Clinica Chimica Acta, 205 (1992) 19-30 © 1992 Elsevier Science Publishers B.V. All riots reserved 0009-8981/92/$05.00 CCA 05195
Discriminant analysis of urinary calculi by near-infrared reflectance spectroscopy Evelyne Peuchant, Xavier Heches, Daniel Sess and Michel Clerc Department of Biochemistry, Saint-Andrd Hospital, 33075 Bordeaux Cedex ( Fran~.e) (Received 20 May 1991; revision received 22 October 1991: accepted 23 October 1991)
Key words: Urinary calculi: Near-infrared reflectance analysis; Infrared spectroscopy
Summary Near-infrared reflectance analysis has been used to determine the qualitative and semi-quantitative composition of urinary calculi. This simple method requires calibration of the most common urinary calculi. Analysis time is short (< I rain) and only very small amounts of calculus in powder form (< !00 ~,g) are required without the use of any reagent. Moreover, when compared to infrared spectroscopy, this method provides sufficient accuracy to identify mixed calculi containing two or three components, and permits semi-quantitative determinations. The speed and simplicity of this technique makes it a powerful method for the routine analysis of urihary calculi in clinical laboratories.
Introduction Analysis of urinary calculi is a key step in determining iithogen metabolic disorders and in the development of a therapy to avoid recurrence [1-3]. Two basic types of technique (chemical and physical) have been proposed to identify the composition of urinary calculi [4]. Chemical methods are actually widely used in biological laboratories owing to the availability of commercial reagents (E. Merck, Darmstadt, FRG). However, these methods are known to have very low sensitivity and-t0 produce false positive or negative results [5]. Physical techniques are the techniques of choice to identify the molecular and crystalline composition of urinary calculi with certainty. Amongst these, X-ray diffraction and infrared (IR) specCorrespondence to: E. Peuchant, Department of Biochemistry, Saint-At~dr6 Hospital, 33075 Bordeaux Cedex, France.
20
troscopy offer the best means of identification [6-10]. However, the instruments are not often available in clinical laboratories due to their low capacity for routine analysis and their cost. The method proposed here is a physical technique using reflectometry in the nearinfrared (NIR) which allows the qualitative and semi-quantitative analysis of urinary calculi. This method is of interest for three reasons. m Firstly, the great accuracy of the results compared to those performed by IR spectroscopy allows the determination of the semi-quantitative composition of mixed calculi. Secondly, this method can be applied to the analysis of a wide range of biological parameters in serum or faeces [11,12]. Finally, the cost of the apparatus is moderate. Materials and Methods
Apparatus IR spectra were obtained on a Perkin-E~mer model 1310 Infrared Spectrophotometer (Norwalk, USA). NIR analysis was performed on a Technicon lnfraalyzer 450 (Bran and Luebbe, Alfa Laval Group Les Clayes sous Bois, France) equipped with a Hewlett Packard 86 B microcomputer. The software was the 'Diagnose' program from Technicon.
Reagents Samples of calcium oxalate monohydrate (whewellite) and calcium orthophosphate (whitocklite) were obtained from BDH Chemicals Ltd (Poole U.K.) and calcium phosphate hydroxide (hydroxyapatite) and calcium monohydrogen phosphate dihydrate (brushite) from Sigma Chemical Co. (St Louis, MO).
Sample preparation After macroscopic examination the calculus was divided in two. This procedure was necessary to determine the homogeneity or heterogeneity of the calculus [I 3]. Different concentric layers observed around a nucleation zone revealed a mixed composition of the calculus, in such cases each layer may be separated using a scalpel and ground into a fine powder before analysis. It may be noted that NIR analysis is not destructive and does not alter the composition of the powder sample and it may be recovered for future investigations as controls used in IR spectroscopy. This is particularly important when only microcalculi are available.
IR analysis A small quantity (< 100/~g) of the powdered sample was mixed with dry KBr (l/100-w/w) and was compressed into a pellet before mounting into the sample beam of the IR spectrometer.
21
The preparation of such a pellet is unnecessary in NIR analysis. The infrared spectrum of the sample was scanned from 200 to 4000 cm -I.
NIR analysis Apparatus The apparatus was fitted with a tungsten halogen lamp and 19 narrow bandwidth (10 nm) filters of fixed wavelength specification from 1445 nm to 2345 nm. High spectral sensitivity was obtained through the use of Kohler optics and an integrating sphere to increase the light measured [14]. The apparatus measures reflectance as the ratio of reflected energy from the sample and the incident radiation. Reflectance was measured at 19 wavelengths and subjected to logarithm conversion.
Sample presentation The calculus in powder form ( < 100 #g) was placed in a cavity (3 mm diameter and 0.5 mm depth) at the upper part of a thin aluminium rod in a conical metallic reflector (Fig. 1). The use of this reflector concentrates all the incident radiation on the sample [15], For improved accuracy, each sample was analysed four times [16].
SOURCE ILLUMXNATION
REFLECTOR
Side
\
View
/ STONE
Top View
\ /
SUPPORT
t j I
i 4¢.m
Fig. I. The special cup in the conical reflector used for calculus analysis.
22
Reference library A total of 40 renal calculi analysed by IR spectroscopy were selected as the reference library for the proposed method, and comprised the following: eight were composed of calcium oxalate (in varying monohydrate and dihydrate forms), seven of uric acid (anhydrous and dihydrate forms), four of ammonium urates, two of xanthine and three of cystine, ten were composed of different types of phosphate (magnesium ammonium phosphate, calcium phosphate, carbonate apatite). The others were composed of a mixture of two or three components.
Principle of the method Classification or identification of a sample by infrared spectroscopy is obtained by matching the location of absorbance peaks on a spectrum to those of known substances. The wavelengths corresponding to the characteristic absorption peaks of different substances in the IR range can be used to identify these substances. In the method proposed here this general principle was applied using 19 fixed filters corresponding to 19 wavelengths covering the near-infrared spectral region to discriminate urinary calculi. For this identification a reference set composed of pure
t~t'atr.e . . . ~
0.740 XanthL, te ,...
LLJ
U
|.*~
Z
Iq01~11 1tl~e.,,,,
++~ 4.
Discriminant analysis of urinary calculi by near-infrared reflectance spectroscopy Evelyne Peuchant, Xavier Heches, Daniel Sess and Michel Clerc Department of Biochemistry, Saint-Andrd Hospital, 33075 Bordeaux Cedex ( Fran~.e) (Received 20 May 1991; revision received 22 October 1991: accepted 23 October 1991)
Key words: Urinary calculi: Near-infrared reflectance analysis; Infrared spectroscopy
Summary Near-infrared reflectance analysis has been used to determine the qualitative and semi-quantitative composition of urinary calculi. This simple method requires calibration of the most common urinary calculi. Analysis time is short (< I rain) and only very small amounts of calculus in powder form (< !00 ~,g) are required without the use of any reagent. Moreover, when compared to infrared spectroscopy, this method provides sufficient accuracy to identify mixed calculi containing two or three components, and permits semi-quantitative determinations. The speed and simplicity of this technique makes it a powerful method for the routine analysis of urihary calculi in clinical laboratories.
Introduction Analysis of urinary calculi is a key step in determining iithogen metabolic disorders and in the development of a therapy to avoid recurrence [1-3]. Two basic types of technique (chemical and physical) have been proposed to identify the composition of urinary calculi [4]. Chemical methods are actually widely used in biological laboratories owing to the availability of commercial reagents (E. Merck, Darmstadt, FRG). However, these methods are known to have very low sensitivity and-t0 produce false positive or negative results [5]. Physical techniques are the techniques of choice to identify the molecular and crystalline composition of urinary calculi with certainty. Amongst these, X-ray diffraction and infrared (IR) specCorrespondence to: E. Peuchant, Department of Biochemistry, Saint-At~dr6 Hospital, 33075 Bordeaux Cedex, France.
20
troscopy offer the best means of identification [6-10]. However, the instruments are not often available in clinical laboratories due to their low capacity for routine analysis and their cost. The method proposed here is a physical technique using reflectometry in the nearinfrared (NIR) which allows the qualitative and semi-quantitative analysis of urinary calculi. This method is of interest for three reasons. m Firstly, the great accuracy of the results compared to those performed by IR spectroscopy allows the determination of the semi-quantitative composition of mixed calculi. Secondly, this method can be applied to the analysis of a wide range of biological parameters in serum or faeces [11,12]. Finally, the cost of the apparatus is moderate. Materials and Methods
Apparatus IR spectra were obtained on a Perkin-E~mer model 1310 Infrared Spectrophotometer (Norwalk, USA). NIR analysis was performed on a Technicon lnfraalyzer 450 (Bran and Luebbe, Alfa Laval Group Les Clayes sous Bois, France) equipped with a Hewlett Packard 86 B microcomputer. The software was the 'Diagnose' program from Technicon.
Reagents Samples of calcium oxalate monohydrate (whewellite) and calcium orthophosphate (whitocklite) were obtained from BDH Chemicals Ltd (Poole U.K.) and calcium phosphate hydroxide (hydroxyapatite) and calcium monohydrogen phosphate dihydrate (brushite) from Sigma Chemical Co. (St Louis, MO).
Sample preparation After macroscopic examination the calculus was divided in two. This procedure was necessary to determine the homogeneity or heterogeneity of the calculus [I 3]. Different concentric layers observed around a nucleation zone revealed a mixed composition of the calculus, in such cases each layer may be separated using a scalpel and ground into a fine powder before analysis. It may be noted that NIR analysis is not destructive and does not alter the composition of the powder sample and it may be recovered for future investigations as controls used in IR spectroscopy. This is particularly important when only microcalculi are available.
IR analysis A small quantity (< 100/~g) of the powdered sample was mixed with dry KBr (l/100-w/w) and was compressed into a pellet before mounting into the sample beam of the IR spectrometer.
21
The preparation of such a pellet is unnecessary in NIR analysis. The infrared spectrum of the sample was scanned from 200 to 4000 cm -I.
NIR analysis Apparatus The apparatus was fitted with a tungsten halogen lamp and 19 narrow bandwidth (10 nm) filters of fixed wavelength specification from 1445 nm to 2345 nm. High spectral sensitivity was obtained through the use of Kohler optics and an integrating sphere to increase the light measured [14]. The apparatus measures reflectance as the ratio of reflected energy from the sample and the incident radiation. Reflectance was measured at 19 wavelengths and subjected to logarithm conversion.
Sample presentation The calculus in powder form ( < 100 #g) was placed in a cavity (3 mm diameter and 0.5 mm depth) at the upper part of a thin aluminium rod in a conical metallic reflector (Fig. 1). The use of this reflector concentrates all the incident radiation on the sample [15], For improved accuracy, each sample was analysed four times [16].
SOURCE ILLUMXNATION
REFLECTOR
Side
\
View
/ STONE
Top View
\ /
SUPPORT
t j I
i 4¢.m
Fig. I. The special cup in the conical reflector used for calculus analysis.
22
Reference library A total of 40 renal calculi analysed by IR spectroscopy were selected as the reference library for the proposed method, and comprised the following: eight were composed of calcium oxalate (in varying monohydrate and dihydrate forms), seven of uric acid (anhydrous and dihydrate forms), four of ammonium urates, two of xanthine and three of cystine, ten were composed of different types of phosphate (magnesium ammonium phosphate, calcium phosphate, carbonate apatite). The others were composed of a mixture of two or three components.
Principle of the method Classification or identification of a sample by infrared spectroscopy is obtained by matching the location of absorbance peaks on a spectrum to those of known substances. The wavelengths corresponding to the characteristic absorption peaks of different substances in the IR range can be used to identify these substances. In the method proposed here this general principle was applied using 19 fixed filters corresponding to 19 wavelengths covering the near-infrared spectral region to discriminate urinary calculi. For this identification a reference set composed of pure
t~t'atr.e . . . ~
0.740 XanthL, te ,...
LLJ
U
|.*~
Z
Iq01~11 1tl~e.,,,,
++~ 4.
