Wavelength Dependence of the Differential Pathlength Factor and the Log Slope in Time-Resolved Tissue Spectroscopy

Abstract

The monitoring of tissue oxygenation by the technique of near infrared spectroscopy (NIRS) was first described by Jöbsis in 1977 (Jöbsis, 1977). The technique relies upon the relative transparency of tissue to near infrared (NIR) light to enable measurements of changes in optical attenuation across many centimetres of tissue. Early NIRS measurements could only derive qualitative changes in tissue and blood oxygenation from the observed variations in tissue attenuation (Brazy et al., 1985). However, data on the optical pathlength of light in tissue, measured by time resolved techniques employing picosecond laser pulses, have now permitted a quantitative analysis of attenuation measurements to be made (Delpy et al., 1988; Wyatt et al., 1990a). By incorporating information on the optical pathlength into a modified Beer-Lambert law it is possible to quantify changes in chromophore concentration from the measured changes in tissue attenuation. The optical pathlength needed in this calculation, the Differential Pathlength (DP) is defined as the local gradient in a plot of the attenuation measured in a scattering medium versus the absorption coefficient of the medium (Cope et al., 1991a; Cope, 1991b). It has been shown in previous studies (Delpy et al., 1988) that the DP can be approximated by measuring the mean distance that a picosecond light pulse travels across the tissue. Furthermore, a dimensionless multiplying factor, the Differential Pathlength Factor (DPF), can be obtained when the DP is divided by the geometric distance between light source and detector on the tissue surface. This factor has been shown, both theoretically and experimentally, to be approximately constant for any tissue once the optode spacing is larger than about 25 mm (van der Zee et al., 1990; van der Zee et al, in press), enabling clinical NIRS measurements to be made with varying optode geometries.