with improved accuracy. The new imaging technique combined with the classification algorithm can address the shortcomings of current urinalysis techniques; and provide quicker and more accurate results.
© (2007) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Amit S. Paranjape, Kenneth Castleman, Thomas E. Milner, and H. Grady Rylander "Advanced imaging technique for automated classification of casts and crystals in urine", Proc. SPIE 6424, Photonic Therapeutics and Diagnostics III, 64240X (23 March 2007);

Back to Top