Kidney stone analysis using FTIR spectrometry

A Kidney Stone Library & Analysis Kit has been developed to assist in the identification and characterisation of kidney stones by spectral analysis.

Mankind has always suffered from calculi in the efferent urinary tract. For example, urinary calculus was found in the pelvic area of a young man in a tomb dating back to 4800 BC near El Amrah, Egypt. However, it was not until the end of the 18th century that the first reports were published on the chemical composition of urinary calculi. At that time, important chemical constituents of urinary calculi were discovered, such as uric acid and cystine.

After the systematic studies by Heller (1847) and Ultzmann (1882), characterisation of urinary calculi by chemical analyses was, in principle, an established routine.

The diagnostic usefulness of information regarding the chemical composition of renal stones has been recognised since the 1950s and has improved so that it is now possible to correlate the results of an analysis with the appropriate diagnosis and therapeutic regimen.

Methods of analysis

No one method is sufficient to provide all the clinically useful information on the structure and composition of kidney stones.

Methods which have been used include infrared spectroscopy, polarisation microscopy, wet or dry chemical analysis, AAS, Roentgen - structural analysis, thermogravimetric analysis, porosity determination, pyrolysis gas chromatography, neutron activation analysis and solid-phase NMR.

A combination of refined morphological and structural examination of kidney stones with optical microscopy, complemented by compositional analysis via infrared spectroscopy of the core, cross-section and surface of calculi, provides a precise and reliable method for identifying the structure and crystalline composition, and permits quantification of stone components while being highly cost effective.

Composition of kidney stones

Stone components may be mineral, organic or both. More than 65 different molecules (including 25 of exogenous origin) have been found in urinary calculi. Extensive analysis of morphology and composition has led to a classification of urinary stones in seven distinctive types and 21 subtypes, including monohydrate (whewellite) and dihydrate (weddelite) calcium oxalates, phosphates, uric acid, urates, protein, and cystine (amino acid) calculi.

Confusing the matter further, the same chemical component may crystallise in different forms. Therefore, proper stone analysis has to identify not only the molecular species present in the calculus, but also the crystalline form.

Most stones are of mixed composition. About 80% are made of a mixture of calcium oxide (CaOx) and calcium phosphate (CaP) in various proportions.

The presence of other compounds, like 2,8 dihydroxyadenine, xanthine, cystine, calcite, etc, places the stone into a specific type of urolithiasis. Quantitative evaluation of all components is needed to provide full diagnostic information.

Quantitative analysis by infrared spectroscopy

There are at least two approaches to the quantitative or semiquantitative analysis of mixtures.

Partial least squares (PLS) techniques will yield highly precise results if the composition of the unknown material is restricted to a reasonably well-defined range, with predictable components present. This procedure is not well suited to kidney stones, because the range of concentrations is wide, and an unpredictable number of components can be present. Modern PLS software (such as TQ Analyst) can be used, but this results in greater complexity. Additionally, unmodelled artefacts cannot be identified using PLS software.

Second, library searching or other comparative algorithms can be used. For this to work, a spectral library of real kidney stones must exist. An unknown sample spectrum is then compared to a number of library spectra and the best correlated spectrum is found. Both the constituents and their concentrations of the matched spectrum are known.

A match value close to 100 indicates that the sample consists of the same components in about the same ratio.

Creation of the software

The Kidney Stone Library & Analysis Kit was created by spectroscopists and medical doctors to allow analysis of kidney stones using Nicolet FT-IR spectrometers with OMNIC software. It consists of three parts: the basic kidney stone library containing approximately 800 spectra; the advanced library containing approximately 18,000 spectra and an algorithm to work with; and the kidney stone guide containing additional analysis information.

The aim of this work was to create an automated FT-IR analyser for kidney stones, which provided a qualitative and quantitative analysis in one step, and then to connect the analysis result directly to information about diagnosis and therapy for the kind of stone found. After discussion with medical doctors, Thermo did not connect the results directly to the diagnosis text, because in diagnosis and therapy, factors other than stone composition are considered. However, this text remains a part of the Kidney Stone Guide in a consultative role.

The first step in building the software was to obtain spectra of all possible kidney stone mixtures. This is theoretically possible, because the number of components is limited and the mixtures build a closed (semi-stoichiometric) set. Even so, the number of possible mixtures is too high to allow collection of real kidney stones in all combinations.

Fortunately, the spectral contribution of each component is strictly additive, so we could take spectra of single component stones and then mix them, building all theoretically possible two- and three-component mixtures.

The concentration of the components in the mixtures ranges from 0-100% with the step of 5% for two-component mixtures and 10% for three-component mixtures.

A software routine based on OMNIC Macros\Pro was created for this purpose.

Consideration of mixtures with more than three components would increase the number of spectra excessively. Fortunately, stones with more than three components are rarely of clinical interest, and, more importantly, this type of stone is rarely found in humans.

It is also known that not all components build mixtures in all possible ratios, so these combinations were excluded.

The calculated spectra of mixtures were used to build two libraries, for use in basic and advanced analyses.

First, a flexible, basic library of approximately 800 of the most frequent mixture types was created, which can be used as a standard spectral library in the OMNIC search routine to identify the major components of an unknown stone. Custom spectra from the end user can be added to this library.

This library is easy to use, but not exhaustive. It contains a high number of similar spectra, so matching can result in high quantitative precision, but may also yield several similar match values when using OMNIC search directly, leading to ambiguous results.

Precision of the analysis

Any discussion of precision in this type of analysis is not simple, since achievable precision varies with type of concrement, percentage of components, baseline correction and the amount of impurities. If the content of a component is about 10%, the results are generally not considered diagnostically reliable. The reproducibility of the result can also be influenced by the inhomogeneity of the stone.

Thermo has optimised the algorithm using about 500 stones, where the concentration of all components was known from other methods. In most cases (about 85%) the accuracy was better than ±5%. According to the literature, an error of 10 to 15% is not of clinical interest, so the accuracy is sufficient.

Thermo strongly recommends that the user visually compare the unknown sample spectrum with the theoretically calculated spectrum. Furthermore, they should consider the pure components interpretation guide (included with the software) and study the morphological features of the sample compared to pictures. Use of an independent reference method is recommended if the reliability factor is not close to 100.

The Kidney Stone Library & Analysis Kit speeds up the analysis and can give excellent results, but a reliable analysis ultimately resides with the end user. This is why additional information is available in the Kidney Stone Guide.

This information includes the interpreted infrared and Raman spectrum of a stone and related pure components, pictures of example stones, discussion of other methods of chemical analysis, discussion of the causes and occurrence of particular stones or components, optical properties, tables of characteristic peaks, structural formulas and other information.

The guide provides information on the medical aspects of kidney stones, including diagnosis and therapy; however, it is not intended to be definitive for making medical decisions.