Automated Fourier Transform Infrared Analysis of Urinary Stones: Technical Aspects and Example of Procedures Applied to Carbapatite/Weddellite Mixtures

• Published in: Applied spectroscopy 2004-06, Vol.58 (6), p.671-678
• Main Authors: Cohen-Solal, Fabrice, Dabrowsky, Béatrice, Boulou, Jean Claude, Lacour, Bernard, Daudon, Michel
• Format: Article
• Language: English
• Subjects: Algorithms; Apatites - chemistry; Autoanalysis - instrumentation; Calcium Oxalate - chemistry; Crystallization; Humans; Spectroscopy, Fourier Transform Infrared - methods; Urinary Calculi - chemistry
• ISSN: 0003-7028; 1943-3530
• DOI: 10.1366/000370204872962

New software (ScanLith) was developed to provide an automated procedure for identifying and quantifying crystalline species in urinary calculi. The first step was to strictly define operative conditions of sample preparation because they have a significant influence on the stability of various crystalline phases. Second, we determined the quantification coefficients of a polynomial curve required to develop the automated procedure. This was illustrated by the study of carbapatite and weddellite, as both constituents represent one of the most frequent associations found in stones analyzed in our laboratory. The following quadratic equation was obtained with a correlation coefficient r2 = 0.9997: Y = −0.5144x2 + 1.5239x – 0.0141, where Y is the absorbance ratio carbapatite/weddellite and x is the percentage of carbapatite in the mixture. The absorbance of carbapatite and weddellite was measured at 1035 cm−1 and 1325 cm−1, respectively. The third step was to validate the ScanLith procedure in routine analysis by comparing computed results with those of an expert. Concordance (r2 = 0.9824) was better than that previously reported using various computerized systems with a mean deviation of 4%. Algorithms developed in ScanLith to identify main and minor components found in urinary stones, even in complex mixtures containing up to seven constituents, allowed us to lower the detection threshold down to 1 to 10% depending on the main component.