New, noninvasive anemia-detection method uses smartphone images of fingernails and app
Biomedical engineers have developed a smartphone app for anemia screening that can assess blood hemoglobin levels through the window of the user’s fingernail. The medical results are based on the coloration of the fingernail bed; the quick and pain-free screening could benefit a vast number of people who are affected by anemia around the world. The team is led by researchers at the Wallace H. Coulter Department of Biomedical Engineering at the Georgia Institute of Technology and Emory University and the Aflac Cancer and Blood Disorders Center of Children’s Healthcare of Atlanta and the Department of Pediatrics at Emory University, in Atlanta, Georgia, and was funded by the National Institute of Biomedical Imaging and Bioengineering (NIBIB).
“Smartphones are part of the communications infrastructure far and wide, including in low-resource settings around the world,” said Tiffani Lash, Ph.D., director of the NIBIB programs in Point-of-Care Technologies and Connected Health (mHealth and Telehealth). “A self-monitoring approach to accurately assessing anemia, a medical condition that affects so many, will have far-reaching, positive impact.”
Anemia is a condition caused by a low red-blood-cell count or a low level of the protein hemoglobin in red blood cells. The iron in hemoglobin carries oxygen to fuel the many processes of metabolism that occur throughout our bodies. Anemia takes many forms, from a mild condition that can be improved by dietary changes, to chronic anemia, that requires vigilant monitoring and possibly transfusions of blood for a lifetime. All told, anemia affects a quarter of the world’s population, causing symptoms from pale coloration, fatigue, and listlessness, to more severe outcomes, including life-threatening heart failure.
People who are chronically anemic require frequent hemoglobin-level testing to monitor their disease and to guide their clinical treatment. The gold-standard test to measure hemoglobin is a complete blood count, based on a blood draw performed by trained technicians and processed in a laboratory using expensive equipment and reagents. Despite the high prevalence of anemia, there has been no noninvasive, inexpensive, and accurate hemoglobin assessment technology available that enables chronic anemia patients to better self-manage their disease.
In his lab at Georgia Tech and Emory University, Wilbur Lam, M.D., Ph.D., combines his expertise as a pediatric hematologist-oncologist and as a bioengineer. His team is dedicated to developing technologies to study, diagnose, and treat blood disorders. The team includes Robert Mannino, Ph.D., who has contended his entire life with an inherited form of chronic anemia, called beta thalassemia major, also known as Cooley’s anemia. He joined the lab while still an undergraduate student, and the smartphone app for anemia detection became his Ph.D. research project some years after, culminating in a paper published Dec. 4, 2018, in Nature Communications.
“When it came time for him to choose his thesis project, Rob came to me to say that he wanted to work on something that is related to his own disease,” said Lam. Lam recognized that Mannino would be the best person to try a pilot experiment based on smartphone image data over the course of several transfusion cycles.
Physicians are taught in medical school that anemia can lead to certain body parts being paler than others, such as the lips, right below the whites of the eyes, the creases on your palms, and finger nails. Fingernail beds contain minimal amounts of melanin compared to other parts of the skin, allowing the method to be adopted by people with any skin tone. Mannino sampled his own skin throughout the transfusion cycles, while he was anemic and not anemic, and compared the images for correlation to his blood draws.
“Sure enough, after a few cycles Rob showed me the data, and there was this perfect match between aspects of the metadata obtained in the phone and the actual hemoglobin values [from blood testing],” Lam said. “I remember saying: Okay, I think you’ve got your PhD project there.”
For the study published this past December, the team extended beyond Mannino’s one-person pilot data to assemble data from 337 male and female children and adults—a combination of 72 healthy individuals and 265 people with anemia from various causes, including sickle cell disease, beta thalassemia, hemophilia, and a variety of cancers. The researchers used a mathematical algorithm to calibrate the data from 237 of the participants that determined a hemoglobin-level value based on nailbed coloration, calibrated to their levels from a complete blood count test obtained during the same visit.
They used data from the remaining 100 participants to test the algorithm, which the researchers incorporated into mobile apps for Android and iOS smartphone devices. The smartphone method calculated hemoglobin levels close to those obtained via the traditional blood draw method (accuracy within 2.4 grams/deciliter and 97% sensitivity).
The researchers performed a study of fingernail image screening in four patients over several weeks using the system.
The algorithm in the app fed more data from the individuals’ coloration became even more fine-tuned in detecting hemoglobin levels for that participant (0.92 grams/deciliter), or just about the same as blood-based hemoglobin tests.
“Getting this out into the public is a major immediate thrust of our efforts,” Lam said. “We are now collaborating with people in the global health space, people in different countries, trying to create impact and constantly make it more accurate. The other thing that is on track is trying to get this out to people in the developing world as well.”
Mannino says that he and others with beta thalassemia major could readily benefit from the app. “Sometimes I’ll show up to my transfusion and my hemoglobin level will be a little bit lower than we’ve predicted it would be,” he explained. “If I had this technology, I could dynamically be setting my transfusion schedule based on when I am anemic and actually need one.”
Lam explained that there are two ways to use this device, either for screening amongst the general population, or for diagnosis in people who are chronically anemic. “The app can learn and weight its data towards the person’s blood counts,” Lam said. “Theoretically, the app would weigh the individual’s data higher, and therefore, recalibrate, which will further increase the app’s accuracy in estimating blood hemoglobin levels.” The study so far has validated the smartphone imaging of fingernail beds as a tool that would be effective for screening. More testing is required for it to be adopted and approved for making individual diagnoses.
The team is planning to make the app publicly available in the coming months, and they will continue to conduct individual calibration studies to confirm the high level of diagnostic accuracy that would be necessary to use the smartphone app in place of blood-based anemia testing.
Since playing his part in developing the smartphone app, Mannino has received the $100,000 Cisco Problem Solver Prize in April 2018 and the $100,000 Student Technology Prize for Primary Healthcare from Massachusetts General Hospital in October 2017.
For more about anemia, go to https://www.nhlbi.nih.gov/health-topics/anemia.
Funding for the project included a grant from the NIBIB (EB025646).
Reference: Smartphone app for non-invasive detection of anemia using only patient-sourced photos. Mannino RG, Myers DR, Tyburski EA, Caruso C, Boudreaux J, Leong T, Clifford GD, Lam WA. Nat Comms. 2018 Dec 4.