Control blood oxygen levels in our own home and without the need for a pulse oximeter. That is the goal pursued by a team of researchers from the University of Washington and the University of California at San Diego. And it seems they have succeeded.
Knowing the oxygen saturation in the blood is something especially useful for people who suffer from a disease that affects the proper functioning of the lungs, such as COPD, heart failure, lung cancer or COVID-19 itself. Also if you have sleep apnea. This test, called pulse oximetry, is done by your doctor in the office or can be done at home with a pulse oximeter. In the future, we could do it ourselves with our mobile phone.
Researchers have managed to get a smartphone to detect blood oxygen saturations of up to 70% . All you have to do is place your finger on the phone’s camera and flash and wait for the result. The smartphone, however, uses a deep learning algorithm to decipher blood oxygen levels.
When the team administered a controlled mixture of nitrogen and oxygen to six subjects to artificially lower their blood oxygen levels, the smartphone correctly predicted whether the subject had low blood oxygen levels 80% of the time.
“Other smartphone apps that do this were developed by asking people to hold their breath. But people get very uncomfortable and have to breathe after a minute or so, and that’s before their oxygen levels drop.” have dropped enough to represent the full range of clinically relevant data,” said study co-senior author Jason Hoffman. “With our test, we were able to collect 15 minutes of data from each subject. Our data shows that smartphones could perform well right in the critical threshold range.”
Another advantage of the system they have developed is that almost everyone has a smartphone .
“In this way, multiple measurements could be made with the device itself at little or no cost ,” said co-author Dr. Matthew Thompson. “In an ideal world, this information could be seamlessly relayed back to the doctor’s office. This would be really beneficial for telemedicine appointments or for triage nurses to quickly determine if patients need to go to the ED or if they can continue.” resting at home and make an appointment with your GP later.”
Six people, three men and three women, between 20 and 34 years old, participated in the study. To collect data and train and test the algorithm, participants wore a standard pulse oximeter on one finger. Another finger of the same hand was placed on the camera and the flash of the mobile. Each participant had this same configuration in both hands simultaneously.
What the mobile camera does is record a video . Every time the heart beats, blood flows through the part illuminated by the flash. The camera also records how much light from the flash that blood absorbs. The color channels it measures are red, green, and blue. The researchers feed those measurements into their deep learning model.
Each participant breathed a controlled mixture of oxygen and nitrogen to slowly reduce oxygen levels. The process took about 15 minutes. For the six participants, the team acquired more than 10,000 blood oxygen level readings between 61% and 100%.
The researchers used data from four of the participants to train the mobile phone algorithm. The rest of the data was used to validate the method and then test it for performance on new subjects.
Further refinement still needs to be done and the study needs to be extended to more subjects, as there are circumstances where the algorithm struggled to make an accurate measurement of oxygen levels. This was the case of one of the participants who had calluses on his fingers. It could also have difficulties if people with different skin tones are given.
“Light from the smartphone can be scattered by all these other components on your finger, which means there’s a lot of noise in the data we’re looking at,” said co-lead author Varun Viswanath. “Deep learning is a really useful technique here because it can see these really complex and nuanced features and help you find patterns that you might not see otherwise.”
The team hopes to continue this research by testing the algorithm on more people. This is a first step towards the development of biomedical devices assisted by machine learning .
Referencia: Hoffman, J.S., Viswanath, V.K., Tian, C. et al. 2022. Smartphone camera oximetry in an induced hypoxemia study. npj Digital Medicine. DOI: https://doi.org/10.1038/s41746-022-00665-y
