The AIF of an optical dye can be measured using a pulse dye densitometer (PDD). However, PDDs are specialized, expensive devices that don't function at the high heart rates of small animals or infants. Now, researchers from the Western University and the Lawson Health Research Institute in London, ON, and Dartmouth College (Hanover, NH) have shown that it's possible to measure the AIF using a pulse oximeter, a device found in virtually every emergency room and intensive care unit worldwide (Phys. Med. Biol. 57 8285).

"Our group is primarily focused on developing optical devices that measure cerebral blood flow quantitatively, especially in the intensive care unit," explained PhD candidate Jonathan Elliott. "We hope to remove some barriers associated with collecting the AIF by describing a pulse oximeter-based approach. For patients such as traumatic brain injury patients and low birth-weight infants, cerebral blood flow monitoring could become standard-of-care in the future."

Determining dye concentration

Pulse oximeters work by recording light absorption at wavelengths corresponding to the absorption of oxyhaemoglobin and deoxyhaemoglobin (the "red" and "infrared" channels). The measured absorption, in combination with the extinction coefficients of the two chromophores, can be used to determine oxygen saturation (SaO2).

When an optical dye is injected into the bloodstream for tracer kinetic modelling, the resulting increase in absorption is interpreted by the pulse oximeter as an increase in deoxyhaemoglobin. This leads to a negative change in the measured SaO2, the intensity of which can be used to quantify the AIF.

To exploit this effect, Elliott and his colleagues derived a mathematical relationship between arterial dye concentration and the change in SaO2. The model, which requires knowledge of the dye's extinction coefficients, the true arterial oxygen saturation (measured prior to dye injection) and the total haemoglobin concentration (determined from a venous blood sample), provides the time-dependent arterial dye concentration (AIF).

For cerebral blood flow monitoring, the team also optically measure the tissue concentration of the dye as it passes through the brain. These data and the calculated AIF can be used in a kinetic model to recover a series of physiological parameters. For this work, the AATH (adiabatic approximation to the tissue homogeneity) model was used to determine parameters including blood flow, capillary transit time and permeability-surface area product.

Error analysis

The team performed an error analysis to highlight potential limitations of the approach. They first simulated the achievable measurement resolution (minimum detectable change in dye concentration), using a small-animal pulse oximeter and the FDA-approved optical tracer indocyanine green (ICG).

Varying the red channel of the dye from 600 to 900 nm revealed that the highest resolution occurred when the wavelength corresponds to the dye's absorption peak (802 nm for ICG in plasma). At this wavelength, a pulse oximeter with a measurement resolution of 0.1% can resolve 0.01 mM changes in dye concentration. The authors note that a standard pulse oximeter with a 650 nm source resolves concentrations of about 0.5 mM.

The relative effect of the dye on the signal will depend on the total haemoglobin concentration (tHb). Thus the researchers simulated varying tHb from 5 to 25 g/dl, at fixed SaO2 and 760 nm. The resolution varied from 0.08 mM to 0.5 mM as tHb increased from the lower to the upper end of this range.

These findings suggest that resolution can be improved by using higher dye concentrations, constructing custom oximeter probes with a wavelength at the peak absorption of the dye being used, and being aware of the potential tHb-related pitfall.

In vivo comparisons

The researchers used a rabbit model to compare the pulse oximetry method with the PDD. Four anaesthetized rabbits were intravenously injected with ICG solution. During each injection, data were acquired simultaneously by oximeter and PDD probes placed on separate hind limbs.

Plots of calculated AIF appeared similar for the two devices. The researchers then used the AIF from the pulse oximeter to extract kinetic parameters from the "true" concentration curve generated from PDD-measured AIF. No significant differences were seen in the parameters obtained using the two AIFs. Across the four rabbits, mean differences in peak concentration, blood flow, permeability surface-area product and capillary transit time were –7.4%, –8.5%, 1.7% and 1.5%, respectively.

The team concluded that it should be fairly easy to convert an off-the-shelf pulse oximeter into a dye densitometer that can acquire AIF for tracer kinetic studies. "While kinetic modelling has been a mainstay of MRI, CT and PET for decades, it is a relatively new concept in optics and only a handful of groups have explored the tremendous possibility of quantitative dynamic contrast-enhanced optical imaging," Elliott told medicalphysicsweb.

Related articles in PMB
Arterial input function of an optical tracer for dynamic contrast enhanced imaging can be determined from pulse oximetry oxygen saturation measurements
Jonathan T Elliott et al Phys. Med. Biol. 57 8285
Tracer kinetic modelling in MRI: estimating perfusion and capillary permeability
S P Sourbron and D L Buckley Phys. Med. Biol. 57 R1