Smartphone cameras and flashes could be used to measure blood oxygen levels in the future

The proof-of-principle study conducted by University of Washington and University of California San Diego researchers has been published in the journal npj Digital Medicine. The researchers have applied for a patent for the technology.

To measure the blood oxygen levels, study participants placed their fingers over the camera and flash module of a Google Nexus 6P smartphone. With every heartbeat, fresh blood flows through the part of the finger illuminated by the flash. The camera records a video to measure how much of the light from the glass is absorbed by the blood across three channels—red, green and blue.

This data from some of the participants were used to train a deep-learning algorithm to measure blood oxygen levels from the values. They then validated the model on the other participants. Not only did the technology accurately predict blood oxygen levels but it continued to do so when the research team brought down the blood oxygen levels of subjects by giving them a controlled mixture of nitrogen and oxygen.

There have been several previous studies on using smartphones for SpO2 levels, and there are also apps that claim to do the same but the purpose of the new study was to validate these measurements across a full range of clinically relevant SpO2 values.

“There were a few on the app store early in 2020, but some were removed due to concerns about their accuracy in the mid-2020s. There may still be apps available, but none are approved for medical use, and they are mainly marketed as fitness apps, rather than health apps. These can be dangerous if users do not understand the potential inaccuracy, and thus we believe our study paves a path forward to potentially improve the accuracy in the future,” said co-lead author Jason Hoffman to indianexpress.com in an email. Hoffman is a doctoral student at the University of Washington.

While the technology and configuration has currently only been tested with one smartphone model, the researchers are optimistic that it can be used on a larger variety of phones in the future. “We haven’t tested enough phone models to validate any hardware limitations, so this is speculation, but we are optimistic that this technique would work on any modern smartphone with a camera and flash that are co-located,” said Hoffman.

Also, since the researchers needed to make configuration changes to the smartphone before using it to take measurements, there is a need for the technology to be tested on different smartphones before it can see wider use.

“If we want to become the first FDA-cleared test for SpO2 on an unmodified smartphone, we’d need to perform a larger study to validate on a larger population, and then send our data to the FDA for evaluation and approval. We’d estimate this whole process would take 18-24 months after the commencement of the new study, though we don’t have concrete plans to perform this study at this time,” added Hoffman.

The data collection portion of the study happened several years ago when the Nexus 6P was a current model. The researchers were spurred by the urgency of the COVID-19 pandemic but it took several years to achieve the results. After that, it took two more years for the paper to make it past the peer review process. Also, the data collection method is expensive, meaning the researchers won’t have the resources to re-collect the data with a different model unless a follow-up study is conducted.

Co-authors of the paper include Xinyi Ding, a doctoral student at Southern Methodist University; Eric Larson, associate professor at Southern Methodist University; Caiwei Tian, who completed this research as a UW undergraduate student; and Shwetak Patel, a UW professor. The research was funded by the University of Washington.