Measuring oxygen saturation in the retina with a modified fundus camera

Oxygen deprivation to the brain (hypoxia) is an identified problem for military
pilots and can be caused by G-forces and flying at altitude. Hypoxia can lead
to impaired decision making and even unconsciousness. Unfortunately, there is
still no suitable technology available to measure cerebral oxygen supply in
pilots. Over the years, TNO and the Center for Man and Aviation (CML) have
tested several non-invasive sensors, such as the Finapres® (measuring blood
pressure with a finger clip sensor) and NIRS (Near InfraRed Spectroscopy
through the skin). The problem with these sensors, however, is that they do not
measure close enough to the brain (finger clip sensor), or provide relative
values **(NIRS), which are not predictive of the oxygen supply in a pilot's
brain.
In a previous feasibility study (TNO 2021 R12701, Feasibility Study on Retinal
Pulse Oximetry using Near InfraRed (NIR) Light) we investigated a new, more
direct way of measuring cerebral oxygen saturation and arterial blood pressure,
using a visually accessible offshoot of the brain: the retina. The brain and
retina are supplied with oxygenated blood by the same artery and are
geometrically close to each other, making it plausible that the oxygen supply
and blood pulsation/pressure in the brain are reflected in the retina. Although
measuring the oxygen content of the retinal vessels (retinal oximetry) had been
done before, our approach differs due to the wavelengths used. Unlike clinical
monitoring of retinal oxygenation, our method needs to be invisible to the
military user, so we use near-infrared (NIR) light that is invisible to the
human eye. In this wavelength range, blood absorbs little, which makes it
challenging to extract reliable information from the retina. That's why it has
never been done. To vary the atmospheric pressure, the feasibility study used
the hypobaric chamber of the CML. Altitudes of 0, 6000, 9000 and 12000 ft have
been simulated, leading to an oxygen saturation (SpO2) varying between
approximately 80-100%. A comparison with a finger clip sensor (Nonin WristOx
3150 with 8000SM-WO2 soft sensor) suggests that it is quite possible to
estimate SpO2 based on NIR images of the retina. In contrast, the estimated
heart rate (HR) was unreliable, possibly due to interfering ambient light. In
the current study, we want to expand the dataset with measurements on a larger
group of test subjects, improve the lighting conditions and filter wavelengths
of the fundus camera.

The importance of this research is that this technology, in collaboration with
industry, could eventually be integrated into the helmet mounted display (HMD)
of pilots. Such technology can detect oxygen deficiency in the brains of pilots
at an early stage and thus prevent accidents.

Primary Objective:
The primary objective is to investigate whether the retinal SpO2 - measured
with invisible, NIR light - changes as a function of simulated height in the
same way as the SpO2 measured with the standard finger clip.

Secondary Objective:
The second goal is to investigate whether improved lighting conditions compared
to the earlier feasibility study make it possible to measure HR in the retina
with NIR light (in addition to SpO2).

It concerns a validation study of a measurement method. The criterion validity
of a new measurement method for measuring oxygen saturation and heart rate is
mapped by determining correlations with the gold standard method. Using
repeated measurements, participants are exposed to four simulated heights.

The burden on participants is low, the procedures are standard and the risks,
partly due to the medical examination prior to participation, are low.
Participants are instructed to report if they are not feeling well and oxygen
is administered. In addition, participants are monitored during the experiment
by a physician familiar with hypoxia-related phenomena. The retinal images are
non-invasive and are made with a modified CE-certified fundus camera, using a
limited amount of NIR light.