Improved Automated Urinalysis

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Current automated urinalysis instruments are unable to recognize crystals, casts, and their subtypes in urine sediment samples. This proposal seeks to develop a technique for automated classification of the types and subtypes of casts and crystals by exploiting birefringence phenomena. The resulting imaging and classifying technique will permit a quantum improvement in currently available automated urinalysis instruments. The casts and crystals in the urine are clinically important indicators of abnormalities. Current automated techniques are unable to classify the sub-types of casts and crystals. As a result, manual examination of suspect urine samples is necessary, thus wasting time and resources. The solution is to improve the classification techniques that such instruments use. This project aims to develop novel optical and pattern recognition techniques for microscopic particle analysis of casts and crystals in urine sediment. Since current urinalysis instruments using flow cytometry and image analysis are limited in their ability to classify the types and subtypes of casts and crystals, this project specifically intends to add functionality for sub-typing casts and crystals in urine. Our approach is to use polarization microscopy to obtain new features to aid classification of crystals, casts and other particles. Another area of interest is to explore the magnetic properties of these crystals and to use novel techniques to detect and identify the crystals by the influence that magnetic fields have on their birefringence properties. The aim of the Phase I project is to demonstrate the utility of birefringence as a unique 'signature' that can give rise to new features for classifying urine particles. If the Phase I feasibility criteria are met, then in Phase II we will build an automated urine particle analyzer incorporating polarization optics combined with current flow microscopy and digital image processing techniques. The resulting automated instrument will provide quicker and more accurate results, and will thus become an even more important tool for urine screening in clinical diagnostics. Current automated urinalysis instruments are unable to recognize crystals, casts, and their subtypes in urine sediment samples. This proposal seeks to develop a technique for automated classification of the types and subtypes of casts and crystals by exploiting birefringence phenomena. The resulting imaging and classifying technique will permit a quantum improvement in currently available automated urinalysis instruments.