One important physiological measurement that a doctor wants to know is the amount of oxygen consumed in different parts of the body. The little instrument that your doctor clips on the end of your finger, a pulse oximeter, can measure SpO2, peripheral oxygen saturation, by looking at the change in the absorption of a specific wavelength of light due to the pulsing of arterial blood using a photodetector. This measurement is called a PPG, a photoplethysmogram. This measurement is correlated with SaO2, arterial oxygen saturation. The SpO2 is related to the amplitude modulation of the pulse signal.

However, to get a better measure of tissue oxygen consumption and blood flow, a measurement of SvO2, venous blood oxygenation, is also needed. Knowledge of both SpO2 and SvO2 would help physicians identify shock or cardiac failure. The usual method of monitoring SvO2 is to sample blood from a pulmonary catheter. This is invasive and has associated risks, especially in ill or injured patients.

In the paper, Simultaneous Estimation of Arterial and Venous Oxygen Saturation Using a Camera, by Mark van Gastel, Hanbing Liang, Sander Stujik, and Gerard de Haan, from the Department of Electrical Engineering at Eindhoven University of Technology, The Netherlands, the authors investigate making arterial and venous oxygen saturation measurements using FluxData’s FD-1665-MS3 multispectral camera. Successful results would mean that doctors could monitor oxygen saturation in a non-invasive way without requiring any instruments to contact the body.

Optical measurement of SvO2 has been studied before. Because veins have thinner walls and are less elastic than arteries, venous blood flow changes due to respiration and the volume of blood flow can be manipulated by applying pressure to parts of body. Taking advantage of this, one optical method used a strain gauge applied to a finger to modulate the venous flow and thus differentiate it from the arterial flow measured using PPG. Another method is to apply a static pressure to a limb to temporarily stop the venous flow of blood and differentiate SvO2 and SpO2 in that way. However, this method can lead to complications.

In this study, the FD-1665-MS3 was used to record five regions of interest (ROIs) on healthy subjects faces. The camera imaged the faces at three infrared bands centered at 760, 800, and 890 nm with 25, 25, and 50 nm bandwidths, respectively. The camera was recording at 15 fps. Nine incandescent bulbs, at a distance of 1.5 m from the face, were used to create diffuse homogeneous illumination.

An auditory breathing pattern was used. The subjects synchronized their breathing to this pattern during the 3-minute recording sessions. Because the venous blood flow volume changes with respiration, the experimenters could use this pattern to help differentiate SvO2 and SpO2. To help check the validity of the results, a finger pulse oximeter measuring PPG and pulse was used.

To reduce noise, the data in each ROI was averaged. The signals in each spectral band were normalized. Using temporal Fourier analysis, the experimenters could look at the amplitude of the signals at differenct frequencies and isolate those that are modulated by the pulse (40-240 bpm) and respiration (10-40 bpm) to calculate the SpO2 and SvO2 respectively. Other signal processing steps include the elimination of motion artifacts and the reduction of irrelevant signals.

A) DC-normalized signals from the 3 channels, B) corresponding spectra with peaks at both breathing and pulse rate, C) spectrogram 890nm channel from the full recording, D) signal after projection to eliminate motion artifacts, E) corresponding spectrum after amplitude tuning, F) spectrogram from the full recording.

The results were very promising. The camera-based measurements of the SpO2 corresponded well with the PPG from the finger pulse oximeter. The SvO2 measurements were fairly equal for the facial ROIs but showed higher variation than the SpO2 measurements and were in the range reported in the literature. (There was no empirical comparison for the SvO2 as there was for the SpO2.) The ability to measure both arterial and venous oxygenation with this non-contact camera method shows promise because it allows for greater patient comfort, fewer complications, and the ability to measure parts of the body where PPG is not possible. Follow-up studies are needed to further validate (with the invasive SvO2 gold standard) and refine this technique and to see how this method works with patients undergoing surgery with spontaneous or artificial respiration.

To view the full research article, click here.