Applicability of the Kubelka-Munk theory for the evaluation of reflectance spectra demonstrated for haemoglobin-free perfused heart tissue.

Reflectance spectrometry is a useful tool for studying in vivo kinetic changes in the oxygen saturation of haemoglobin and myoglobin as well as the redox state of cytochromes. A method is given which allows the quantification of tissue reflectance spectra using multicomponent analysis. This method utilizes the Kubelka-Munk theory for modelling the measured tissue spectra. To test this approach, reflectance spectra of a haemoglobin-free perfused guinea pig heart were measured by a fast scanning spectrophotometer (100 spectra/s, spectral resolution 1.0 nm) and evaluated using the component absorbance spectra measured separately. A relative mean spectral residual error of 0.15% was achieved by least-squares fitting. Using statistical error propagation, oxygenation of myoglobin is obtained within a relative precision of 1%, and the redox state of cytochromes aa3 and c are determined simultaneously within a margin of 3%; the results for the redox-state of cytochrome b, however, are less precise. Special component error functions are presented to provide a reliability measure for the concentration prediction using this multicomponent assay. The consistency of the theory and the component absorptivity data is tested by regressing the actual concentrations obtained for each of the redox pair components during the various states of tissue oxygenation. A method is described for the recognition and reduction of systematic errors.