Researchers at the Georgia Institute of Technology have developed a smartphone app that was able to estimate hemoglobin levels by analyzing images of fingernail beds, creating, they wrote, an "on-demand system [that] enables anyone with a smartphone to download an app and immediately detect anemia anywhere and anytime."

The team detailed its work in the Dec. 4, 2018, online issue of Nature Communications.

One of the goals of senior author Wilbur Lam, who is both a pediatric hematologist at Emory University School of Medicine and a professor of engineering at the Georgia Institute of Technology, is to develop minimally invasive or noninvasive point-of-care diagnostics for hematologic and oncologic indications "with our eye on global health settings," Lam said. "So these devices have to be simple, cheap and at the same time still accurate."

The right person for the job

Lam told BioWorld that Robert Mannino, first author of the Nature Communications paper and now a postdoctoral scientist in Lam's laboratory, is quite possibly the only person in the world who could have developed the app.

"He's a brilliant computer programmer," Lam said, "with beta-thalassemia major... He gets transfused every month."

Right before those transfusions, Mannino is anemic. Right after them, he is not.

While beta-thalassemia is a comparatively rare disorder, anemia, which is a lack of red blood cells and the hemoglobin they transport, is most definitely not. It is the most common blood disorder in the world, affecting roughly a quarter of the global population, or somewhere around 1.6 billion people.

Its causes are extremely disparate. Factors from genetic mutations to iron deficiency to chemotherapy can bring on anemia.

No matter what its cause, though, one of its symptoms, and one that doctors use as an initial screen for anemia, is that "when a patient is anemic, there are certain body parts that are pale," Lam said.

Those body parts include the whites of the eyes, creases of the arms, beds of the fingernails, and others.

Mannino's illness enabled him to gather multiple datasets by repeatedly photographing himself before and after transfusion, and test algorithms that could separate out anemia-induced pallor from the multiple other factors that influence how skin color will be rendered in a photograph.

Using those data, the team showed that it was the color of the nail beds that correlated best with hemoglobin levels. In further work, they validated their algorithm that estimated hemoglobin levels from nail bed pictures with data from 100 individuals of different ages and ethnicities – Lam said that one advantage of using fingernail pictures is that "people with different skin colors, with different ethnicities, their fingernails are actually pretty similar."

The app currently works with smartphones that are roughly three years old or newer, and Lam said his team is also working to create a conversion "so new phones can adapt to the database... [we think we can] control for future models."

The app is only one of several minimally invasive or noninvasive blood test devices that Lam's team has developed. At the annual meeting of the American Society for Hematology this week, Lam's colleagues presented data on Anemocheck (Sanguina LLC), which was approved by the FDA in 2017. Anemocheck is minimally invasive, requiring less than a drop of blood to diagnose anemia.

Unlike Anemocheck, however, and despite the fact that the team demonstrated that with personal calibration, the algorithm could reach an accuracy that would enable clinical diagnosis, Lam and his group do not plan to seek FDA approval and turn the app into an FDA-approved device, at least not initially.

Instead, the plan is to make the app available as rapidly as possible – Lam said it will most likely be ready for distribution in the spring of 2019 – and follow the strategy used by the Apple watch of making physiological measurements available, but refraining from health claims about what those measurements mean.

The reason for doing so, Lam said, is that his group is prioritizing broad reach for the work. "The biggest impact this can have is the fact that there are many populations, demographics, that are at risk for anemia," he said. Being able to quickly determine whether feeling run down is a case of the blahs or a sign of anemia could be useful for populations including pregnant women, cancer patients and public health workers who work with populations at risk of lead poisoning, which can also cause anemia. "Now, people can screen themselves – once we get this on the app store – any time they want."